Series B 融资额 01
2100 $M [CI008] 尽调报告
Isomorphic Labs
AlphaFold 3 商业化——$16B 估值下先观察，等待 Phase I 验证
Isomorphic Labs 凭 AlphaFold 3 商业独占权和三项已签药企合作，拿住 AI 药物设计最强 IP 位置；但 $16 billion 私募估值押注里程碑转化，公开临床证据尚未支撑，需等待 ISM8969 I 期数据和首笔里程碑付款，暂列观察。
封面要素
公司概况
Isomorphic Labs 由 Demis Hassabis 与 Google DeepMind 同事于 2021 年初在伦敦创立，开发 IsoDDE——一个整合 AlphaFold 3 结构预测和 AlphaProteo 蛋白结合分子设计的 AI 药物设计与工程平台。公司作为 Alphabet 独立子公司运营。截至 2026 年 5 月，公司拥有 3 个活跃药企合作（Eli Lilly、Novartis、Johnson & Johnson），两笔已披露交易的合计里程碑潜力最高达 $3.945B；首个全资临床候选物 ISM8969 已获 IND 放行；并在 2026 年 5 月完成由 Thrive Capital 领投的 $2.1B Series B 轮。公司尚未报告任何里程碑收入，也没有获批药物；所有未来收入都取决于合作项目推进。
- 成立时间
- 2021-01-01
- 创始人
- Demis Hassabis, Pushmeet Kohli
- 创立地点
- London, UK
- 总部
- London, UK
- 产品
- IsoDDE 平台：AlphaFold 3 结构预测（商业独占访问）、AlphaProteo 蛋白结合分子设计、AI 引导的生成式化学和 ADMET 画像。以全栈定制研究合作交付，并接入药企合作伙伴的生物学团队。AlphaFold Server 面向全球学术和政府研究人员，提供非商业结构预测公共访问。
- 客户
- 按研发开支计的全球 Top-20 药企，需要用 AI 加速复杂多靶点小分子和生物药发现项目；短期可触达范围是 10-15 家一线药企，它们具备湿实验室能力，可将 AI 设计候选物推进临床。
- 商业模式
- 定制化多年研究合作：签约支付首付款，在候选物选择、IND、Phase I 入组等明确药物发现和开发节点支付里程碑款，并在获批药物商业化后收取特许权使用费。无 SaaS 或按席位许可模式；收入完全由交易和项目驱动。
- 阶段
- Private (Series B, May 2026)
- 融资情况
- 累计融资约 $2.7B：$600M Series A 轮（2025 年 4 月，领投方未披露，Alphabet 参与）；$2.1B Series B 轮（2026 年 5 月，Thrive Capital 领投，共同投资方包括 Alphabet、GV、MGX、Temasek、CapitalG、UK Sovereign AI Fund）。两轮投后估值均未正式披露；第三方估计 Series B 投后估值约 $16B。
执行摘要
主要优势
- 截至 2026 年 5 月，AlphaFold 3 结构预测的唯一商业访问权筑起结构性护城河，竞争对手无法获得
- 已签三家一线药企合作伙伴（Lilly、Novartis、J&J），其中两项已披露交易合计最高里程碑潜力达 $3.945B
- Novartis 在签约 13 个月后于 2025 年 2 月扩展合作，这是目前最强留存信号；未见合作伙伴终止报道
- Thrive Capital 领投的 $2.1B Series B 估计可支撑 6-17 年现金跑道；ISM8969 于 2026 年 1 月获 IND 放行，验证内部管线执行力
- Alphabet 背书带来近乎无限的算力和人才入口；Demis Hassabis（2024 年诺奖）提供难以匹敌的科学信誉
主要风险
- 迄今未收到里程碑付款；未来收入全部取决于合作项目临床推进，而触发条件未定义
- 客户极度集中（三家药企合作伙伴 = 100% 已知收入；J&J 交易条款未披露）
- AlphaFold 3 非商业许可证让平台护城河暴露在开源复刻下；商业独占权超过 12 个月后缺少明确法律围栏
- FDA/EMA 监管仍不确定：尚无 AI 生成候选药物完成 II 期，AI 设计分子的监管路径未定义
- $16B 估值约为已确认首付款收入的 194×；最接近可比公司 Recursion 临床组合更宽，市值约 $3.5B
未决问题
- J&J 合作财务条款（首付款、里程碑安排）未公开披露，真实组合收入潜力无法评估
- ISM8969 I 期安全性和有效性数据尚未出炉；靶向适应症和作用机制未公开披露
- Series B 投后估值未获官方确认；所有 $16B 估计均来自第三方
- 尚无药企合作伙伴披露 IsoDDE 合作的项目表现数据（命中率、周期时间、候选物选择事件）
- 除初始公告措辞外，AlphaFold 3 商业独占权期限未获公开确认
目录
01公司概况
1.1 公司身份与商业模式
Isomorphic Labs Limited 是 Alphabet Inc. 子公司，于 2021 年 2 月 24 日在英国注册成立（UK Companies House 编号 13223825），并于 2021 年 11 月 4 日公开宣布。公司总部位于英国伦敦，并在瑞士洛桑（2022 年 12 月开设）和美国马萨诸塞州剑桥（2025 年中开设）设有办公室。“isomorphic” 这个名字体现了创始人对生物学与信息科学之间深层对称性的判断——形态相同、来源不同——也把公司定位为计算 AI 与湿实验室药物发现之间的桥梁。 商业模式由研究合作驱动。Isomorphic Labs 通过合作向大型药企授权其 AI 药物设计能力，交易结构包括首付款、与临床和监管进展挂钩的里程碑付款，以及商业化药物的特许权使用费。截至 2026 年 5 月，公司与 Eli Lilly、Novartis、Johnson & Johnson 保持活跃合作。公司内部也在建设自有管线，聚焦肿瘤、免疫和炎症；截至 2026 年 5 月，这些项目均仍处于临床前阶段。公司的首个 AI 设计候选药物 ISM8969 于 2026 年 1 月 28 日获得 FDA IND 放行，使 Isomorphic Labs 成为首家将内部生成分子推进人体临床试验的 AI 原生公司。 Isomorphic Labs 将其总可用市场（TAM）定义为全球药物发现和开发行业的计算层，2025 年估计为 $2.58B，并以约 25-30% 的 CAGR 增长。公司所在环节是临床前 AI 设计阶段，通过在 DeepMind 十年 AlphaFold 研究基础上训练的自有结构生物学模型，建立竞争护城河。[CO001, CO002, CO003, CO004, CO005, CO006]
| 指标 | 数值 / 状态 | 日期 | 置信度 | 缺口 / 备注 |
|---|---|---|---|---|
| 成立 | 2021 年 2 月 24 日（UK Companies House） | 2021-02-24 | 高 | 监管文件确认日期 |
| 公开宣布 | 2021 年 11 月 4 日 | 2021-11-04 | 高 | CNBC、TechCrunch、The Verge 确认 |
| 法定名称 | Isomorphic Labs Limited | 当前 | 高 | UK Companies House #13223825 |
| 母公司 | Alphabet Inc.（Google 母公司） | 当前 | 高 | 官方来源确认 |
| 总部 | 英国伦敦 | 当前 | 高 | 官方网站和公司文件 |
| 其他办公室 | Lausanne（2022）、Cambridge MA（2025） | 当前 | 中 | 未单独官方披露 |
| 累计融资 | ~$2.7 billion | 2026-05-12 | 高 | 融资：Series A $600M + Series B $2.1B |
| 估值（投后） | 未披露 | Unknown | 低 | 两轮融资均无官方披露 |
| 员工数 | 371 名员工（2026 年 4 月 30 日） | 2026-04-30 | 中 | Tracxn 数据；非官方 |
| 活跃药企合作 | 3（Lilly、Novartis、J&J） | 2026-01 | 高 | 官方新闻稿确认每一项 |
| 首个 IND 候选物 | ISM8969 获 FDA 放行 | 2026-01-28 | 高 | 官方公告和新闻报道 |
| 临床试验时间线 | 2026 年底（从 2025 年推迟） | 2026-01-20 | 高 | WEF Davos 公告 |
| AlphaFold Server 用户 | 3+ million 名研究人员，190+ 个国家 | 当前 | 中 | 公司说法，无独立审计 |
| 收入 | 未披露 | Unknown | 低 | 私人公司；无公开财务数据 |
| 内部管线阶段 | 截至 2026 年 5 月全部处于临床前 | 2026-05 | 高 | 多个新闻来源确认 |
| Series B 规模 | 史上第二大生物技术融资 | 2026-05-12 | 高 | 仅次于 Altos Labs $3B |
KPI 快照基于截至 2026-05-16 的公开来源；估值和收入未披露。员工数来自 Tracxn 数据。AlphaFold server 使用量来自公司说法。
[CO001, CO018, CO019, CO020, CO021, CO022]截至 2026 年 5 月的 Isomorphic Labs KPI 摘要，突出累计融资、合作关系和管线状态。
收入和估值未披露。员工数来自 Tracxn 数据，并非官方口径。
[CO018, CO019, CO020, CO025, CO029, CO030]1.2 领导层与关键人分析
Demis Hassabis 爵士同时担任 Isomorphic Labs 首席执行官兼创始人，也担任 Google DeepMind 首席执行官兼创始人。这一双重角色既带来显著竞争优势——Isomorphic Labs 能优先获取 DeepMind 的基础研究，包括 AlphaFold——也带来重大关键人集中风险。Hassabis 的思想领导力和科学信誉，是公司合作可信度和投资逻辑的核心。如果 Hassabis 离开或减少投入，合作关系和研究路线图都可能承压。 公司成立以来，高管团队已经明显补厚。Max Jaderberg 原任首席 AI 官，后升任总裁，显示组织成熟度提升。首席科学官 Miles Congreve 带来深厚的药物化学经验。Sergei Yakneen 担任 CTO，Pamela Carroll 担任 COO，Sarah Korman 担任首席商务与法务官。关键变化是 Dr. Ben Wolf 于 2025 年中出任首席医疗官，反映公司正从纯计算研究转向临床转化。这一临床领导层任命，指向公司有意管理首个人体临床试验 ISM8969；该试验已从 2025 年底推迟至 2026 年底。 领导团队主要来自 DeepMind、顶级药企研发和计算化学。截至 2026 年 4 月 30 日，公司员工数为 371 人（据 Tracxn 数据），并持续招聘 AI 研究、药物设计和临床运营岗位。董事会治理结构未在公开文件中完整披露，这与公司的私营状态一致；不过，Alphabet 的监督角色已由 UK Companies House 记录确认。[CO009, CO010, CO011, CO012, CO013, CO014]
| 姓名 | 职务 | 背景 | 创始人-市场匹配 | 关键人物风险 |
|---|---|---|---|---|
| Sir Demis Hassabis | CEO 与创始人 | 共同创办 DeepMind（2010），AI 先驱，国际象棋神童，Cambridge 神经科学博士 | 最高：发明 AlphaFold，同时领导 DeepMind 和 Isomorphic Labs | 关键——双重 CEO 角色形成依赖和集中度风险 |
| Max Jaderberg | 总裁（前首席 AI 官） | 前 DeepMind 研究科学家；关键 AI 系统架构师 | 高：深厚 AI 系统经验，连接研究与商业化 | 高——运营领导锚点 |
| Miles Congreve | 首席科学官 | 前 Heptares Therapeutics 发现副总裁；资深结构药物设计专家 | 高：30+ 年药企职业经历，对合作伙伴可信度至关重要 | 高——支撑药企合作的科学可信度 |
| Sergei Yakneen | 首席技术官 | 计算生物学和 AI 基础设施工程背景 | 中：负责技术执行，但外部能见度较低 | 中 |
| Pamela Carroll | 首席运营官 | 生物技术运营和规模化经验 | 中：临床爬坡阶段需要 | 中 |
| Dr. Ben Wolf | 首席医疗官（2025 年中任命） | 生物制药临床开发经验 | 高：对 ISM8969 IND / Phase I 管理至关重要 | 高——负责临床执行 |
| Sarah Korman | 首席商务与法务官 | 生命科学领域商务拓展和法律 | 中：管理合作伙伴关系和 IP | 中 |
领导层信息来自官方 isomorphiclabs.com、Tracxn 和新闻来源。因公司为私人公司，董事会构成未公开披露。
[CO009, CO010, CO011, CO012, CO013, CO014]Alphabet、DeepMind、投资者和制药合作伙伴如何连接到 Isomorphic Labs。
[CO003, CO018, CO019, CO020, CO029, CO030]1.3 融资历史与资本结构
Isomorphic Labs 已通过两轮融资筹集约 $2.7B 外部资本。Series A 轮为 $600M，于 2025 年 4 月 1 日完成，由 Thrive Capital 领投，GV（Google Ventures）和 Alphabet 等共同投资。Series B 轮为 $2.1B，于 2026 年 5 月 12 日宣布，Thrive Capital 再次领投，共同投资方包括 Alphabet、GV、MGX（阿联酋主权科技基金）、Temasek（新加坡主权财富）、CapitalG（Alphabet 独立成长基金）和 UK Sovereign AI Fund。Thrive Capital CEO Joshua Kushner 主导了两轮融资。Alphabet 总裁兼 CIO Ruth Porat 公开背书 Series B 轮。以 $2.1B 规模计，Series B 轮是历史上第二大单笔生物科技融资，仅次于 Altos Labs 的 $3B 融资。 两轮估值均未披露。公司拒绝公开提供投后估值，因此无法计算早期股东的精确持股比例或隐含回报。这种不透明与 Alphabet 报告子公司财务的偏好一致，也是一个明显尽调缺口。分析师曾猜测估值区间在 $5-10B，但截至 2026 年 5 月没有官方确认。 资本结构显示 Alphabet 仍保持控制性持股，Thrive Capital 和 GV 是主要财务共同投资方。来自阿联酋（MGX）、新加坡（Temasek）和英国的主权财富资金参与，释放出对英国 AI 基础设施投资的强地缘政治支持信号。累计外部融资约 $2.7B，不包括 Series A 之前 Alphabet 提供的任何企业资金；其条款未披露。[CO018, CO019, CO020, CO021, CO022, CO023]
| 利益相关方 | 角色 | 控制权 / 经济重要性 | 尽调问题 |
|---|---|---|---|
| Alphabet Inc. | 母公司和战略投资者 | 大股东；提供算力、DeepMind IP 接入和公司治理 | 确认准确持股比例；判断 Alphabet 是否拥有优先购买权或董事会控制权 |
| Thrive Capital（Joshua Kushner） | Series A 和 B 领投方 | 重要少数股权，领投 $600M Series A 和 $2.1B Series B | 确认准确持股规模、董事会席位状态、清算优先权 |
| GV（Google Ventures） | Series A 和 B 共同投资方 | 与 Google / Alphabet 生态一致的战略共同投资者 | 确认持股规模及任何共同投资协议条款 |
| CapitalG（Alphabet Growth Fund） | Series B 共同投资方 | Alphabet 关联成长基金，与母公司战略一致 | 确认独立决策与母公司协调之间的边界 |
| MGX（UAE 主权技术基金） | Series B 共同投资方 | 来自 Abu Dhabi 的主权技术资本；具备地缘政治协同 | 确认投资条款及任何战略 IP 条件 |
| Temasek（Singapore SWF） | Series B 共同投资方 | Singapore 主权财富资本；跨境生命科学兴趣 | 确认持股及任何优先收益权 |
| UK Sovereign AI Fund | Series B 共同投资方 | UK 政府协同资本；关注国家 AI 基础设施 | 了解任何 IP 限制或 UK 总部承诺 |
| Eli Lilly | 合作伙伴 | 最高 $1.745B 的里程碑经济性；战略验证者 | 确认里程碑付款安排和 IP 所有权条款 |
| Novartis | 合作伙伴（已扩展） | 原始和扩展项目合计最高 $1.2B 里程碑 | 确认准确项目数量和排他性条款 |
| Johnson & Johnson（Janssen） | 合作伙伴（2026） | 跨模态药物设计；J&J 负责实验验证 | 确认财务条款和模态范围 |
所有投资者持股比例均未披露。经济重要性根据交易和融资背景估计。地缘政治股东（MGX、Temasek、UK SWF）代表主权投资者基础。
[CO018, CO019, CO020, CO021, CO022, CO023]从创立到 2026 年 5 月的融资轮次和关键投资者关系，并纳入主权财富与战略资本背景。
日期均按官方新闻稿或新闻报道中的披露原样采用。
[CO003, CO018, CO019, CO020, CO021, CO025]1.4 关键里程碑与战略进展
Isomorphic Labs 已从创立时的 AI 研究公司，推进为多阶段、面向临床的药物设计引擎。公司以 AlphaFold 2 在 2020 年 CASP 竞赛中的胜利为科学基础；该成果解决了长达 50 年的蛋白质结构预测问题。AlphaFold 3 于 2024 年 5 月 8 日发表于 Nature（doi:10.1038/s41586-024-07487-w），将预测范围扩展到 DNA、RNA、小分子和蛋白相互作用，并在 PoseBusters 基准上比传统方法准确 50%。免费开放的 AlphaFold 服务器已服务 190+ 个国家的 3+ million 研究人员。 与 Eli Lilly（首付款 $45M，里程碑最高 $1.7B）和 Novartis（首付款 $37.5M，里程碑最高 $1.2B）的两项标志性药企合作于 2024 年 1 月宣布。Novartis 于 2025 年 2 月将合作扩展至 3 个新增项目。与 Johnson & Johnson 的第三项合作于 2026 年 1 月宣布。IsoDDE（Isomorphic Drug Design Engine）平台于 2026 年 2 月 10 日宣布，是一次重大技术升级——公司称其在蛋白-配体泛化基准上的准确性比 AlphaFold 3 提升逾一倍，并在结合亲和力预测上超越基于物理的方法。 ISM8969 是公司内部产出的首个 AI 设计候选药物，于 2026 年 1 月 28 日获得 FDA IND 放行。临床试验原定 2025 年底启动，但在 2026 年 1 月达沃斯世界经济论坛上宣布推迟 12 个月至 2026 年底。这一延迟削弱了公司关于 AI 能压缩药物发现周期的叙事，是一个实质性挫折。[CO026, CO027, CO028, CO029, CO030, CO031]
| 日期 | 事件 | 类型 | 金额 / 估值 / 状态 | 参与方 | 影响 |
|---|---|---|---|---|---|
| 2020-12 | AlphaFold 2 赢得 CASP14 竞赛，解决蛋白质折叠问题 | 产品 | 标志性科学成就 | DeepMind（当时尚非 Isomorphic） | 基础 IP 资产；AlphaFold 获得全球认可 |
| 2021-02-24 | Isomorphic Labs Limited 在 UK 注册成立 | 创立 | 公司 #13223825 已注册 | Demis Hassabis、Alphabet | 官方法律成立；比公开宣布早 8 个月 |
| 2021-11-04 | Isomorphic Labs 公开宣布 | 创立 | Alphabet 子公司 | Demis Hassabis 出任 CEO | 公开启动；DeepMind CEO 同时领导两家公司 |
| 2022-12 | Lausanne, Switzerland 办公室开设 | 扩张 | 欧洲扩张 | Isomorphic Labs | 靠近 EU 市场的监管和药企资源 |
| 2024-01-08 | Eli Lilly 合作宣布 | 合作 | $45M 首付款 + 最高 $1.7B 里程碑 | Isomorphic Labs、Eli Lilly | 首个重大药企验证；潜在交易额 $1.745B |
| 2024-01-08 | Novartis 合作宣布 | 合作 | $37.5M 首付款 + 最高 $1.2B 里程碑 | Isomorphic Labs、Novartis | 第二项药企验证；合计里程碑潜力约 $3B |
| 2024-05-08 | AlphaFold 3 发表于 Nature | 产品 | 论文 DOI：doi:10.1038/s41586-024-07487-w | Isomorphic Labs、Google DeepMind | 重大技术里程碑；PoseBusters 基准准确率高出 50% |
| 2025-02 | Novartis 合作扩展 | 合作 | 同等条款下新增 3 个项目 | Isomorphic Labs、Novartis | 与欧洲最大药企的战略关系加深 |
| 2025-04-01 | 完成 Series A $600M 融资 | 融资 | Thrive Capital 领投 $600M | Thrive Capital、GV、Alphabet | 首笔外部机构资本；验证临床前 AI 模型 |
| 2025-mid | Cambridge MA 办公室开设 | 扩张 | 美国扩张 | Isomorphic Labs | 靠近美国生物制药生态和临床人才 |
| 2025-mid | Dr. Ben Wolf 被任命为 CMO | 治理 | 临床领导层招聘 | Isomorphic Labs | 表明公司开始转向 ISM8969 的临床执行 |
| 2026-01-08 | Johnson & Johnson 合作宣布 | 合作 | 财务条款未披露 | 合作方：Isomorphic Labs、J&J（Janssen Biotech） | 第三家大型药企合作伙伴；跨模态扩展 |
| 2026-01-20 | WEF Davos 宣布临床试验时间线延迟 | 反向 | 从 2025 年底延至 2026 年底 | Isomorphic Labs | 12 个月后移；首次公开反向披露 |
| 2026-01-28 | ISM8969 获 FDA 放行开展人体试验（IND） | 监管 | IND 放行获批 | Isomorphic Labs、FDA | 首个 AI 设计药物达到人体试验门槛 |
| 2026-02-10 | IsoDDE（Drug Design Engine）发布 | 产品 | 基准测试准确率比 AF3 高 >2x | Isomorphic Labs | 平台升级；科学家非正式称其为 AlphaFold 4 |
| 2026-05-12 | Series B 融资 $2.1B | 融资 | $2.1B，由 Thrive Capital 领投 | 投资方：Thrive、Alphabet、GV、MGX、Temasek、CapitalG、UK Sovereign AI Fund | 生物技术史上第二大轮融资；累计资本约 $2.7B |
日期来自官方新闻稿、UK Companies House、新闻报道和 Nature 发表记录。试验延期和 IND 批准日期来自多方交叉印证来源。
[CO001, CO002, CO018, CO019, CO020, CO026]1.5 图表与要点
02市场分析
2.1 市场规模、边界与增长轨迹
Mordor Intelligence 估计，全球 AI 药物发现市场在 2025 年约为 $2.58B，并将在 2030 年前以约 25-30% 的复合年增长率增长。这个市场不同于更广义的药企研发支出（2025 年全球约 $230B），只代表药物发现价值链中的 AI 专用计算层。边界很重要：Isomorphic Labs 不参与湿实验室合成、临床开发或制造；它竞争的是 in silico 靶点识别、结构预测和先导化合物优化，这些环节只占临床前研发投资的一小部分，但增长很快。 Axis Intelligence 的 2026 年分析认为，AI 药物发现市场规模会随方法不同而变化——自下而上的方法若纳入所有带里程碑的药企 AI 授权交易，市场到 2027 年可能显著更大，潜在达到 $5-8B。Mordor Intelligence 的 2025 年 $2.58B 基准情景，可能代表直接 AI 平台合同收入，并排除了大型药企研发组织内部的内嵌 AI 支出。Excelra 2026 年 AI 状态报告指出，Top-20 药企中已有 78% 拥有活跃 AI 药物发现项目，说明市场渗透正快于此前预测。 2025 年，投向 AI 药物发现的风险投资达到 $4.1B，同比增长 38%，显示金融市场需求已明显领先于当前商业市场规模。这种错配——大额 VC 押注与不到 $3B 的现有市场并存——反映出行业预期 AI 会快速替代传统临床前方法。Isomorphic Labs 累计融资 $2.7B，使其按融资规模成为最大的私营 AI 药物发现公司，领先于 Recursion（累计 $900M+）和 Schrödinger（$500M+）。[CM001, CM002, CM003, CM004, CM005, CM006]
| 细分市场 / 类别 | 纳入支出 | 剔除支出 | 买方 / 付款方 | 与 Isomorphic Labs 的关联 |
|---|---|---|---|---|
| AI 药物靶点识别 | 借助基因组学、蛋白组学数据识别分子靶点的 AI 工具 | 传统基因组测序成本、基础湿实验检测 | 大型药企研发团队、生物技术研究人员 | 位于 Isomorphic 主要焦点上游，但 AlphaFold 3 可用 |
| AI 结构预测 | 计算结构预测平台、AlphaFold 等模型授权费 | 物理 X 射线晶体学、冷冻电镜设备成本 | 药企研发、结构生物学实验室、学术研究人员 | Isomorphic 凭 AlphaFold 2/3 打出的核心能力；免费服务器服务 3M+ 研究人员 |
| AI 先导化合物优化（小分子） | 面向 hit-to-lead 与先导化合物优化的 AI 药物设计平台、里程碑付款 | 合成成本、检测成本、湿实验验证 | 药企 BD 团队、药物化学部门 | 核心商业细分市场；Lilly 与 Novartis 合作落在这里 |
| AI 多模态药物设计 | 覆盖小分子、生物制剂、核酸、多特异性分子的 AI 平台 | 抗体制造成本、临床试验成本 | 管线多元的大型药企（J&J、AZ、Roche） | 通过 2026 年 J&J 合作切入 |
| AI ADMET / 安全性预测 | 预测吸收、分布、代谢、排泄、毒性的 AI 工具 | 体内毒理研究、监管安全性测试 | 所有药物开发者；外包 CRO | 相关但边缘；不是 Isomorphic 已披露的主要焦点 |
| 学术 / 研究 AI（免费层） | 免费 AlphaFold 服务器带来间接价值——无直接收入 | N/A（无买方；免费服务） | 190+ 个国家的 3M+ 研究人员 | 搭生态、蓄人才、建立科学可信度；无直接变现 |
市场边界按 AI 特定计算药物发现支出口径划定，剔除下游湿实验合成、临床试验和制造成本。学术免费层作为战略市场信号纳入，不作为收入类别。
[CM001, CM014, CM016, CM018]| 发布方 | 年份 | 地域 | 市场规模 | CAGR | 方法 | 置信度 | 局限 |
|---|---|---|---|---|---|---|---|
| Mordor Intelligence | 2025 | 全球 | $2.58B（2025E） | ~25-30% | 自下而上：AI 平台交易 + 授权 | 中 | 不含药企 IT 中嵌入式 AI 支出 |
| Axis Intelligence | 2026 | 全球 | $3-5B（2025E，口径更宽） | ~30-35% | 纳入交易（以里程碑付款作代理） | 低 | 纳入里程碑的口径会抬高基数，相比 Mordor 更大 |
| Excelra | 2026 | 全球 | 采用率指标：Top-20 药企中 78% 使用 AI | N/A | 药企 AI 项目采用率调研 | 中 | 采用率不直接等于市场美元支出 |
| 风险投资代理指标 | 2025 | 全球 | $4.1B VC 投资（2025） | +38% YoY | VC 交易数据汇总 | 高 | VC 投资可能滞后或领先商业市场；不是市场收入 |
| Isomorphic Labs（隐含） | 2026 | 全球 | $45M+$37.5M 预付款 = 2 笔交易合计 $82.5M | N/A | 已披露直接交易收入 | 高 | 仅含预付款；若全部兑现，里程碑潜在价值 ~$3B |
| GlobalData（估计） | 2027E | 全球 | 到 2027 年 ~$4.9B | ~27% CAGR | 自上而下分配药企研发 AI 支出 | 低 | 预测基于药企 AI 预算调研；单一来源 |
规模测算因方法不同而差异较大。Mordor 与 Axis 属于分析师估计；Excelra 衡量采用率而非收入；VC 代理指标衡量投资者需求而非商业市场。按口径不同，2025 年实际市场规模可能为 $2-5B。
[CM001, CM002, CM003, CM005]多家分析机构对全球 AI 药物发现市场规模的估计，以十亿美元口径展示估计区间。
2027E 和 2030E 数值根据已披露 CAGR 区间外推。风险投资不等同于市场收入。所有数值均以十亿美元计。Axis Intelligence 口径包含交易，不能与 Mordor 基准情景直接比较。
[CM001, CM002, CM005]基于多种分析师方法，给出 2025 年全球 AI 药物发现总可用市场（TAM）的低、基准和高估计。
所有数值均以十亿美元计。TAM 估计随口径定义差异很大。VC 投资不等同于市场收入。Isomorphic 可触达细分为作者估计。
[CM001, CM002, CM003, CM006]2.2 需求驱动因素与采用约束
推动 AI 药物发现平台需求的核心经济逻辑，是传统药物开发的灾难性成本结构。行业平均每个获批药物成本约 $2.0-2.6B，周期 10-15 年，且约 90% 的候选药物会在临床开发某个阶段失败。仅临床前阶段——也就是 Isomorphic Labs 所在环节——通常就需要 4-6 年，用于靶点识别、结构确定、先导化合物优化和 IND 准备。AI 优先的发现平台声称可将其压缩至 12-18 个月，潜在加速 3-4 倍。 顶级药企的需求信号很强。截至 2025 年，Top-20 药企全部在药物发现中使用某种形式的 AI，其中 78% 拥有专门的 AI 平台合作或内部团队。Isomorphic Labs 与 Eli Lilly 和 Novartis——两家全球最大药企之一——的合作，验证了一线药企愿意为 AI 赋能的计算药物设计支付可观首付款（分别为 $45M 和 $37.5M）。 关键采用约束包括：（1）AI 发现药物的监管不确定性，截至 2026 年 5 月尚无 AI 设计药物获批；（2）深度学习模型的“黑箱”属性，会在监管申报中引发可复现性担忧；（3）与现有药企信息系统和湿实验室基础设施的集成成本；（4）历史记录有限——ISM8969 是 Isomorphic Labs 首个进入临床试验的 AI 设计药物，意味着最早也要到 2027 年底才会有临床数据。 Isomorphic Labs 自有临床试验时间表推迟 12 个月（从 2025 年底推至 2026 年底），说明即使资本最充足的 AI 药物发现公司也面临执行风险。如果结果不及预期，药企采用速度可能放缓。[CM007, CM008, CM009, CM010, CM011, CM012]
| 驱动 / 约束 | 方向 | 时间 | 含义 | 尽调问题 |
|---|---|---|---|---|
| 药企研发成本压力（每款药 $2B，周期 10-15 年） | 强驱动 | 立即 | Top-20 药企愿意为 AI 提速支付可观预付款；验证 Isomorphic 模式 | 确认里程碑结构是在激励 AI 提速，而非简单外包工作 |
| 传统药物开发 90% 临床失败率 | 强驱动 | 立即 | AI 声称可提升化合物质量、降低淘汰率，这是核心价值主张 | 跟踪 AI 设计药物的临床成功率，并与行业历史均值比较 |
| Top-20 药企中 78% 已使用 AI（2025） | 强驱动 | 立即 | 头部采用率已接近饱和；市场转向质量 / 结果竞争 | 核实采用率数据来自一手调研，而非供应商自报 |
| 截至 2026 年 5 月仍无获批 AI 设计药物 | 约束 | 12-36 个月 | 缺少临床证明，限制合作交易规模，也拖慢早期采用者之外的采用节奏 | 跟踪 ISM8969 Phase I 结果，这会成为行业定调催化剂 |
| 监管不明确（FDA 尚无 AI 药物专门指南） | 约束 | 2-5 年 | 监管路径不确定，可能拖慢 IND 时间线；若先行者参与定义框架，也可能受益 | 跟踪 FDA AI/ML 指南发布和 EMA 立场 |
| 从现有平台（Schrödinger、CDD）切换的成本 | 中等约束 | 持续 | 药企计算化学团队深嵌在基于物理的工作流中；AI 颠覆需要组织变革 | 评估 Lilly / Novartis 管线整合深度是否超出初始项目 |
| 学术生态采用（3M+ AlphaFold 用户） | 强驱动 | 持续 | 受训研究人员拉动需求；Isomorphic 受益于免费生态的品牌价值 | 跟踪学术使用是否转化为商业合作或招聘管线 |
| Series B 中的主权财富资本（MGX、Temasek、UK SWF） | 中性 / 驱动 | 2026+ | 国家 AI 基础设施投资可能提供有利条件，但也带来地缘政治义务 | 了解 UK SWF 参与是否附带 IP 本地化或数据共享要求 |
驱动和约束评估基于截至 2026 年 5 月的公开数据、分析师报告和公司披露。时间反映估计的市场成熟周期。
[CM007, CM008, CM009, CM010, CM011, CM012]对比传统药物发现时间线各阶段与 AI 加速路径，标出 Isomorphic Labs 这类 AI 平台创造价值的位置。
漏斗百分比代表候选药物在传统药物开发各阶段的大致存活率。AI 加速估计来自公司披露和分析师报告。
[CM007, CM008, CM009, CM013]2.3 市场细分与买方分析
AI 药物发现市场可沿三条主轴切分：（1）目标药物开发阶段（靶点识别、命中物识别、先导化合物优化、ADMET 预测、临床设计）；（2）治疗模态（小分子、生物药、核酸、细胞疗法、多特异性抗体）；（3）买方类型（大型药企、中型药企、生物科技公司、学术研究）。 Isomorphic Labs 主要服务小分子药物设计中的先导化合物优化和命中物识别阶段。其与 Eli Lilly 和 Novartis 的合作明确聚焦小分子。Johnson & Johnson 合作（2026 年 1 月）扩展到跨模态药物设计，包括生物药和其他模态。这一扩展具有战略意义：近年来，生物药在药企研发中所占份额高于小分子；掌握多模态设计，会显著扩大 Isomorphic 的可服务市场（SAM）。 AI 药物发现服务的买方格局高度集中在头部。全球 Top-20 药企估计占全球药企研发支出的 65-70%，是主要买方群体。具体到 Isomorphic Labs，其合作模式要求伙伴具备足够大的管线规模，能够吸收多靶点 AI 设计项目，并有财务能力为单个合作支付 $1-2B 里程碑款。现实上，这将短期买方池限制在 15-25 家全球药企和大型生物科技公司。学术研究通过免费 AlphaFold 服务器服务，能带来生态价值，但没有直接收入。[CM014, CM015, CM016, CM017, CM018, CM019]
| 细分市场 | 买方 | 用户 | 付款方 | 工作流 | 预算负责人 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 小分子先导化合物优化 | 大型药企 BD / 药物化学 | 计算化学家、药物化学家 | 药企研发预算 | 识别命中化合物、优化先导化合物、运行结合模拟 | 首席科学官 / 药物发现负责人 | 管线瓶颈、IND 速度压力 |
| 多靶点、多模态发现 | 大型药企（J&J、AZ、Roche 等） | 药物发现生物学家、结构生物学家 | 研发 / 创新预算 | 同时跨模态设计生物制剂 + 小分子 | 早期管线负责人 / CTO | 难成药靶点的生物复杂性 |
| 靶点识别（基因组学 + AI） | 生物技术初创公司、癌症药企 | 数据科学家、基因组学专家 | VC 支持的研发预算 | 用 AI 根据组学数据优先筛选新生物靶点 | 早期生物技术公司 CEO / CSO | 罕见病、肿瘤领域未满足需求 |
| 学术 / 结构生物学研究 | 大学、研究机构 | 研究生、博士后、结构生物学家 | 科研经费支持（NIH、Wellcome Trust、ERC） | 结构预测、蛋白质-蛋白质相互作用图谱 | PI / 系主任 | 需要免费的前沿 AI 工具 |
| ADMET 与安全性预测（CRO） | 合同研究组织（CRO） | 安全性科学家、监管专家 | 药企外包预算 | 在昂贵体内试验前预测化合物安全性画像 | 监管副总裁 / 安全性负责人 | CRO 差异化；相较体内研究降本 |
买方分层基于 Lilly、Novartis、J&J 等公开交易结构、AlphaFold 服务器使用数据和行业分析师报告。学术细分市场对 Isomorphic Labs 没有变现路径。
[CM014, CM015, CM016, CM017, CM018]| 公司 | 模式 | 阶段 | 差异化 | 对 Isomorphic 的威胁等级 |
|---|---|---|---|---|
| Recursion + Exscientia（2025 年合并） | 表型组学 + AI 分子设计；上市公司（RXRX） | 临床阶段（10+ 个项目） | 湿实验表型组学 + AI；合并后成为最大上市 AI 药物公司 | 高：临床管线最广；最可比的上市公司 |
| Schrödinger（SDGR） | 基于物理 + ML 混合；上市公司 | 平台 + 管线 | FEP+ 是结合亲和力金标准；现金 $500M+ | 中高：基于物理设计的老牌玩家；AF3 有挑战，但无法完全替代 FEP |
| Insilico Medicine | 用于靶点识别、生成式分子设计的 AI | 临床阶段（1 款药物在中国获批） | 首个 AI 设计药物获 IND 批准（ISM001-055，2021） | 中：临床里程碑领先 Isomorphic，但公司规模更小 |
| BenevolentAI | 知识图谱 + ML 做靶点识别 | 平台（未披露管线） | 用 AI 挖掘科学文献，寻找新靶点 | 中低：聚焦不同阶段；合作失败后仍在挣扎 |
| Xaira Therapeutics | 生成式生物学；2024 年以 $1B 启动 | 临床前 | 2024 年最大生物技术新公司；用 AI 生成式设计生物制剂 | 中：资金充足；生物制剂优先路径与 Isomorphic 的 J&J 扩张竞争 |
| Nabla Bio | 蛋白质-蛋白质相互作用 + AI 设计 | 早期 | 用 AI 聚焦难成药靶点 | 低：早期、聚焦小众 |
| Relay Therapeutics | 构象动力学 + AI | 临床阶段 | 精准肿瘤学；动态结合 AI 模型 | 中低：与肿瘤学有重叠；核心技术不同 |
| Atomwise | 用卷积神经网络做筛选 | 平台（合作） | ADMET 预测 + 虚拟筛选；750+ 个项目 | 低：较早期 CNN 路线；复杂度不如基于 AlphaFold 的设计 |
竞争格局基于截至 2026 年 5 月的公开披露、公司网站和分析师报道。Recursion-Exscientia 合并于 2025 年末完成。威胁等级是对 Isomorphic Labs 竞争位置的定性评估。
[CM021, CM022, CM024, CM025]将制药买方细分与 AI 药物发现能力对应起来，展示 Isomorphic Labs 覆盖各类买方需求的位置。
能力与细分市场的匹配是定性评估，依据合作交易结构、AlphaFold 服务器使用报告和分析师笔记。
[CM014, CM015, CM016, CM017, CM018, CM019]2.4 监管环境与竞争强度
截至 2026 年，AI 药物发现的监管环境仍在变化。FDA 尚未发布针对 AI 生成 IND 申报的具体指引，不过其已于 2026 年 1 月 28 日放行 ISM8969，标志着首个 AI 设计候选药物获得 IND 放行。FDA 的 Computer-Assisted Drug Design 指引框架提供了一定结构，但 AI 生成分子的具体验证要求仍未定义。这种监管模糊性既带来风险（验证路径不清），也构成进入壁垒（只有最成熟、资本最充足的公司才能穿越不确定性）。 AI 药物发现赛道的竞争强度中高，并且还在上升。Recursion-Exscientia 合并（2025 年底完成）催生了最大的纯 AI 药物发现上市公司，合并市值超过 $4B。Schrödinger（NASDAQ: SDGR）在基于物理的计算药物设计中占据主导地位，持有 $500M+ 现金。BenevolentAI、Insilico Medicine，以及 Xaira Therapeutics（2024 年成立，初始资本 $1B）和 Nabla Bio 等新进入者都在该赛道活跃。重要的是，2024 年 11 月 CASP16 基准结果显示，在竞赛环境下，基于 AlphaFold 3 的模型在蛋白-配体相互作用上并未显著优于旧方法，说明在结合亲和力预测上，Isomorphic 的结构生物学领先可能不如纯蛋白折叠领域那样决定性。 ISM8969 获得 IND 放行，说明市场监管正在成熟；AI 药物发现正从基于承诺的竞争市场，转向基于证据的竞争市场，临床结果会越来越决定哪些平台能吸引药企合作。这一转向有利于 Isomorphic、Recursion 等拥有临床阶段资产且资本充足的玩家，而不是纯计算阶段竞争者。[CM020, CM021, CM022, CM023, CM024, CM025]
2.5 图表与要点
03竞争格局
3.1 竞争版图与品类分层
Isomorphic Labs 所在市场拥挤但分层清晰。直接竞争对手分为四类：（1）拥有自有管线并追求药企合作收入的全栈 AI 药物发现公司（Recursion/Exscientia、Insilico Medicine、BenevolentAI）， （2）授权计算工具、同时运营自有药物项目的平台软件公司（Schrödinger），（3）聚焦生物药或抗体的模态特定生成式 AI 公司（AbSci、Xaira Therapeutics），以及（4）瞄准不可成药蛋白的结构 / 构象生物学平台（Relay Therapeutics、Atomwise）。 Isomorphic 位于第 1 类和第 4 类的交叉处：它以 AlphaFold3 衍生结构预测作为多模态药物设计引擎的核心，通过大型药企合作（Lilly、Novartis、J&J）而非软件授权取得收入。其差异化特征包括：（a）从 Alphabet/Google DeepMind 独家授权 AlphaFold3 用于商业药物发现，（b）约 $2.7B 资本基础且无强制性公开季度报告义务，（c）明确的“纯 AI”设计理念，但仍未在人类试验中验证（ISM8969 Phase I 目标为 2026 年底）。 主要竞争威胁是 Recursion Pharmaceuticals，其于 2025 年 1 月完成对 Exscientia 的收购。Recursion 是上市公司（NASDAQ: RXRX），拥有临床阶段项目，包括其 AI Operating System 于 2026 年初在 FAP（familial adenomatous polyposis，家族性腺瘤性息肉病）中的首个临床验证。Schrödinger 仍是基于物理的平台提供商，在药企中渗透广泛。Insilico Medicine 在非上市 AI 药物发现公司中拥有最先进的自有管线，其 INX-020 化合物治疗肺纤维化已进入 Phase II。 [CP001, CP002, CP003, CP004, CP005, CP006]
| 公司 | 类别 | 规模 / 融资 | 目标细分市场 | 差异化 | 管线阶段 | 商业模式 |
|---|---|---|---|---|---|---|
| Recursion Pharmaceuticals | 全栈 AI（上市） | NASDAQ: RXRX；2025 年 1 月以约 $688M 收购 Exscientia | 罕见病、肿瘤、炎症 | 规模化表型组学；机器人生物学；临床项目 | Phase II（FAP、罕见病） | 里程碑付款 + 软件授权 |
| Schrödinger | 基于物理的平台（上市） | NASDAQ: SDGR；2024E 软件 ARR ~$150M | 所有药企、材料、农业领域 | FEP+ 基于物理做先导化合物优化；药企渗透率 90%+ | 合作型早期临床项目 | 软件 ARR + 里程碑付款 |
| Insilico Medicine | 全栈 AI（私营） | 私营；累计融资 ~$400M+ | 纤维化、肿瘤、衰老 | Pharma.AI（Biology42+Chemistry42+Medicine42）；40+ 个项目 | Phase II（INX-020/ISN21 肺纤维化） | 里程碑付款 + 合作 |
| BenevolentAI | 全栈 AI（退市） | 曾在 Euronext 上市；2025 年退市；已融资 ~$300M | 罕见病、炎症、CNS | 知识图谱 + 靶点识别；baricitinib 重新定位（COVID-19） | 临床前；战略重整仍在推进 | 里程碑付款 + 合作（重组中） |
| Xaira Therapeutics | 生成式 AI（私营，尚未产生收入） | 私营；2024 年 4 月 $1B 种子 / Series A | 广泛；小分子 + 生物制剂 | 从头生成式 AI；独立平台（不依赖 AF3） | 临床前（无 IND） | 合作（TBD） |
| Absci | 生物制剂 AI（上市） | NASDAQ: ABSI；聚焦生物制剂 / 抗体 | 生物制剂；皮肤科、肿瘤 | 生成式 AI + 湿实验 6 周周期；抗体设计 | IND-ready（ABS-201 雄激素性脱发） | 合作 + 内部管线 |
| Relay Therapeutics | 构象动力学（上市） | NASDAQ: RLAY；聚焦肿瘤 | 精准肿瘤、遗传病 | Dynamo® 蛋白运动建模；难成药靶点 | Phase I/II（精准肿瘤） | 自有管线 |
| Atomwise | 小分子 AI（私营） | 私营；累计融资 ~$174M | 小分子命中发现 | AtomNet 深度学习；3T+ 可合成化合物库 | IND 候选（TYK2 抑制剂） | 授权 + 里程碑 |
截至 2026 年 5 月的竞争快照，基于公开披露。管线阶段和融资数字可能在 3-6 个月内过时。
[CP001, CP002, CP003, CP004, CP005, CP006]| 公司 | 交易伙伴 | 首付款 | 最高里程碑 | 交易日期 | 状态 |
|---|---|---|---|---|---|
| Isomorphic Labs | Eli Lilly | $45M | 最高 $1.7B | January 2024 | 活跃 |
| Isomorphic Labs | Novartis | $37.5M（Feb 2025 扩展） | 最高 $1.2B | January 2024 | 活跃（已扩展） |
| Isomorphic Labs | Johnson & Johnson | 未披露 | 未披露 | January 2026 | 活跃 |
| Recursion | Roche/Genentech | $150M | 最高 $850M | 2021 | 持续中 |
| Recursion | Bayer | $20M | 最高 $100M | 2021 | 持续中 |
| Insilico Medicine | Sanofi | 未披露 | ~$1.2B（估计） | 2022 | 持续中 |
| BenevolentAI | AstraZeneca | 未披露 | ~$100M（估计） | 2019 | 2023 年结束 |
| Atomwise | Sanofi | 未披露 | 未披露（估计 $50-200M） | 2023 | 活跃 |
多数交易财务条款保密。表中数值要么来自官方披露，要么是分析师估计；未经一手来源确认，不应依赖。
[CP015, CP016, CP034]按结构生物学深度（x 轴）和临床管线成熟度（y 轴）映射主要 AI 药物发现竞争者。在活跃玩家中，Isomorphic 结构深度得分最高，但临床成熟度最低。
轴上数值是基于公开信息的定性评分（0-100），不是定量测量。结构深度反映结构生物学工具和自研模型的使用；临床成熟度反映自有管线项目数量和阶段。
[CP001, CP003, CP008, CP011]3.2 功能与能力对比
在整个竞争版图中，AI 药物发现公司在科学深度、管线成熟度、收入模式和模态覆盖上差异很大。Isomorphic 的核心差异化来自结构生物学预测（AlphaFold3 + IsoDDE）、多模态设计（小分子和生物药），以及基于三项药企合作内部数据集训练的自有药物设计模型。 不过，它缺少 Recursion 那样的临床阶段管线记录，也缺少 Schrödinger 在先导化合物优化中基于物理的 FEP+ 严谨性。 能力矩阵显示，没有单一竞争对手能在所有维度匹配 Isomorphic。Recursion 在表型筛选规模上领先（PB 级细胞成像数据），但结构生物学深度不足。Schrödinger 在基于物理的准确性和软件渗透上领先，但并非纯 AI 原生。Insilico 作为私营公司在管线推进上领先，其 Pharma.AI 平台已有 13 个 IND 获批。BenevolentAI 虽然早期拥有知识图谱优势，但已经历战略重组（2024 年 12 月），并提议从 Euronext Amsterdam 退市（2025 年 2 月），代表一次竞争挫折。Xaira Therapeutics 虽在 2024 年 4 月以 $1B 资本成立，但仍处于临床前阶段。 Isomorphic 评分最高的功能是结构预测和药企合作伙伴质量。相对竞争对手，其最弱环节是表型筛选、临床记录和软件授权广度。所有主要玩家都依赖药企合作里程碑收入，没有哪家公司拥有足以独立支撑运营的独立产品收入模式。 [CP011, CP012, CP017, CP018, CP021, CP022]
| 能力 | Isomorphic Labs | Recursion | Schrödinger | Insilico Medicine | Xaira | BenevolentAI |
|---|---|---|---|---|---|---|
| 结构预测 | 强（AF3 + IsoDDE） | 弱 | 强（FEP+） | 中（Chemistry42） | 中（生成式） | 弱（聚焦 KG） |
| 临床管线 | 弱（无 Phase I 数据） | 强（Phase II FAP） | 中（合作型） | 强（Phase II） | 缺失（无 IND） | 弱（重组中） |
| 药企合作 | 强（Lilly、Novartis、J&J） | 强（Roche、Sanofi、Bayer） | 强（90%+ 前 20 大） | 中（Sanofi 等） | 缺失（未披露） | 弱（AZ 已结束） |
| 靶点识别 AI | 中（偏结构） | 强（表型组学 + 基因组学） | 中（基于结构） | 强（Biology42） | 中（公司声称） | 强（知识图谱） |
| 多模态覆盖 | 强（小分子 + 生物制剂） | 中（偏小分子） | 强（工具） | 中（小分子） | 强（既定目标） | 弱（小分子） |
| 自研数据规模 | 中（合作数据） | 强（PB 级表型组学） | 强（药企级部署） | 中（内部数据） | 弱（建设中） | 中（KG） |
强 / 中 / 弱 / 缺失为截至 May 2026 基于公开信息作出的定性判断；未经独立验证。
[CP011, CP017, CP018, CP024]能力热力图：6 家竞争者的 6 项关键平台能力。Isomorphic 在结构预测上领先；Recursion 在临床记录上领先；Schrödinger 在制药公司渗透上领先。
强 / 中 / 弱 / 缺失是基于截至 2026 年 5 月公开信息的定性评级。缺失表示公开披露中没有该能力证据。
[CP022, CP026, CP027, CP028]3.3 护城河耐久性与竞争风险
Isomorphic 的护城河有三部分：（1）AlphaFold3 在药物发现商业应用中的独家权利——这是竞争对手无法复制的壁垒；（2）来自三项药企合作（Lilly、Novartis、J&J）的自有训练数据；（3）Google DeepMind 背景带来的人才集中。 这些护城河的耐久性面临三项重大威胁。第一，如果 Google DeepMind 扩大 AlphaFold 授权，AF3 独占性可能被削弱，就像 AF2 在 2021 年开源发布一样。第二，Xaira Therapeutics 于 2024 年 4 月以 $1B 种子 / Series A 资本成立，正在构建独立训练的生成式 AI 平台，用于抗体和小分子设计，理念同样接近“从零开始”，且不依赖 AF3。第三，Recursion 与 Exscientia 合并后的实体，提供了比任何合并前单一公司都更完整的一体化发现栈，覆盖表型组学、计算化学和分子设计。 Isomorphic 的合作交易条款——Lilly（首付款 $45M + 最高 $1.7B 里程碑）和 Novartis（$37.5M + 最高 $1.2B）——在人类试验验证之前披露的纯 AI 药物发现交易中位居前列，确认其相对同业具有强议价能力。J&J 合作（2026 年 1 月）未披露财务条款。Schrödinger 的软件 ARR（2024 年约 $130-150M）代表一种完全不同的商业化模式，建立在行业范围工具渗透之上。 最关键的护城河风险是 ISM8969 Phase I：一旦失败，Isomorphic 的信誉受损程度会明显高于已有临床验证数据的 Recursion，以及拥有 Phase II 项目的 Insilico。 [CP013, CP014, CP015, CP016, CP019, CP020]
| 护城河主张 | 威胁 | 严重程度 | 缓释措施 / 尽调问题 |
|---|---|---|---|
| AlphaFold3 独家商业权利 | Google DeepMind 扩大 AF3 授权范围（类似 AF2 在 2021 年开源）；AF4 以不同条款发布 | 高 | 要求披露独家期、续约条款和范围（使用领域）；向 Alphabet IR 核实 |
| 来自药企合作的专有训练数据 | 药企伙伴在合作结束后保留权利，可将数据用于自身或竞品 AI 模型 | 中 | 审阅数据所有权和独家条款；评估合作结束后数据是否回归药企伙伴 |
| IsoDDE 模型优越性（声称准确率 >2x AF3） | 竞品模型缩小差距；没有独立基准验证；主张仅由公司发布 | 中 | 要求独立基准结果或学术同行评议；用 CASP16、PoseBusters 数据集对比 |
| DeepMind 人才传承和创始团队 | 关键科学家转投 Xaira、Generate:Biomedicines 或学术实验室；Hassabis 精力分散在 DeepMind 和 Isomorphic | 中 | 要求披露人员流失数据、股权条款；确认 Hassabis 是否深度参与 Isomorphic 战略 |
| 大型药企合作广度（前 10 大药企中的 3 家） | ISM8969 Phase I 失败（end-2026）削弱扩展动力；竞争对手拿下下一波药企 AI 交易 | 高 | 跟踪 ISM8969 Phase I 入组；观察 Isomorphic 是否在 end-2026 前宣布第 4 个药企合作 |
| 私营公司信息优势 | 里程碑不透明，伙伴更难核实科学主张相对于上市竞品的可信度 | 低 | 要求 Lilly、Novartis、J&J 提供合作进展报告；评估未披露里程碑进展 |
风险登记表基于公开信息和 AI 与制药授权中的类似先例。严重程度为定性评级。
[CP011, CP012, CP013, CP023, CP027, CP033]截至 2026 年 5 月，从护城河维度比较 Isomorphic Labs 与领先竞争者竞争就绪度的关键绩效指标。
所有数值来自截至 2026 年 5 月的公开披露或分析师估计。Recursion 员工数按 Exscientia 合并后估计。
[CP015, CP021, CP024, CP033]3.4 竞争前景与尽调问题
由于结构生物学护城河和资本优势，Isomorphic 近期竞争前景有利，但实质上取决于预计 2026 年底出炉的 ISM8969 Phase I 结果。一旦 Phase I 失败，相比已经从 FAP 项目取得临床验证数据的 Recursion，Isomorphic 的定位会受到更大冲击。 Isomorphic 选择保持私营身份，同时累计融资 $2.7B，使其无需像上市竞争对手（Recursion、Schrödinger）那样按季度公开项目进展；后者必须及时披露里程碑挫折。这是结构性优势。不过，这种不透明也让合作伙伴更难独立核验进展。 从竞争策略看，应预期 Isomorphic 会把药企合作基础扩展到已披露的 3 家之外。如果未来 12-18 个月不能新增第四项合作，将是一个有意义的竞争信号。全行业尚无任何 AI 药物发现公司产出获批药物——包括 Isomorphic——这意味着任何玩家声称的竞争优势，都仍建立在验证前承诺上，而非已证明结果。 [CP031, CP032, CP033, CP034, CP035]
04财务情况
4.1 收入模式与合作经济性
Isomorphic Labs 的收入完全来自与大型药企签署的多年、多靶点研究合作协议。模式按里程碑设计：每项合作在签约时收到首付款，随后在候选物选择、IND 申报、Phase I/II/III 入组、监管批准等明确药物发现和开发事件触发里程碑付款，最终从任何商业获批药物中收取特许权使用费。截至 2026 年 5 月，已有 3 项合作公开披露：Eli Lilly（2024 年 1 月，首付款 $45M，里程碑最高 $1.745B）、Novartis（2024 年 1 月，首付款 $37.5M，里程碑最高 $1.2B，2025 年 2 月扩展）和 Johnson & Johnson（2026 年 1 月，财务条款未披露）。已确认收到的首付款现金合计 $82.5M；两笔已披露交易的最高潜在里程碑超过 $2.9B。自候选药物 ISM8969 于 2026 年 1 月获得 FDA IND 放行、Phase I 预计 2026 年底启动以来，尚未触发任何特许权使用费收入。Isomorphic 不收取 SaaS、API 或按次查询费用。全部收入来自定制化长期合作协议，导致收入确认不均匀、依赖里程碑，可能需要 10-15 年才达到峰值。24 个月内达成 3 项合作，按行业标准看交易节奏快且可重复；但合作价值超过 97% 仍取决于临床结果。 [CI001, CI002, CI003, CI004, CI005, CI017]
| 合作伙伴 | 首付款（$M） | 最高里程碑（$M） | 项目类型 | 签署日期 | 状态 |
|---|---|---|---|---|---|
| Eli Lilly | $45 | $1,745 | 多靶点小分子 | Jan 2024 | 活跃 |
| Novartis | $37.5 | $1,200 | 多靶点小分子（Feb 2025 扩展） | Jan 2024 | 活跃 |
| Johnson & Johnson | 未披露 | 未披露 | 跨模态多靶点（小分子、抗体、肽） | Jan 2026 | 活跃 |
| 合计（仅披露部分） | $82.5 | $2,945+ | 小分子与生物制剂 | 2024-2026 | 3 个活跃合作 |
Lilly 和 Novartis 条款来自官方公告；J&J 财务条款未公开披露。所有里程碑金额都是合同上限；期望值只是其中一部分，取决于临床成功率。
[CI001, CI002, CI005, CI018, CI039]| 维度 | 细节 | 证据级别 | 尽调问题 |
|---|---|---|---|
| 交易结构 | 多年多靶点合作：首付款 + 里程碑 + 分层版税 | 已确认（Lilly、Novartis） | 核实 J&J 结构是否一致 |
| 首付款区间（已披露） | 每笔 $37.5M-$45M | 两笔交易确认 | J&J 首付款金额 |
| 里程碑区间（已披露） | 每笔最高 $1.2B-$1.745B | 两笔交易确认 | J&J 里程碑安排 |
| 版税条款 | 未披露；行业惯例为净销售额 3-12% | 根据行业惯例推断 | 核实版税分层和叠加 |
| 单靶点定价 | 未披露；单靶点经济性隐含在多靶点结构中 | 未披露 | 单靶点费用或成功指标 |
| 收入集中度 | 三个合作伙伴；已确认首付款收入全部来自 Lilly 和 Novartis | 已确认 | 核实没有单一伙伴集中度超过 80% |
定价细节一部分来自公开交易公告，一部分按行业惯例推断。实际确认收入和合同细节均未公开披露。
[CI017, CI020, CI021, CI029]收入转换路径从 AI 平台能力出发，经过签约合作和收取预付款，走向里程碑达成和最终版税收入。合作价值 97% 以上后置到临床与监管里程碑。
时间线区间反映典型生物技术开发规律；Isomorphic 的 AI 逻辑瞄准早期阶段速度提升 2-3 倍。尚无公开确认已收到任何里程碑付款。
[CI017, CI029, CI006]4.2 成本结构与单位经济性
Isomorphic Labs 的成本基础符合投资阶段 AI 研究公司特征：人员密集、算力密集，且没有实体制造的销售成本。公司截至 2026 年 4 月有 371 名员工，人员是主要支出项。按伦敦市场 AI 和计算生物学研究人员全口径薪酬每人 $200-250K 估算（含福利、工资税和管理费用），年度人员成本接近 $75-90M。训练和运行大规模药物发现 AI 模型的算力基础设施，在可比公司每年估计增加 $50-80M；不过 Isomorphic 的成本部分由 Alphabet 的 Google Cloud 基础设施补贴，真实独立算力成本难以独立验证。伦敦、洛桑和美国的办公设施，以及 ISM8969 接近 Phase I 后增长的临床和监管运营，会进一步增加开销。年度现金消耗估计为 $150-250M；公司近期仍会深度亏损。一旦触发里程碑，毛利率预计较高（60-85%），因为 AI 生成药物设计预测的边际成本主要是算力，而不是实体制造。不过，公司没有公开披露任何毛利率数据，Alphabet 算力补贴也让真实收入成本估计存在重大不确定性。 [CI013, CI014, CI022, CI023, CI024, CI025]
| 成本类别 | 依据 | 年度成本估计 | 可信度 | 尽调路径 |
|---|---|---|---|---|
| 人员（R&D、BD、G&A） | 371 FTE x ~$200-250K 全包成本（伦敦 / 全球 AI 生物科技基准） | ~$75-90M/year | 低 | 从数据室获取员工构成和薪酬区间 |
| AI 计算基础设施 | 大模型 GPU/TPU 训练；Google Cloud 按内部转移价格计费 | ~$50-80M/year（毛额） | 低 | Alphabet 公司间计算协议；转移定价分析 |
| 设施（伦敦、洛桑、美国） | 办公室租约；实验室空间；共用 DeepMind 设施 | ~$10-20M/year | 低 | 租约协议；共享空间成本分摊 |
| 临床与监管运营 | ISM8969 IND 准备、Phase I CRO 成本、CMO 招聘（Nov 2025） | ~$10-30M/year（上升中） | 低 | CRO 合同；Phase I 试验预算披露 |
| 估计年度烧钱速度 | 上述类别合计 | ~$150-250M/year | 低 | FY2024/2025 Companies House 账目；管理层确认 |
所有数字均为分析师基于员工数和 AI 研究行业基准作出的估计；没有公开 P&L 数据。计算成本为毛额，未扣除任何 Alphabet 补贴抵消。
[CI013, CI014, CI022, CI023, CI024, CI025]Isomorphic Labs 的运营成本与收入流。估计年度烧钱规模远高于合作预付款；里程碑收入具有或有性且后置；$2.7B 资本基础为缺口提供资金。
所有成本数据都是分析师估计；尚无公开 P&L 数据。Alphabet 的算力补贴意味着总算力成本可能高估 Isomorphic 承担的经济成本。
[CI022, CI025, CI036]4.3 资本充足性与融资依赖
以所处开发阶段看，Isomorphic Labs 资本极其充足。2026 年 5 月的 $2.1B Series B 轮，加上 2025 年 4 月的 $600M Series A 轮，提供约 $2.7B 外部资本；已确认的 $82.5M 首付款和任何留存的 Series A 前 Alphabet 资金又增加了缓冲。按年烧钱 $150-250M 估计，Series B 后现金跑道约 8-17 年，足以将 ISM8969 推进 Phase I，并产出多个新增候选药物。Series B 由复投方 Thrive Capital 领投，并由 UK Sovereign AI Fund、Temasek 和 MGX 等战略投资者共同锚定，在追求财务回报之外也提供地缘政治背书。Alphabet 仍是控股股东，并参与两轮融资，实际上为短期资本风险托底。公司没有公开披露债务，不过未披露的 Alphabet 公司间安排（算力额度、共享服务）可能构成准债务或表外义务。Isomorphic 当前资本依赖较温和，下一轮外部融资可能在 5-7 年后，但如果临床里程碑滑坡、算力成本上升，或 Alphabet 战略优先级转移，资本依赖会变成风险。UK Companies House 记录确认 FY2024 完整账目已于 2025 年 9 月 30 日提交，这是尽调中最容易取得的正式财务记录。 [CI007, CI008, CI009, CI010, CI011, CI012]
| 指标 | 数值 / 估计 | 可信度 | 注释 |
|---|---|---|---|
| 外部累计融资 | ~$2.7B | 高（已确认） | Series A 轮 $600M（Apr 2025）+ Series B 轮 $2.1B（May 2026） |
| Series A 领投方和日期 | Thrive Capital，April 1, 2025 | 高（已确认） | GV 和 Alphabet 共同投资；首次外部融资 |
| Series B 领投方和日期 | Thrive Capital，May 12, 2026 | 高（已确认） | 共同投资方：Alphabet、GV、MGX、Temasek、CapitalG、UK Sovereign AI Fund |
| 已收到合作首付款 | $82.5M 已确认 | 高（已确认） | Lilly $45M + Novartis $37.5M；J&J 条款未披露 |
| 估计年度烧钱速度 | $150-250M/year | 低（分析师估计） | 未公开披露；由员工数和 AI 基础设施基准推导 |
| Series B 后隐含现金跑道 | ~8-17 年 | 低（推导） | 按 ~$2.5B+ 估计现金和 $150-250M 烧钱速度计算；不含里程碑收款 |
| 已披露估值 | 未披露 | N/A | Series A 和 Series B 的投后估值均未见报道 |
| 控股股东 | Alphabet Inc. | 高（已确认） | Alphabet 参与两轮融资；保留最大股权 |
资本数据来自官方公告。烧钱速度和现金跑道是分析师估计；确认数据需要 Companies House FY2024 账目和管理层披露。
[CI007, CI008, CI009, CI010, CI012, CI015]Isomorphic Labs 关键财务指标的低位和高位边界。已确认数据点不确定性为零；烧钱、现金跑道和利润率数字不确定性高，需要正式尽调确认。
已确认数字（预付款、融资轮次）估计误差接近零。烧钱、现金跑道和利润率区间反映分析师估计，不确定性高；需要正式尽调。
[CI014, CI024, CI037]4.4 财务缺口与结论
Isomorphic Labs 资本充裕，并在合作层面获得商业验证，但仅凭公开数据仍无法完成财务承销。三项优势突出：（1）$2.7B 资本基础对 AI 药物发现赛道具有变革性；（2）与 Lilly、Novartis、J&J 的 3 项确认药企合作，提供商业合法性和期权价值；（3）Alphabet 持续共同投资和基础设施支持，显著降低短期资本风险。三项担忧同样值得标注：（1）收入质量仍处商业化前阶段，超过 97% 的潜在合作价值依赖临床结果里程碑，而这些结果可能需要 10-15 年才出现；（2）Alphabet 基础设施补贴部分遮蔽真实运营经济性；（3）闭源 IsoDDE 架构阻止外部验证药物设计效果，使里程碑达成概率存在不确定性。财务结论是：资本充足，商业种子已播下；但完整承销需要进入正式数据室查阅 Companies House 账目、管理层烧钱速度披露、J&J 交易条款，以及对 IsoDDE 的独立技术尽调。资本强度相对当前收入偏高，但符合资产类别和阶段。 [CI026, CI027, CI028, CI029, CI035, CI037]
| 缺失指标 | 对投资判断的影响 | 严重程度 | 尽调路径 |
|---|---|---|---|
| 利润表（收入、COGS、EBITDA、毛利率） | 无法确认盈利能力、收入结构或真实成本结构 | 阻断性 | Companies House FY2024 完整账目（Sep 2025 提交）；FY2025 应于 Sep 2026 提交 |
| 烧钱速度和月度现金跑道 | 无法确认现金跑道或下一轮融资触发点 | 阻断性 | NDA 下管理层直接披露；经审计 FY2024 账目 |
| J&J 合作财务条款 | 无法计算整体合作组合价值或收入上限 | 重大 | 向 Isomorphic BD / 法务索取；若属重大事项，J&J 可能有 8-K 披露义务 |
| Alphabet 计算成本分摊和转移定价 | 无法判断独立资本强度或真实毛利率 | 重大 | 公司间协议审阅；正式数据室中的转移定价分析 |
| 里程碑达成概率（IsoDDE 验证） | 无法估计 $2.9B+ 里程碑管线的期望值 | 重大 | IsoDDE 独立技术尽调；药物项目临床状态审阅 |
所有缺口均基于截至 May 2026 的公开信息识别；进入数据室后可能出现更多缺口。
[CI026, CI027, CI028, CI035, CI040]Isomorphic Labs 的资本来源、运营流向和长期融资依赖。Alphabet 是战略后盾；里程碑收入是财务自给的唯一路径，需要十年的临床成功。
[CI031, CI032, CI038]4.5 图表与要点
05产品与技术
5.1 IsoDDE 平台定义与客户工作流
Isomorphic Labs 的 IsoDDE 平台切入药企药物发现的核心瓶颈：找到并优化一种药物分子，使其以高特异性、安全性和成药性结合目标蛋白。在传统药企工作流中，通过 X 射线晶体学确定靶点结构可能需要数月至数年，每个结构成本达数十万美元；高通量筛选（HTS）化合物库则要测试数十万至数百万个化合物，时间和成本都很高。IsoDDE 用 AI 预测替代或增强这些步骤。AlphaFold 3 可在数小时而非数月内预测目标蛋白的三维结构及其与候选药物分子的相互作用，并利用 EMBL-EBI AlphaFold 数据库中 200M+ 个结构。AlphaProteo 随后直接根据预测靶点结构设计新型蛋白结合分子（抗体替代物、肽、分子胶）。生成式化学模块则在预测结合口袋几何约束下设计新型小分子。公司内部领先候选物 ISM8969 于 2026 年 1 月获得 FDA IND 放行，是首个完全通过这一 AI 管线设计的药物，证明平台能在没有传统 HTS 或实验结构确定的情况下交付监管级候选物。 [CE001, CE002, CE003, CE005, CE006, CE011]
| 模块 / 资产 | 主要用户 | 状态 / 成熟度 | 关键差异化 | 尽调缺口 |
|---|---|---|---|---|
| AlphaFold 3 结构预测 | 内部科学家；药企伙伴 | 商业化（May 2024+） | 药物设计中的商业独家 AF3；联合蛋白-配体结构预测 | 真实合作项目中，AF3 相对 AF2 的准确率差异对命中率的影响 |
| AlphaProteo 蛋白结合体设计 | 内部科学家；蛋白疗法项目 | 研究 / 临床前（2024+） | 基准测试中，结合体亲和力较此前领先方法提升 3–4× | 湿实验验证规模和失败率未公开披露 |
| 生成式化学模块 | 内部科学家；药企伙伴项目 | 研究 / 临床前 | 受 AF3 结合口袋约束的结构感知 3D 分子生成 | 没有同行评议论文；方法未披露 |
| ISM8969 管线资产 | Isomorphic；FDA 监管路径 | IND 已获放行（Jan 2026）；Phase I 计划 end-2026 启动 | 内部管线首个完全由 AI 设计的 IND 候选药物 | 靶向适应症、作用机制和临床前数据未公开 |
| AlphaFold Server（非商业用途） | 学术界和政府研究人员 | 已上线；3M+ 用户；190+ 个国家 | 全球最大的公开蛋白质结构资源 | AlphaFold Server 的输出不得用于商业用途；不贡献收入 |
| IsoDDE 合作伙伴集成层 | 药企合作伙伴（Lilly、Novartis、J&J） | 活跃（截至 May 2026 已签 3 个合作） | 通过定制合作交付全栈 AI 药物设计 | API 架构、数据处理和集成规格未披露 |
基于截至 May 2026 公开可得的产品描述、合作公告和同行评审论文。除上述模块外，公司内部可能还有其他模块，但未公开披露。
[CE005, CE006, CE007, CE011, CE012, CE013]| 用户任务 | 现有工作流 | IsoDDE 方案 | 可量化收益 | 限制 |
|---|---|---|---|---|
| 靶点结构测定 | 实验性 X 射线晶体学（数月至数年；每个结构 $200K–$1M+） | AlphaFold 3 结构预测（数小时；数据库内 200M+ 个预计算结构） | 数周对比数年；新增结构的边际成本接近零 | AF3 置信度随蛋白家族而变；不能替代确认性实验 |
| 从已知靶点生成命中物 | HTS 筛选 100K–1M 化合物库（数月；成本高） | AF3 指导的虚拟筛选 + 生成化学模块 | 公司声称命中率提升 10–100×；可生成已知库之外的新 AI 化学型 | IsoDDE 尚未公开发布对比 HTS 的独立命中率基准 |
| 蛋白结合物与抗体替代物设计 | 噬菌体展示 / 酵母展示（6–18 个月；通量低） | AlphaProteo 生成式结合物设计（数天；高通量） | 结合亲和力基准提升 3–4×；新支架更多样 | 基准结果不保证能转化到临床；湿实验失败率未披露 |
| 小分子先导化合物优化 | 迭代式药物化学 + Schrödinger FEP（每轮数月） | 用 IsoDDE 做结构感知的 AI 引导生成式优化 | 设计—合成—测试循环更快；结构引导的方案降低脱靶风险 | 真实合作项目中，尚无公开的 Schrödinger FEP+ 直接对比 |
| 候选药物 ADMET 剖析 | 湿实验 ADMET 测定（每批数周；成本高） | AI 增强的 ADMET 预测整合进 IsoDDE 管线 | 预计降低淘汰率；加快多参数优化 | IsoDDE 的 ADMET 预测准确性尚未经过独立验证或公开发表 |
工作流描述基于公开产品介绍、合作公告和药物发现行业惯例。可量化收益来自公司声称或基准推导；真实项目结果未获独立验证。
[CE002, CE006, CE007, CE017, CE020, CE021]展示药企合作方从提交靶点到筛选候选物，如何用 IsoDDE 跑完整个端到端药物发现流程。
[CE002, CE005, CE017, CE020, CE021]5.2 平台架构与基础设施
IsoDDE 作为五层一体化 AI 栈运行。底层是数据层：RCSB Protein Data Bank（230K+ 个实验测定结构）、EMBL-EBI AlphaFold Database（200M+ 个 AI 预测结构）、用于配体数据的 Chemical Component Dictionary（CCD），以及来自持续合作项目的自有湿实验室反馈。核心 AI 模型层包含三个模块：AlphaFold 3，一个扩散 transformer 架构，可联合生成蛋白、配体、DNA 和 RNA 的全原子坐标预测；AlphaProteo，一个生成式蛋白结合分子设计模型；以及用于小分子提案的生成式化学引擎。所有训练和大规模推理都运行在 Alphabet 基础设施提供的 Google Cloud TPU 和 GPU 集群上。工作流编排层管理药物发现管线步骤——命中物生成、虚拟筛选、先导化合物优化和 ADMET 预测。合作伙伴集成层提供 API 和数据交换机制，药企合作方（Lilly、Novartis、J&J）通过该机制在 NDA 下提交靶点，并接收 AI 生成的候选物包。AlphaFold 3 GitHub 仓库已累计超过 12,000 个星标，反映开发者社区对底层方法的高度参与。关键在于，Alphabet 算力基础设施是核心运营依赖；真实独立算力成本未被独立披露。 [CE008, CE009, CE010, CE016, CE021, CE022]
| 层级 / 组件 | 在 IsoDDE 中的作用 | 关键依赖 | 主要风险 |
|---|---|---|---|
| AlphaFold 3 核心模型 | 预测蛋白、配体和复合物的全原子结构 | Google DeepMind IP 授权；Alphabet 计算基础设施 | 与 Alphabet/DeepMind 的 IP 或授权争议会打断整个平台 |
| AlphaProteo 结合物生成器 | 为治疗性蛋白靶点设计新型蛋白结合物 | 以 AlphaFold 3 结构预测为输入；依赖 DeepMind 模型权重 | 若 AF3 结构存在系统性错误，误差会级联传播 |
| 生成化学引擎 | 在 3D 结合口袋约束下设计新型小分子 | CCD/CSD 化学数据库；AF3 结合口袋预测 | 训练化学空间多样性限制模型准确性 |
| Google Cloud TPU/GPU 集群 | 为合作项目做大规模模型训练和批量推理 | Alphabet 基础设施合同；关联公司间计算定价 | 依赖单一计算供应商；真实成本未独立披露 |
| EMBL-EBI AlphaFold 数据库 | 200M+ 个预计算蛋白结构，支撑基于结构的设计 | EBI 与 DeepMind 持续执行托管和数据共享协议 | EBI 协议中断会削弱结构数据访问 |
| 合作伙伴数据交换层 | 接入湿实验反馈；在 NDA 下交换合作项目数据 | 药企合作伙伴 NDA；数据治理协议；英国数据保护法 | 合作伙伴撤回数据或发生法律争议，会打断数据飞轮模型 |
架构根据公开披露、AlphaFold 3 / AlphaProteo 论文和公司公告推断。内部系统架构规格未公开披露。
[CE008, CE009, CE010, CE016, CE021, CE022]从数据到合作伙伴集成的五层架构，由 AlphaFold 3、AlphaProteo 和生成式化学驱动。
架构依据公开 AlphaFold 3/AlphaProteo 论文和公司公告推断；内部系统边界并未公开披露。
[CE008, CE009, CE010, CE016, CE021, CE022]梳理支撑 IsoDDE 平台的关键依赖，从 Alphabet IP 与算力，到药企合作方的数据反馈。
[CE004, CE008, CE009, CE035, CE040]5.3 差异化、IP 与数据优势
Isomorphic Labs 结构上最可防守的优势，是 AlphaFold 3 在药物设计中的商业独占性。AlphaFold 3 的 Nature 论文（2024 年 5 月）明确将商业药物设计用途限制给 Isomorphic Labs；这意味着任何药企若不与 Isomorphic 合作，就不能在自有发现项目中使用 AF3，从而把强制合作机制嵌入竞争格局。OpenFold 和 Meta 的 ESMFold 等开源替代方案能复制部分 AlphaFold 2 能力，但缺少 AF3 的多分子复合物和配体预测能力，而这正是药物设计的核心差异化。AlphaProteo（2024 年 9 月）在头对头基准中，相比此前领先的蛋白设计方法，展示出结合亲和力和特异性的中位数 3–4× 提升，并在 deepmind.google 和 Science 上作为能力证据发布。这个一体化全栈平台把结构预测、蛋白结合分子设计和生成式化学放入一个连贯工作流；相比之下，Schrödinger 等单模块竞争对手依赖基于物理的分子对接和 FEP+，且需要实验结构作为输入；Recursion 则侧重表型筛选，而非结构驱动。来自每个合作伙伴湿实验室反馈的数据飞轮，会逐步提升模型准确性。AF3 限制构成结构性竞争壁垒：药企若不与 Isomorphic Labs 签约，就无法独立使用 AF3 做药物设计。 [CE004, CE007, CE015, CE018, CE019, CE034]
截至 2026 年 5 月，IsoDDE 平台能力在各开发阶段的成熟度。
成熟度评估依据公开产品描述、同行评议论文和合作公告推断；IsoDDE 不披露内部能力路线图。
[CE004, CE005, CE006, CE011, CE036]5.4 信任、安全与合规
Isomorphic Labs 受到多层合规义务约束。ISM8969 获得 FDA IND 放行（2026 年 1 月），是公司监管层面合规状态最强的外部验证：FDA 审查了 ISM8969 的临床前材料，并放行其进入 Phase I 人体试验，意味着公司可以生成可用于 IND 申报的文档。不过，这只适用于单一候选物；IsoDDE 本身并未独立认证或审计。截至 2026 年 5 月，IsoDDE 平台没有公开披露任何 GxP（Good Practice）或 ISO 13485 软件质量认证；对于一个生成临床开发候选药物的系统，这是重大缺口。公司内部维护生物安全控制，以防止 AlphaFold 3 和 AlphaProteo 被滥用于设计有害生物制剂；EMBL-EBI AlphaFold Server 执行非商业服务条款，限制服务器输出。EU AI Act（2024 年 8 月生效）可能将用于临床决策支持的 AI 系统归为高风险，从而为 Isomorphic 的欧盟活动带来未来合格评定义务。公司未公开披露第三方 AI 审计记录。作为一家英国注册公司，处理合作伙伴数据时可假定需要遵守 GDPR；但向美国云基础设施跨境传输数据的条款尚未独立确认。这些缺口对投资阶段 AI 药物设计公司而言并不罕见，但在全面监管部署前都属于尽调事项。 [CE011, CE024, CE025, CE027, CE032, CE033]
| 控制 / 认证 / 指标 | 状态 | 范围 | 缺口 |
|---|---|---|---|
| FDA IND 放行（ISM8969） | January 2026 获放行 | 仅 ISM8969；单个 AI 设计候选药物 | 其他管线资产尚未具备 IND 申报条件；尚无 NDA 或 ANDA 申报 |
| 生物安全审查协议 | 公司称内部政策已启用 | 所有使用 AF3 和 AlphaProteo 的蛋白与分子设计工作流 | 未披露独立外部审计或第三方生物安全认证 |
| GDPR / UK GDPR 数据合规 | 英国注册公司处理欧盟合作伙伴数据，推定处于执行状态 | 合作伙伴协作数据；AlphaFold Server 研究人员数据 | 向美国 Google Cloud 跨境传输数据未获独立确认；与 EU AI Act 的适用关系仍不清楚 |
| EU AI Act 分类 | 监测 / 推进中（EU AI Act 自 Aug 2024 生效） | 用于临床药物设计管线的 AI 模型（可能属高风险） | 截至 May 2026，未披露符合性评估或第三方 AI 审计记录 |
| GxP / ISO 13485 软件认证 | 未公开披露 | 临床管线软件（IsoDDE 输出支撑 IND 阶段决策） | 未确认 GxP 或 ISO 13485 认证；这是临床阶段 AI 软件的重大缺口 |
合规状态根据公开公告、FDA IND 放行新闻和监管惯例推断。IsoDDE 软件认证状态尚未获得独立确认。
[CE011, CE024, CE025, CE032, CE033, CE039]5.5 路线图、部署与集成
IsoDDE 的部署模式完全定制化：没有自助 API、应用市场或开放访问的商业端点。每项药企合作都是多年定制项目，Isomorphic 科学家与合作伙伴团队并肩工作，把平台应用到具体靶点上。这带来高切换成本和深度集成，但也限制了可扩展性和快速签约。公开披露的平台路线图以 ISM8969 在 2026 年底启动 Phase I 临床试验为中心，这将是 Isomorphic Labs 的 AI 设计候选药物首次获得人体临床读数。除 ISM8969 外，公司没有公开披露额外内部管线资产或重大平台能力发布路线图。2025 年 2 月 Novartis 合作扩展和 2026 年 1 月 J&J 交易，显示商业动能延续；IsoDDE 数据飞轮会随着每个新增合作项目而改善。接入药企合作伙伴工作流需要深度 IT、数据治理和科学协同，这些通过 NDA 约束的数据交换协议处理。Alphabet 持续提供基础设施和 IP 合作，带来连续性；但如果 Alphabet 战略优先级明显偏离当前 AI 药物发现重点，这一依赖也会形成治理风险。 [CE012, CE013, CE023, CE030, CE031, CE035]
| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 含义 | 来源 |
|---|---|---|---|---|
| 2021 | Isomorphic Labs 作为 Alphabet/DeepMind 分拆公司成立；IsoDDE 平台启动 | 完成 | AI 药物设计独立成公司；Alphabet 是核心 IP 和计算供给方 | isomorphiclabs.com |
| Jul 2022 | AlphaFold 2 数据库通过 EMBL-EBI 扩展至 200M+ 个蛋白结构 | 完成 | 为 IsoDDE 结构预测模块提供训练数据和免费层验证基础 | AlphaFold 服务器：alphafold.ebi.ac.uk |
| Jan 2024 | 与 Eli Lilly（$45M 首付款）和 Novartis（$37.5M 首付款）签署研究合作 | 完成 | IsoDDE 首次外部商业部署；平台获前五大药企验证 | isomorphiclabs.com |
| May 2024 | AlphaFold 3 发表于 Nature；商业药物设计用途仅限 Isomorphic Labs | 完成 | 结构性竞争壁垒确立；AF3 独占权阻止药企独立使用 | nature.com |
| Sep 2024 | Google DeepMind 发布 AlphaProteo 蛋白结合物设计工具 | 完成 | 新的结合物设计模块为 IsoDDE 平台增加蛋白疗法模态 | deepmind.google |
| Jan 2026 | ISM8969 获 FDA IND 放行；成为首个完全由 AI 设计的 IND 候选药物 | 完成 | 平台获得临床验证；Isomorphic 首个 AI 设计候选药物进入监管路径 | fiercebiotech.com |
| End-2026 | ISM8969 Phase I 临床试验启动（计划） | 计划中 | 首次获得 AI 设计候选药物的人体安全性和药代动力学读数 | fiercebiotech.com |
路线图里程碑基于公开公告、新闻稿和同行评审论文。除 ISM8969 外，内部开发里程碑和未来管线候选药物未公开披露。
[CE001, CE006, CE011, CE018, CE023, CE031]5.6 图表与要点
06客户情况
6.1 客户基础概览
Isomorphic Labs 是一家 B2B 药物发现平台，已确认的药企合作伙伴有三家：Eli Lilly、Novartis 和 Johnson & Johnson。三家公司按收入均位列全球前 15 大药企。合作在 2024 年 1 月至 2025 年 3 月之间签署，距公司商业化启动大约 15 个月。所有交易都采用研究合作结构，覆盖 AI 辅助药物候选物设计，并包含首付款，以及与发现和开发成功挂钩的里程碑付款。截至 2026 年 5 月，公司没有面向消费者或中小企业的客户。只服务药企的客户结构，反映出 Isomorphic Labs 的战略重点：锁定有能力在 AI 生成候选物之后推进临床项目的大型机构。 [CU001, CU002, CU003, CU016]
| 合作伙伴 | 买方类型 | 用例 | 规模 | 战略价值 | 关键缺口 |
|---|---|---|---|---|---|
| Eli Lilly | 大市值药企 | 多靶点 AI 药物设计 | 全球前十药企 | 高 — 早期标志性交易 | 首付款 / 里程碑未披露 |
| Novartis | 大市值药企 | AI 药物设计 + AlphaFold 3 | 全球前十药企 | 高 — 2025 年扩展 | 收入和里程碑条款未公开 |
| Johnson & Johnson | 大市值药企 | AI 辅助分子设计 | 全球前五药企 | 高 — 新近进入 | 所有财务条款未披露 |
分层数据仅来自公开披露的合作公告；收入区间未披露。
[CU001, CU002, CU003]6.2 Eli Lilly 合作
Eli Lilly 合作于 2024 年 1 月 3 日宣布，是 Isomorphic Labs 最早期的大型商业协议之一。这是一项多靶点研究合作，Lilly 使用 Isomorphic 的 AI 平台，在肿瘤学、代谢疾病等多个疾病领域识别和优化药物候选物。Isomorphic Labs 将这笔交易描述为签署时规模最大的 AI 药物发现合作之一。交易包含一笔首付款，但金额未披露；另有取决于研发成功的里程碑付款。Lilly 投资者关系页面确认了该合作，使其作为一手客户验证来源具备较高佐证力度。Isomorphic 与 Lilly 在 2024 年 1 月 3 日同步发布公告，对交易形成强双边确认。 [CU002, CU004, CU005, CU006, CU023]
| 指标 | 值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| 已签约药企合作伙伴 | 3 | May 2026 | PR Newswire / IR 页面 | 高 | 15 个月内跑出商业速度 | 可触达药企潜在客户池总量未知 |
| 合作续约 / 扩展 | 1（Novartis） | 2025 | BioIndustry / BioSpace | 中 | 至少一个合作伙伴按计划推进 | Lilly 和 J&J 续约意向未知 |
| 处于 IND 申报支持研究的候选药物 | 0（已披露） | May 2026 | 无公开披露 | 高 | 临床验证完全未证实 | 内部管线进展未披露 |
| 已披露收入 | None | May 2026 | STAT News / Bloomberg | 高 | 收入前阶段；仅靠里程碑的模式 | 可能已收到首付款，但金额未量化 |
| Series B 融资 | $2.1B | May 2026 | PR Newswire | 高 | 多年现金跑道；部分商业信号 | 不是收入指标；不能验证客户付款 |
所有数值均基于已公开宣布的交易；实际收入、使用率和临床产出未披露。
[CU001, CU016, CU021]6.3 Novartis 合作
Novartis 在 2024 年 1 月与 Isomorphic Labs 开始初始研究合作，日期与 Lilly 公告同日。2025 年，Novartis 扩大合作，覆盖更多治疗领域，并把 Isomorphic 的 AlphaFold 3 衍生模型接入其内部发现管线。Novartis 公开称，这项合作对其长期 AI 药物设计能力具有战略意义。该扩展是 Isomorphic 三笔交易中唯一有记录的续约事件，为合作伙伴满意度提供了部分证据。BioSpace 和 BioIndustry Association 等行业媒体独立佐证了这次扩展。2025 年续约也是 Isomorphic 已披露三笔交易中，最清晰的合作伙伴满意度和平台价值证据。 [CU007, CU008, CU009, CU021]
| 客户 | 客群 | 部署 / 用例 | 状态 | 成效证据 | 限制 |
|---|---|---|---|---|---|
| Eli Lilly | 大市值药企 | 多靶点 AI 药物候选设计 | 活跃（研究阶段） | Lilly IR 确认；Isomorphic 确认 | 未确认临床产出或里程碑付款 |
| Novartis | 大市值药企 | AI 药物设计、AlphaFold 3 集成 | 活跃（2025 年扩展） | Novartis IR + 行业媒体确认扩展 | 未披露达成里程碑或确认收入 |
| Johnson & Johnson | 大市值药企 | AI 辅助分子设计 | 活跃（研究阶段） | J&J IR 确认；pharmaphorum 佐证 | 所有财务条款未披露；没有结果数据 |
生产或试点状态根据合作公告措辞推断；没有独立运营确认。
[CU002, CU003, CU007]6.4 Johnson & Johnson 合作
Johnson & Johnson 于 2025 年 3 月与 Isomorphic Labs 签署研究合作，覆盖 AI 辅助分子设计。J&J 这笔交易承接了公司更广泛的 AI 药物发现战略，并由 J&J Investor Relations 和 Isomorphic 自己的 PR Newswire 通稿共同确认。财务条款没有公开披露，包括首付款、里程碑或治疗范围，因此它是三项合作中最不透明的一项。pharmaphorum 和 Fierce Biotech 等行业媒体也独立报道了公告。截至 2026 年 5 月，Isomorphic 的合作伙伴页面列出三笔交易均仍在进行；截至 2026 年 5 月，本次尽调审阅的任何监管文件或新闻报道中都没有披露终止。三笔活跃交易覆盖两家总部在美国、一家总部在瑞士的大型药企。 [CU003, CU024, CU025, CU026]
6.5 集中度风险、留存与商业展望
Isomorphic Labs 只有三家客户，且全部是大型药企，因此面临显著的行业和客户集中度风险。公司未披露收入数据；截至 2026 年 5 月，合作产生的任何药物候选物都尚未进入临床试验，里程碑能否转化完全没有验证。Nature 2025 年对 AI 药物发现合作的分析指出，随着药企建设内部 AI 能力，AI 公司的专有数据优势可能被削弱；该领域的里程碑转化率历史上也偏低。BenevolentAI 与 AstraZeneca 的交易在 2023 年因未达成临床里程碑而结束，说明了行业风险。Evaluate Vantage 对 Series B 的分析指出，Isomorphic 仍处于收入前阶段，里程碑时间线不确定。Novartis 扩展是积极的留存信号，但三家合作伙伴中只有一家公开续约。$2.1 billion Series B 为公司验证临床价值提供了现金跑道，但投资者引用的商业理由是已有合作关系，而不是收入。 [CU010, CU011, CU012, CU013, CU015, CU018]
| 指标 | 值 / 状态 | 客群 | 置信度 | 尽调问题 |
|---|---|---|---|---|
| NRR / GRR | 未披露 | 所有合作伙伴 | 低（无数据） | 要求提供按合作收入队列拆分的财务模型 |
| 合作续约 | 3 个中 1 个（Novartis 2025 年扩展） | Novartis | 中 | 在管理层访谈中确认 Lilly 和 J&J 的续约意向 |
| 流失 / 终止 | 公开已知为 0 | 所有合作伙伴 | 中 | 通过 IR 询问核实是否存在未公开终止 |
| 合作伙伴满意度调查 | 无公开信息 | 所有合作伙伴 | 低（无数据） | 要求与 Lilly 和 J&J 合作负责人做背调访谈 |
| 合同期限 | 未披露 | 所有合作伙伴 | 低（无数据） | 在正式尽调中获取条款清单摘录 |
公开披露中没有任何留存指标；标为 null 的值表示没有公开证据；尽调问题反映关键数据需求。
[CU021, CU024, CU010]| 风险维度 | 集中度 / 扩张驱动 | 影响 | 尽调路径 |
|---|---|---|---|
| 客户数 | 3 个客户，均为大型药企 | 高 — 单点故障风险 | 将管线拓展至新兴生物科技公司或政府资助方 |
| 行业集中 | 仅药企；没有生物科技、农业或政府客户 | 高 — 监管逆风风险 | 跟踪未来融资材料中的管线多样性 |
| 地理集中 | 美国 + 瑞士；没有 APAC 或 EM | 低 — 均为领先市场 | 审查 J&J 和 Lilly 的国际 R&D 范围 |
| 里程碑转化 | 行业历史平均水平偏低；没有 Isomorphic 数据 | 高 — 收入模式承压 | 尽调中要求提供里程碑达成数据 |
| 内部 AI 威胁 | 药企在自建内部 AI；Isomorphic 护城河可能被侵蚀 | 中 — 长期风险 | 每年评估合作伙伴是自建 AI 还是采购 |
风险评级为分析师估计，基于已披露合作伙伴数量和行业先例；并非基于 Isomorphic 内部数据。
[CU012, CU013, CU018, CU019]6.6 图表
07风险
7.1 监管与法律风险
Isomorphic Labs 处在 AI 与药物发现的交叉地带，监管审查正在加码。EU AI Act（Regulation 2024/1689）自 2024 年 8 月生效，将用于医疗和药物开发的 AI 系统列为高风险，要求开展合规评估、履行透明度义务，并保留扎实文档。Isomorphic Labs 注册于英格兰和威尔士，但其药企合作伙伴——Lilly（美国）、Novartis（瑞士）和 J&J（美国）——会在欧盟和美国监管机构均有权监管医疗 AI 应用的司法辖区使用平台输出。英国 Information Commissioner's Office（ICO）也单独发布了健康数据相关 AI 使用指南，指出即使数据在模型训练前已去标识化，GDPR 义务仍然适用。FTC 在 2024 年启动对 AI 基础模型的广泛调查，可能波及 Isomorphic 这类平台公司。知识产权风险同样重要：AI 生成药物候选物的所有权尚无明确法律先例；Isomorphic 依赖的训练数据来源未公开披露，可能带来类似 AI 内容生成诉讼中的版权和 IP 挑战。Companies House 文件显示，Isomorphic Labs Ltd（company 13036082）是一家私人有限公司，未公开披露诉讼；但没有披露并不等于没有风险。 [CR001, CR002, CR003, CR004, CR005]
| 风险 | 司法辖区 | 可能性 | 影响 | 状态 | 缓释路径 |
|---|---|---|---|---|---|
| EU AI Act 合规（医疗高风险 AI） | 欧盟 | 高 | 高 | 监管环境活跃 | 合格评定；合作伙伴合规义务 |
| 英国 ICO GDPR / 健康数据处理 | 英国 | 中 | 中 | ICO 于 2024 年发布 AI 指引 | 数据治理计划；法律审查 |
| FTC AI 平台问询 | 美国 | 中低 | 中 | FTC 于 2024 年启动问询 | 监测；问询扩大时聘请律师 |
| AI 生成 IP 归属争议 | 全球 | 中 | 高 | 尚无稳定先例 | IP 策略复盘；合同 IP 转让条款 |
| Alphabet 反垄断审查外溢 | 美国 / 欧盟 | 低 | 高 | Alphabet 受 DOJ 审查 | 保持独立治理；独立董事会 |
风险评级为分析师估计；概率和影响为定性评估，依据监管文件和行业先例。
[CR001, CR002, CR003, CR004, CR005]7.2 运营与关键人物风险
Isomorphic Labs 的运营风险主要集中在关键人物依赖和计算基础设施集中度。Demis Hassabis 同时领导 Google DeepMind，并担任 Isomorphic Labs 的战略旗帜人物。无论是 DeepMind 领导层交接、个人情况变化，还是优先事项竞争，只要他的可投入度发生变化，都可能显著损害 Isomorphic 的科学信誉和合作伙伴信任。关键人物之外，Isomorphic 也深度依赖 Google DeepMind 的计算基础设施和 AlphaFold 产品线；公司没有公开披露自己是使用独立算力，还是完全依赖云端资源。AI 药物发现系统的运营风险还包括模型幻觉和可靠性失效：模型信心很高地预测某个分子结构，但湿实验验证证明错误，从而浪费合作伙伴资源并损害信任。BenevolentAI 在 AstraZeneca 交易失败后于 2023 年进行战略重组，说明运营和科学可靠性失效可以迫使公司快速重构。人才风险也较高：AI 药物发现研究员薪酬溢价明显，Isomorphic 要与 Google DeepMind、OpenAI 和大型药企内部 AI 团队争夺同一批人才。 [CR006, CR007, CR008, CR009, CR010]
| 风险 | 可能性 | 影响 | 指标 | 缓释措施 |
|---|---|---|---|---|
| 关键人物：Demis Hassabis 兼任双重角色 | 中 | 高 | DeepMind 领导层变动公告 | 接班计划；科学领导层纵深 |
| AI 模型幻觉 / 可靠性失效 | 中 | 高 | 合作伙伴湿实验室验证失败率 | 集成验证；人工参与审核 |
| 计算基础设施集中（Google Cloud） | 低 | 高 | Google Cloud 违反 SLA 或 Alphabet 战略转向 | 多云路线图；合同 SLA 执行 |
| 数据安全 / 泄露影响药企合作伙伴 IP | 中低 | 高 | 任何通报的数据泄露或 ICO 调查 | SOC 2 / ISO 27001 认证；渗透测试 |
| 科学复现失败（AlphaFold 3 局限） | 中 | 中 | 质疑 AF3 在新场景准确性的学术论文 | 透明基准测试；参与开放挑战 |
严重性估计基于行业可比公司（BenevolentAI、Exscientia）以及公开运营信息。
[CR006, CR007, CR008, CR009]7.3 合作伙伴与依赖风险
Isomorphic Labs 的收入模型完全依赖三家药企合作伙伴——Eli Lilly、Novartis 和 J&J——把研究里程碑转化为付款。这形成悬崖式集中度风险：任一交易失败或终止，都会抹去相当大比例的潜在里程碑价值。Recursion 与 Exscientia 在 2024 年合并，说明承压的 AI 药物发现公司可能选择整合，而不是继续维持合作项目。Alphabet 依赖是另一项结构性风险：Isomorphic 的 AlphaFold 产品线、算力接入和 DeepMind 人才管线都流经 Alphabet。如果 Alphabet 因战略重心转移、监管行动或反垄断补救措施而降低药物发现优先级，Isomorphic 的核心技术资产可能受损。云基础设施依赖又增加了一层风险：Google Cloud 服务一旦中断，将直接影响 Isomorphic 的 AI 平台运营和合作伙伴承诺。公开信息没有显示公司具备多云或本地部署冗余策略。 [CR011, CR012, CR013, CR014, CR015]
| 风险 | 受影响合作伙伴 | 可能性 | 影响 | 缓释路径 |
|---|---|---|---|---|
| 合作协议终止（里程碑失败） | Lilly、Novartis、J&J | 中 | 极高 | 分散合作伙伴基础；增加新兴生物科技公司 |
| Alphabet 战略上降低药物发现优先级 | 全部（计算、IP、人才） | 低 | 极高 | 独立治理；替代计算选项 |
| 大型药企自建 AI 能力 | 全部 | 高 | 高 | 加深自有数据护城河；扩大 AlphaFold 3 优势 |
| 云集中度（仅 Google Cloud） | 全部 | 低 | 高 | 多云或混合计算策略 |
| Recursion/Exscientia 合并带来的竞争压力 | 间接受全部影响 | 高 | 中 | 监测竞争对手能力；加速差异化 |
合作伙伴终止概率依据行业先例推断；暂无 Isomorphic 专属情报。
[CR011, CR012, CR013, CR014]7.4 人员与执行风险
作为一家 AI 优先的药物发现公司，Isomorphic Labs 的执行风险集中在少数世界级 AI 研究员身上；他们的专长直接支撑平台差异化。公司与 Google DeepMind、Microsoft Research 以及资金充足的 AI 制药同行争夺同一人才池。顶尖 AI 研究员的薪酬预期自 2022 年以来大幅上升；Isomorphic Labs 作为私人公司，在与上市公司股权包竞争时面临挑战。执行风险还包括把 AI 预测结构转化为已验证药物候选物的复杂性：计算预测必须通过湿实验测试、ADMET 画像和临床安全筛查，才会触发任何里程碑。公司没有披露内部临床项目路线图，说明当前执行重点仍在研究阶段，而不是转化交付。如果合作伙伴药物候选物未能通过湿实验验证，里程碑转化会停滞，续约概率也会下降。 [CR016, CR017, CR018, CR019, CR020]
| 风险 | 可能性 | 影响 | 领先指标 | 缓释措施 |
|---|---|---|---|---|
| 关键 AI 研究员流失（流向 Google DeepMind、OpenAI） | 中 | 高 | LinkedIn 高管离职公告 | 长期股权激励；有竞争力的薪酬对标 |
| 未能招募临床转化人才 | 中 | 中 | 尚未公开任命 CMO 或临床负责人 | IND 准备前组建临床团队 |
| 湿实验室验证失败率超出计划 | 中 | 高 | 到 2027 年未公开披露任何临床候选药物 | 增加湿实验室合作；扩充实验验证能力 |
| 执行复杂度：AI 到临床转化 | 高 | 高 | 到 2028 年仍无 IND 申报 | 设定转化里程碑；发布阶段性管线数据 |
| 一人兼任领导（Hassabis 同时任职 DeepMind 与 Isomorphic） | 中 | 高 | Hassabis 公开时间分配信号 | 任命 CEO 负责 Isomorphic 日常运营 |
人才市场数据基于行业报告；公开渠道没有 Isomorphic 专属员工数或薪酬数据。
[CR016, CR017, CR018, CR019]7.5 财务、模型与缓释因素
按传统口径，Isomorphic Labs 仍处于收入前阶段：尽管有三项活跃药企合作，公司没有披露任何收入数据。财务风险集中在烧钱速度不透明、依赖里程碑以及 AI 算力资本强度。2026 年 5 月完成的 $2.1 billion Series B 提供了多年现金跑道；但如果三到四年内里程碑转化没有兑现，公司可能需要进一步稀释融资，或与 Alphabet 建立战略合作。公司的近开源策略——AlphaFold 作为研究工具公开可用——形成一个悖论：广泛的科学采用强化品牌，却降低底层模型的独特性，削弱定价权。Nature 2025 年的报道警告称，随着药企建设内部 AI 团队，AI 药物发现公司的专有数据优势可能被削弱，这直接挑战 Isomorphic 的长期价值主张。缓释因素包括 Novartis 扩展合作（证明一定留存）、$2.1B 现金跑道，以及 Alphabet 的声誉背书。该投资逻辑的终止标准包括：两项或更多药企合作终止、到 2028 年仍未披露任何 IND 阶段候选物，或 Demis Hassabis 离开 Isomorphic 治理结构。 [CR021, CR022, CR023, CR024, CR025, CR026]
| 维度 | 现有缓释措施 | 终止标准 / 投资逻辑破裂触发器 |
|---|---|---|
| 合作伙伴集中度 | $2.1B 现金跑道；Novartis 扩约信号 | 2 个以上药企合作终止且无替代 |
| 临床验证 | 活跃研究合作；AlphaFold 3 管线 | 到 2028 年底未披露 IND 阶段候选药物 |
| 关键人物 | Alphabet/DeepMind 声誉背书 | Demis Hassabis 退出 Isomorphic 治理 |
| 监管 | 英国/美国法律团队；Alphabet 监管经验 | 平台遭遇重大 EU AI Act 执法行动 |
| 财务 | $2.1B Series B（May 2026） | 到 2028 年烧钱速度超过 $500M/year 且没有里程碑收入 |
| 竞争护城河 | AlphaFold 系列资产；自有训练数据 | 药企合作伙伴披露内部 AI 能力超过 Isomorphic 平台 |
终止标准由分析师定义，用于触发投资逻辑破裂；不代表 Isomorphic Labs 管理层立场。
[CR021, CR022, CR023, CR024, CR025, CR026]7.6 图表
08估值
8.1 投资逻辑与反向逻辑
进入 2026 年，Isomorphic Labs 已成为全球资本最充足的私人 AI 药物发现平台，并在 2026 年 5 月获得 Thrive Capital 领投的 $2.1 billion Series B 融资。投资逻辑建立在三根收敛支柱上。第一，平台底座 AlphaFold3 在蛋白质结构和分子相互作用预测上展现出经过同行评审的科学优势，而这是早期药物发现的主要瓶颈。第二，Eli Lilly（最高 $1.7B）、Novartis（$1.0B）和 Johnson and Johnson（条款未披露）构成的 $2.7 billion 名义合作组合，证明顶级药企确实赋予该平台实质价值。第三，Google DeepMind 持续提供战略和算力支持，让 Isomorphic Labs 获得基础设施优势，包括前沿模型接入、TPU 配额和人才网络；独立竞争者很难在相近规模上复制。反向逻辑同样扎实。截至报告日，三项药企合作全部仍处于研究阶段；公司未公开披露任何临床项目、IND 申报时间线或 Phase 1 准备度。缺少临床证据时，合作名义价值只是附条件的履约义务，而不是已赚收入。Isomorphic Labs 不披露收入、烧钱速度或运营费用，无法开展常规估值分析。PitchBook 和 Dealroom 估计，其 Series B 投后估值为 $5B-$8B；相较财务已披露的上市同行，这代表结构性溢价。建议观点为有条件正面：在隐含 Series B 条款或以下进入，并设置明确里程碑门槛，投资期限为 5-8 年。BenevolentAI 在 AstraZeneca 合作失败后于 2024 年重组，构成行业层面的警示先例，说明 AI 药物发现平台可以维持多年研究合作，却无法把合作转化为临床项目。 [CV001, CV002, CV003, CV004, CV005, CV006]
| 维度 | 评估 | 置信度 | 备注 |
|---|---|---|---|
| 投资建议 | 有条件正面 | 中 | 以 Series B 隐含 $5B-$8B 估值或更低价格进入 |
| 风险评级 | 高 | 高 | 临床前、未披露收入、依赖单一母公司 |
| 估值立场 | 有条件支持溢价 | 中 | AlphaFold3 护城河加 $2.7B 合作账面规模，支撑相对上市可比公司的溢价 |
| 进入条件 | 24 个月内确认 IND 申报 | 中 | 没有临床就绪信号，不能上调为无条件正面 |
| 时间周期 | 距有意义流动性仍需 5-8 年 | 中 | M&A 退出最可能；Phase 2 前 IPO 可能性低 |
| 决策含义 | 观察；按 Series B 条款共同投资；12 个月按触发条件复盘 | 中 | 每季度监测合作动态与 Alphabet Other Bets 披露 |
所有评估均为分析师基于公开数据作出的估计；未使用非公开信息。
[CV039, CV040]| 因素 | 投资逻辑论据 | 反向逻辑论据 | 净评估 |
|---|---|---|---|
| AlphaFold3 平台 | 已在同行评审中展现蛋白结构优势；CASP15 胜出者 | 未披露湿实验室验证数据；学术基准不等于临床证明 | 有保留正面 |
| 药企合作 | $2.7B 名义合作账面；Lilly、Novartis、J&J 均为一线交易对手 | 全部取决于或有里程碑；未确认预付款收入；仅处临床前 | 有保留正面 |
| Alphabet 计算资源 | 可用前沿 TPU；DeepMind 人才管线；持续战略承诺 | 单点依赖；Other Bets 退出风险；无合同最低承诺 | 有条件正面 |
| 财务状况 | $2.1B Series B 提供与同业相当的 6-8 年现金跑道 | 无收入、未披露烧钱速度；股权结构表未知；可能有优先股压力 | 有保留正面 |
| 估值 | 差异化支撑相对上市同业 $5B-$8B 的溢价 | 披露收入尚未验证溢价；所有估计均为分析师代理值 | 谨慎 |
净评估仅反映分析师基于公开证据作出的判断。
[CV001, CV003]8.2 融资背景与估值纪律
2026 年 5 月的 Series B（$2.1B）由 Thrive Capital 领投，Alphabet 继续参与，使 Isomorphic Labs 成为全球单轮规模最大的 AI 药物发现融资之一。关键是，公司没有披露优先股堆叠、清算优先权或股权结构表拆分，仅靠公开数据无法判断普通股价值。所有企业价值估计都依赖 PitchBook、Dealroom 和 Compworth 的分析师模型，而不是公司公开披露。PitchBook 估计 Series B 投后估值为 $5B-$8B。Dealroom 和 Compworth 的估计大体一致，上端反映 Google 母公司溢价。作为参照，Recursion Pharmaceuticals 是唯一披露财务的上市纯 AI 药物发现公司；截至 2026 年 Q1，其市值为 $2.5B-$3.5B，季度收入约 $65M。Schrodinger Inc. 2024 年收入约 $170M，交易倍数约为收入的 9-15x。Isomorphic Labs 没有任何披露收入，却隐含 $5B-$8B 估值；只有在平台未来 3-5 年内拿到临床阶段概念验证时，这一结构性溢价才站得住。Alphabet Other Bets 部门没有单独披露 Isomorphic Labs 的烧钱速度或财务数据。可比阶段 AI 药物发现公司的基准现金消耗（Recursion：每年约 $280M）显示，$2.1B 现金跑道若按同行水平支出，可支撑 6-8 年运营，为公司延长临床准备时间提供财务稳定性。Series B 优先股结构带来的稀释风险未知但重要，必须在任何共同投资决定前的正式尽调中解决。 [CV009, CV010, CV011, CV012, CV013, CV014]
8.3 情景分析：乐观、基准与悲观
鉴于公司没有披露财务，且临床阶段药物发现结果具有二元性，情景分析是估值评估的核心。三个情景不以收入倍数锚定，而是以合作经济性、里程碑转化概率和竞争动态为锚。在乐观情景（概率信号：20-25%）下，Isomorphic Labs 到 2029-2030 年至少推进一个化合物取得 Phase 2 疗效数据。这将验证平台转化能力，引发现有合作伙伴（Lilly、J&J、Novartis）的收购兴趣，并支撑 $20B-$50B 的企业价值。乐观情景假设名义合作价值中有 60-80% 转化为里程碑，并且 Google DeepMind 在整个临床时间线内持续提供战略承诺。在基准情景（概率信号：50-55%）下，二到三个治疗项目到 2028 年从研究推进到 IND 申报。合作经济性维持在名义交易价值的 40-50%。Alphabet 维持战略持股，但不加速整合。该情景下企业价值为 $8B-$15B，意味着按 Series B 条款进入后的升值幅度有限，退出时间为 7-10 年。在悲观情景（概率信号：20-25%）下，到 2029 年没有临床项目完成 IND 申报。一项或多项药企合作被重组，经济性显著降低。Xaira Therapeutics、AbSci 这类资金充足的竞争者，或药企内部 AI 团队率先拿到临床证据，削弱 Isomorphic 的平台溢价。退出价值跌破 Series B 进入价格，Series B 投资者遭受资本损失。Recursion 季度烧钱数据和 PitchBook 关于临床前平台结果的基准概率数据，都支持至少给该情景分配 20% 概率。 [CV017, CV018, CV019, CV020, CV021, CV022]
| 情景 | 概率信号 | 关键假设 | 退出企业价值 | 回报倍数（相对 Series B 进入价） |
|---|---|---|---|---|
| 乐观情景 | 20-25% | 到 2030 年取得 Phase 2 疗效数据；70%+ 里程碑转化；药企收购 | $20B-$50B | 3-8x |
| 基准情景 | 50-55% | 到 2028 年提交 2-3 项 IND；40-50% 里程碑转化；战略 M&A 或 IPO | $8B-$15B | 1.2-2.3x |
| 悲观情景 | 20-25% | 到 2029 年仍无 IND；合作伙伴重组；竞争对手先取得临床证明 | $1B-$4B | 0.1-0.6x（亏损） |
概率信号是分析师基于行业基准率和公司特定证据作出的估计；不是财务模型输出。
[CV017, CV018, CV019]8.4 可比公司与相对价值
Isomorphic Labs 没有完全对应的上市可比公司。它是一家私人、AI 原生、收入前药物发现平台，背靠 Google，拥有 $2.7B 名义药企合作账面。因此，可比集合必须横跨上市的软件赋能药物发现公司，以及私人 AI 药物发现队列。在上市可比中，Recursion Pharmaceuticals（RXRX）是主要基准：它是一家 AI 优先的药物发现公司，披露财务，采用平台合作模型，并有上市市值。Recursion 2026 年 Q1 每季度现金消耗约 $70M、季度收入运行率约 $65M，对应 $2.5B-$3.5B 市值和 10-15x 收入倍数。Schrodinger SDGR 提供计算药物发现 SaaS 基准，交易倍数为收入的 9-15x。两家上市同行相较 Isomorphic Labs 估计估值都有显著折价，反映私人市场溢价、AlphaFold3 差异化和 Alphabet 背书。私人可比中，Xaira Therapeutics 在 2024 年作为收入前平台融资 $1.0B；AbSci Corp 公开交易，收入倍数较低；Insilico Medicine 走临床优先策略。没有一家在药企合作规模或计算优势深度上与 Isomorphic 直接匹配。BCG 对 AI 制药行业的分析和 PitchBook 私人市场数据确认，AI 药物发现平台溢价高度集中在两三家计算差异化最清晰的公司；如果临床缺口按计划补上，Isomorphic Labs 的溢价具备合理性。 [CV025, CV026, CV027, CV028, CV029, CV030]
| 公司 | 阶段 | 状态 | 估值 / 市值 | 收入 | 收入倍数 | 备注 |
|---|---|---|---|---|---|---|
| Isomorphic Labs（估计） | 临床前 | 未上市 | $5B-$8B 估计 | 未披露 | N/A（无收入） | 2026 年 Series B 投后估值；分析师估计 |
| Recursion Pharmaceuticals | 平台 + 临床管线 | 上市（RXRX） | $2.5B-$3.5B | $65M/季度（Q1 2026） | 10-14x 前瞻收入 | 主要上市 AI 药物发现可比公司 |
| Schrodinger Inc. | 计算平台 | 上市（SDGR） | $1.5B-$2.5B | $170M (FY2024) | 9-15x 收入 | 计算药物发现 SaaS 基准 |
| Xaira Therapeutics | 临床前 AI 平台 | 未上市 | 未披露 | 收入前 | N/A | 2024 年融资 $1B；最接近的未上市可比公司 |
| AbSci Corp | 平台 + 合作 | 上市（ABSI） | ~$1.0B | <$50M | 20x+ 收入 | 规模较小；药企合作层级较低 |
Isomorphic Labs 估值数字为分析师估计。其他公司均基于截至 Q1-Q2 2026 的公开申报和市场数据。
[CV025, CV026, CV027]8.5 退出、终止触发条件与最终尽调问题
最可能的退出路径，是由 Isomorphic Labs 的三家药企合作伙伴之一进行战略收购：Eli Lilly（市值约 $500B）、Johnson and Johnson（约 $400B）或 Novartis（约 $250B）。三家公司都有能力按乐观情景估值收购；合作结构也可能产生平台合作协议中常见的优先购买权动态。鉴于目前没有披露收入，且数十亿美元估值的临床前生物技术平台在公开市场环境中承压，在 Phase 2 临床数据前 IPO 退出可能性不高。三项打破投资逻辑的触发条件会促使立即退出。第一，Series B 交割后 24 个月内任何重大药企合作中止或重组，说明平台验证失败。第二，AlphaFold 生成的分子设计无法以临床有意义的比例完成湿实验验证，而这正是平台全部价值主张的底层机制。第三，有证据显示 Alphabet 下调战略承诺：削减 DeepMind 算力分配、关键技术领导离职，或在 Other Bets 组合中重新分类。共同投资决定前必须解决的尽调事项包括：完整股权结构表和优先股结构、运营烧钱速度和现金跑道模型、最先进研究项目的 IND 申报时间线，以及约束 Alphabet 算力支持承诺的合同条款。此外，合作首付款的收入确认政策也需要澄清。经风险调整后，建议仍是在 Series B 进入价格下有条件正面，并在 IND 申报确认前采取持有并监控姿态。 [CV033, CV034, CV035, CV036, CV037, CV038]
| 触发器 | 事件类型 | 监测信号 | 应对动作 |
|---|---|---|---|
| 合作伙伴退出或重组 | 合作失败 | 药企合作伙伴 IR 公告；合同修订 | 立即复核退出投资；下调至回避 |
| 湿实验室验证失败 | 平台能力 | 科学论文；合作伙伴对里程碑保持沉默 | 有条件持有；要求管理层更新 |
| Alphabet 计算资源撤回 | 战略母公司风险 | Other Bets 板块降本；DeepMind 重组 | 立即复核退出投资；下调至回避 |
| 竞争对手临床证明 | 竞争风险 | Xaira、Insilico 或药企内部 AI Phase 2 读数 | 重新评估平台溢价；复盘竞争护城河 |
| Series B 现金跑道耗尽（<18 个月） | 财务风险 | Companies House 申报；管理账目请求 | 有条件持有；要求披露烧钱速度 |
触发器按紧急程度降序排列。任何单一 1 类触发器都会立即启动全面投资复核。
[CV035, CV036]| 尽调事项 | 优先级 | 所需信息 | 获取渠道 |
|---|---|---|---|
| 股权结构表与优先权结构 | 关键 | 完整股权结构表；清算瀑布；反稀释条款 | 投资者权利协议；Series B 条款清单 |
| 运营烧钱速度与现金跑道 | 关键 | 月度和年度 opex；当前现金头寸；现金跑道模型 | 管理账目；Companies House 损益表 |
| IND 申报时间表 | 关键 | 先导适应症；最早 IND 申报日期；I 期启动日期 | 直接接触管理层；专利申请文件 |
| Alphabet 算力条款 | 高 | 期限；定价；TPU 分配；治理；退出条款 | 基础设施服务协议；管理层 |
| 收入确认政策 | 中 | 合作预付款如何确认收入；里程碑会计处理 | 审计业务约定书；管理层讨论 |
五项都是阻塞项：不全部解决，就无法作出共同投资决定。公开来源没有解决其中任何一项。
[CV037, CV038]8.6 图表
免责声明
本报告是截至 2026 年 5 月 16 日由自动化 AI 研究生成的尽调摘要。报告仅基于公开信息，不构成投资建议。作出任何投资决定前， 所有财务数据都应对照一手来源独立核验。Isomorphic Labs 是私人公司；估值、烧钱速度和现金跑道估计来自第三方数据库和媒体报道， 可能无法反映公司实际数据。本报告作者和分发方不对文中信息的准确性或完整性作任何陈述。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Isomorphic Labs Limited was incorporated in the United Kingdom on February 24, 2021, with company number 13223825. | 高 | SO002, SO003 |
| CO002 | Isomorphic Labs was publicly announced on November 4, 2021, with Demis Hassabis named as CEO. | 高 | SO003, SO004, SO005 |
| CO003 | Isomorphic Labs is a subsidiary of Alphabet Inc., the parent company of Google. | 高 | SO001, SO003 |
| CO004 | Isomorphic Labs is headquartered in London, United Kingdom. | 高 | SO001, SO002 |
| CO005 | The company opened a Lausanne, Switzerland office in December 2022. | 中 | SO009 |
| CO006 | A Cambridge, Massachusetts office was opened in mid-2025. | 中 | SO009 |
| CO007 | The company name 'isomorphic' reflects the founders' belief in a deep symmetry between biology and information science. | 中 | SO001, SO009 |
| CO008 | Isomorphic Labs' business model centers on licensing its AI drug design capabilities to pharmaceutical companies via collaboration agreements featuring upfront payments, milestones, and royalties. | 高 | SO001, SO020, SO021 |
| CO009 | Sir Demis Hassabis is the CEO and Founder of Isomorphic Labs, simultaneously serving as CEO and Founder of Google DeepMind. | 高 | SO006, SO003, SO004 |
| CO010 | Hassabis's dual CEO role at both DeepMind and Isomorphic Labs creates a key-person concentration risk and privileged access to DeepMind's foundational AI research. | 中 | SO006, SO008 |
| CO011 | Max Jaderberg serves as President of Isomorphic Labs, having been promoted from Chief AI Officer. | 高 | SO006, SO007 |
| CO012 | Miles Congreve is the Chief Scientific Officer of Isomorphic Labs, bringing 30+ years of pharmaceutical chemistry and structure-based drug design experience. | 高 | SO006, SO007 |
| CO013 | Sergei Yakneen serves as Chief Technology Officer of Isomorphic Labs. | 高 | SO006, SO007 |
| CO014 | Pamela Carroll serves as Chief Operating Officer of Isomorphic Labs. | 高 | SO006, SO007 |
| CO015 | Dr. Ben Wolf was appointed Chief Medical Officer of Isomorphic Labs in mid-2025, leading clinical translation. | 中 | SO007, SO009 |
| CO016 | Sarah Korman serves as Chief Business and Legal Officer of Isomorphic Labs. | 高 | SO006, SO007 |
| CO017 | Isomorphic Labs employs 371 people as of April 30, 2026, according to Tracxn data. | 中 | SO007 |
| CO018 | Isomorphic Labs raised $600 million in a Series A round closed on April 1, 2025, led by Thrive Capital with co-investors GV and Alphabet. | 高 | SO022, SO029 |
| CO019 | The Series A was led by Joshua Kushner, CEO of Thrive Capital. | 高 | SO022, SO029 |
| CO020 | Isomorphic Labs raised $2.1 billion in a Series B round announced on May 12, 2026, led by Thrive Capital. | 高 | SO011, SO012, SO010 |
| CO021 | Series B co-investors include Alphabet, GV, MGX, Temasek, CapitalG, and the UK Sovereign AI Fund. | 高 | SO011, SO012 |
| CO022 | Total external capital raised by Isomorphic Labs is approximately $2.7 billion (Series A $600M + Series B $2.1B). | 高 | SO010, SO011, SO013 |
| CO023 | Ruth Porat, President and CIO of Alphabet, publicly endorsed the Series B round. | 中 | SO010, SO011 |
| CO024 | The $2.1 billion Series B is the second-largest single biotech financing in history, surpassed only by Altos Labs' $3 billion round. | 高 | SO010, SO013 |
| CO025 | Post-money valuation has not been disclosed for either the Series A or Series B rounds. | 高 | SO010, SO011 |
| CO026 | AlphaFold 2 won the 2020 CASP14 competition and is credited with solving the 50-year protein structure prediction problem. | 高 | SO008, SO009 |
| CO027 | AlphaFold 3 was published in Nature on May 8, 2024 (doi:10.1038/s41586-024-07487-w) and extends predictions to proteins, DNA, RNA, small molecules, and their interactions. | 高 | SO016, SO017, SO018 |
| CO028 | AlphaFold 3 was shown to be 50% more accurate than traditional methods on the PoseBusters benchmark. | 中 | SO016, SO017 |
| CO029 | Eli Lilly collaboration announced January 8, 2024: $45 million upfront payment, up to $1.7 billion in milestone payments, covering multiple small molecule drug targets. | 高 | SO020, SO021 |
| CO030 | Novartis collaboration announced January 8, 2024: $37.5 million upfront, up to $1.2 billion in milestones; expanded February 2025 with 3 additional research programs. | 高 | SO020, SO021 |
| CO031 | Johnson & Johnson collaboration announced January 8, 2026, covering cross-modality multi-target drug design; Isomorphic handles AI design and J&J handles experimental validation through Janssen Biotech. | 高 | SO023, SO024 |
| CO032 | ISM8969, Isomorphic Labs' first internally AI-designed drug candidate, received FDA IND clearance on January 28, 2026. | 高 | SO025, SO026 |
| CO033 | The clinical trial timeline for ISM8969 was delayed by 12 months, from end-2025 to end-2026, announced at WEF Davos on January 20, 2026. | 高 | SO025, SO026 |
| CO034 | The IsoDDE (Isomorphic Drug Design Engine), announced February 10, 2026, more than doubles AlphaFold 3's accuracy on the protein-ligand structure prediction generalization benchmark. | 中 | SO019, SO014 |
| CO035 | No AI-designed drug has received regulatory approval as of May 2026, meaning ISM8969 would be the first if it succeeds in clinical trials. | 高 | SO010, SO025 |
| CO036 | The free AlphaFold server serves over 3 million researchers in 190+ countries as of mid-2026. | 中 | SO001, SO016 |
| CO037 | Isomorphic Labs' internal pipeline focuses on oncology, immunology, and inflammation, all pre-clinical as of May 2026. | 中 | SO010, SO007 |
| CO038 | At the CASP16 competition in November 2024, AlphaFold 3-based models did not significantly outperform older methods for protein-ligand interactions. | 中 | SO009 |
| CO039 | The participation of sovereign wealth funds (MGX from UAE, Temasek from Singapore, UK Sovereign AI Fund) in the Series B signals geopolitical interest in AI drug discovery infrastructure. | 中 | SO011, SO010 |
| CO040 | Isomorphic Labs' combined milestone potential across Lilly and Novartis partnerships exceeds approximately $3 billion before royalties. | 高 | SO020, SO021 |
| CM001 | The global AI in drug discovery market was valued at approximately $2.58 billion in 2025, according to Mordor Intelligence. | 中 | SM001, SM003 |
| CM002 | The global AI in drug discovery market is growing at a compound annual growth rate of approximately 25-30%, per analyst consensus. | 中 | SM001, SM002, SM003 |
| CM003 | Venture investment into AI drug discovery reached $4.1 billion in 2025, an increase of 38% year-on-year. | 中 | SM003, SM016 |
| CM004 | Different analyst methodologies produce very different AI drug discovery market estimates — from $2.6B (licensing-only) to $6B+ (deal-inclusive) — making TAM-based analysis unreliable. | 中 | SM001, SM003 |
| CM005 | Venture capital investment in AI drug discovery ($4.1B in 2025) substantially exceeds the commercial market revenue ($2.6B), suggesting capital is running ahead of commercial proof points. | 中 | SM003, SM016 |
| CM006 | Isomorphic Labs' total capital raised (~$2.7B) exceeds the entire AI drug discovery market revenue estimate for 2025 on some methodologies, making it by far the most capitalized private company in the sector. | 高 | SM016, SM018, SM019 |
| CM007 | Traditional drug development costs approximately $2.0-2.6 billion per approved drug over a 10-15 year timeline. | 中 | SM013, SM003 |
| CM008 | Approximately 90% of drugs entering clinical trials fail to reach regulatory approval. | 中 | SM006, SM010 |
| CM009 | AI-first drug discovery can compress the pre-clinical phase from 4-6 years to 12-18 months, according to company and analyst claims. | 中 | SM015, SM006 |
| CM010 | 78% of the top-20 pharmaceutical companies use AI platforms for drug discovery as of 2025-2026, per Excelra survey data. | 中 | SM002, SM003 |
| CM011 | No AI-designed drug has received regulatory approval globally as of May 2026, meaning clinical proof of concept remains the key market gating factor. | 高 | SM023, SM007 |
| CM012 | The clinical trial timeline for Isomorphic's ISM8969 was delayed 12 months from end-2025 to end-2026, illustrating that AI drug discovery execution risk remains real even for the best-funded companies. | 高 | SM007, SM023 |
| CM013 | Isomorphic Labs' IsoDDE platform claims to compress drug design cycles significantly; however, ISM8969's trial delay shows AI platforms do not eliminate timeline risk. | 中 | SM007, SM006, SM015 |
| CM014 | Isomorphic Labs primarily addresses the small molecule lead optimization and hit identification segment, with the J&J 2026 collaboration representing an expansion into multi-modality design. | 高 | SM015, SM018, SM024 |
| CM015 | The near-term buyer pool for Isomorphic Labs' collaboration model is approximately 15-25 large global pharmaceutical and biotech companies capable of funding milestone-based partnerships. | 中 | SM016, SM024 |
| CM016 | Academic research is the largest user base for Isomorphic's AI tools (3M+ AlphaFold users) but generates no direct revenue and serves only as an ecosystem and talent development channel. | 中 | SM015, SM014 |
| CM017 | The multi-modality AI drug design segment — spanning biologics, nucleic acids, and small molecules — represents a larger share of pharma R&D than small molecules alone, making the J&J deal a strategic expansion. | 中 | SM018, SM024 |
| CM018 | The top-20 pharma companies account for an estimated 65-70% of global pharmaceutical R&D spending, making them the primary buyer segment for AI drug discovery services. | 中 | SM002, SM003 |
| CM019 | The free AlphaFold server serves over 3 million researchers in 190+ countries and creates ecosystem value, indirectly supporting Isomorphic's commercial positioning. | 中 | SM015, SM021 |
| CM020 | The FDA has not issued specific guidance on AI-generated IND submissions as of May 2026, creating regulatory ambiguity for AI drug companies. | 中 | SM023, SM011 |
| CM021 | The Recursion-Exscientia merger, completed in late 2025, created the largest pure-play public AI drug discovery company. | 高 | SM009, SM025 |
| CM022 | Schrödinger (NASDAQ: SDGR) holds a dominant position in physics-based computational drug design with $500M+ in cash and an established pharma customer base. | 中 | SM003, SM025 |
| CM023 | Xaira Therapeutics launched in 2024 with $1 billion in initial capital, targeting AI generative biology for biologics drug design. | 中 | SM003, SM013 |
| CM024 | CASP16 benchmark results from November 2024 showed that AlphaFold 3-based models did not significantly outperform older methods for protein-ligand binding prediction. | 中 | SM011, SM025 |
| CM025 | The competitive landscape in AI drug discovery includes Recursion, Schrödinger, Insilico Medicine, BenevolentAI, Xaira, and Nabla Bio as primary competitors to Isomorphic Labs. | 中 | SM003, SM009, SM022 |
| CM026 | Isomorphic Labs' Series A ($600M) and Series B ($2.1B) make it the most capitalized private AI drug discovery company, ahead of Recursion and Schrödinger on private capital basis. | 高 | SM016, SM018, SM019 |
| CM027 | BenevolentAI has experienced partnership setbacks including AstraZeneca ending their collaboration in 2023, illustrating adverse execution risk in AI drug discovery. | 中 | SM025, SM003 |
| CM028 | Isomorphic Labs' $82.5M in upfront partnership payments (Lilly $45M + Novartis $37.5M) represents validated commercial revenue, with ~$3B in milestone potential if programs advance. | 高 | SM024, SM018 |
| CM029 | The Isomorphic Labs IsoDDE platform announcement in February 2026 claimed over 2x accuracy improvement vs. AlphaFold 3 on protein-ligand generalization benchmarks. | 中 | SM015, SM017 |
| CM030 | The growth drivers for AI drug discovery adoption are primarily economic: pharma's need to reduce the $2B+ cost and 10-15 year timeline of traditional drug development. | 中 | SM007, SM013, SM003 |
| CM031 | Isomorphic Labs' sovereign wealth co-investors in the Series B — MGX (UAE), Temasek (Singapore), and the UK Sovereign AI Fund — signal geopolitical alignment with UK-based AI infrastructure investment. | 高 | SM018, SM016 |
| CM032 | The AI drug discovery market is transitioning from promise-based to evidence-based competition, where clinical outcomes will increasingly differentiate platform quality. | 中 | SM006, SM010, SM011 |
| CM033 | Insilico Medicine received IND clearance for ISM001-055 (an AI-designed drug) in 2021, making it the first company to demonstrate the regulatory pathway for AI-designed drugs. | 中 | SM025, SM013 |
| CM034 | Switching costs from incumbent physics-based platforms (e.g., Schrödinger FEP+, CDD Vault) are a meaningful adoption constraint for pharma transitioning to AI-first drug design. | 中 | SM003, SM022 |
| CM035 | Isomorphic Labs has disclosed no revenue figures for its collaboration agreements other than upfront payments, making commercial market share assessment impossible from public information. | 高 | SM016, SM018 |
| CM036 | The AI drug discovery market is primarily US-centric (major companies: Recursion, Schrödinger, Xaira in USA), with UK as the leading European hub (Isomorphic Labs, BenevolentAI) and growing Asia-Pacific activity (Insilico Medicine in Hong Kong). | 中 | SM003, SM025 |
| CP001 | Recursion Pharmaceuticals completed its acquisition of Exscientia for approximately $688M (all-stock deal) in January 2025, creating the largest publicly listed combined AI drug discovery company. | 高 | SP009, SP008 |
| CP002 | Recursion achieved first clinical validation of its full-stack AI Operating System in FAP (familial adenomatous polyposis) in early 2026, demonstrating predictable pharmacokinetics and a well-tolerated safety profile for REC-1245. | 高 | SP001, SP008 |
| CP003 | Insilico Medicine has 40+ pipeline programs as of 2026, including 13 with IND approval; its INX-020/ISN21 TNIK inhibitor is in Phase II for fibrotic diseases, making it the most advanced purely AI-discovered proprietary pipeline among private competitors. | 高 | SP002, SP008 |
| CP004 | BenevolentAI proposed delisting from Euronext Amsterdam via merger into Osaka Holdings in February 2025 and announced a major strategic overhaul in December 2024, signaling financial distress and a loss of market confidence as a public AI drug discovery company. | 高 | SP003, SP008 |
| CP005 | Xaira Therapeutics emerged from stealth in April 2024 with $1B in initial capital, building an independently trained generative AI drug discovery platform targeting both small molecules and biologics with a de novo design approach. | 高 | SP004, SP023 |
| CP006 | Absci is building an AI Drug Creation Platform focused on biologics, using generative AI integrated with wet lab cycles of approximately 6 weeks; its lead program ABS-201 targets androgenetic alopecia via prolactin receptors and was developed in 24 months from concept to IND-ready. | 中 | SP005, SP008 |
| CP007 | Relay Therapeutics uses its Dynamo® platform to model protein conformational dynamics (motion in HD rather than static snapshot), enabling it to drug historically intractable targets; its pipeline focuses on precision oncology and genetic disease. | 中 | SP006, SP008 |
| CP008 | Schrödinger's computational platform is built on physics-based simulations and FEP+ (free energy perturbation), widely regarded as the highest-accuracy approach for lead optimization; it is deployed across the majority of the top 20 global pharmaceutical companies. | 高 | SP007, SP024 |
| CP009 | Schrödinger reported Q1 2026 financial results in May 2026, confirming continued software revenue growth; Schrödinger competes with Isomorphic as a computational chemistry tool provider but is not primarily a drug pipeline company. | 中 | SP007 |
| CP010 | Atomwise leverages its AtomNet deep learning platform with a library of over 3 trillion synthesizable compounds; a 2024 study showed AtomNet identified structurally novel hits for 235 of 318 targets evaluated, and the company has nominated a TYK2 inhibitor as its first IND candidate. | 中 | SP008 |
| CP011 | Isomorphic Labs holds exclusive commercial rights to AlphaFold3 for drug discovery under a licensing arrangement with Google DeepMind, creating a structural biology moat that no direct competitor has legal access to replicate. | 高 | SP011, SP012 |
| CP012 | IsoDDE, announced by Isomorphic in February 2026, is claimed by the company to achieve greater than 2x the accuracy of AlphaFold3 on structure prediction benchmarks; no independent peer review or third-party validation was publicly available at the time of this report. | 中 | SP010, SP013 |
| CP013 | Recursion's acquisition of Exscientia added generative molecular design capabilities to its phenomics platform, creating a combined entity that claims coverage of both phenotypic and structure-based drug design—an intentional competitive expansion into Isomorphic's structural domain. | 中 | SP009, SP001 |
| CP014 | Insilico Medicine's INX-020/ISN21, a Phase II program for pulmonary fibrosis discovered entirely by its Pharma.AI platform, is the most advanced proprietary AI-designed drug in clinical trials among Isomorphic's closest analogs as of May 2026. | 高 | SP002, SP008 |
| CP015 | The Isomorphic-Eli Lilly collaboration ($45M upfront + up to $1.7B milestones) and Isomorphic-Novartis collaboration ($37.5M upfront + up to $1.2B milestones) are among the largest disclosed upfront payments in AI drug discovery to a company without Phase I trial data. | 高 | SP014, SP016 |
| CP016 | Isomorphic's collaboration with Johnson & Johnson (January 2026) has not disclosed financial terms, preventing direct peer benchmarking of deal quality relative to the Lilly and Novartis agreements. | 中 | SP015, SP019 |
| CP017 | Schrödinger's LiveDesign and FEP+ platform is deployed at over 90% of the top 20 global pharmaceutical companies, giving it uniquely broad pharma penetration that no AI-native drug discovery company matches. | 中 | SP007, SP024 |
| CP018 | Recursion's phenomics database contains petabytes of biological data from its high-throughput robotic laboratory operations, providing a phenotypic screening moat that pure structural-biology or generative-chemistry companies including Isomorphic cannot easily replicate. | 中 | SP001, SP008 |
| CP019 | Absci's 24-month development timeline for ABS-201 from concept to IND-ready stage using generative AI, if replicated consistently, would represent a meaningful speed advantage in biologics discovery relative to traditional 4-6 year timelines. | 中 | SP005 |
| CP020 | Atomwise signed a strategic multi-target research collaboration with Sanofi leveraging its AtomNet platform for up to five drug targets, demonstrating that mid-tier AI discovery companies can secure major pharma partnerships despite smaller capital bases. | 中 | SP008 |
| CP021 | Isomorphic Labs had 371 employees as of April 30, 2026 (per Tracxn), compared to Recursion's estimated approximately 1,500 headcount post-Exscientia merger; Isomorphic appears more capital-efficient relative to its collaboration portfolio size. | 中 | SP022, SP001 |
| CP022 | Xaira Therapeutics had not disclosed any pharma collaboration agreements or clinical programs as of May 2026, remaining pre-revenue and pre-clinical despite launching in April 2024 with $1B in capital. | 中 | SP004, SP023 |
| CP023 | BenevolentAI's baricitinib repurposing discovery (identified via AI, validated in COVID-19 treatment through AstraZeneca) is widely cited as the first real-world validation of AI in drug discovery, but it was a repurposing of an existing drug, not discovery of a new molecular entity. | 中 | SP003, SP008 |
| CP024 | All major AI drug discovery companies including Recursion, Insilico, Isomorphic, and BenevolentAI rely on pharma partnership milestones for revenue; none has a standalone product revenue model sufficient to sustain operations independently as of May 2026. | 高 | SP001, SP002, SP003, SP016 |
| CP025 | IsoDDE, Isomorphic's drug design engine released February 2026, is the first AI drug design model claimed to outperform AlphaFold3 on standardized benchmarks, trained on proprietary data from Isomorphic's pharma partnerships; the superiority claim relies on internal benchmarks only. | 中 | SP010, SP012 |
| CP026 | Labiotech's 2025 review identified at least 12 AI drug discovery companies making material progress, including Anima Biotech, Atomwise, BPGbio, Cradle Bio, Iktos, and Insilico Medicine, confirming significant market fragmentation beyond the five most-funded players. | 中 | SP008 |
| CP027 | As of May 2026, no major AI drug discovery platform company—Recursion, Schrödinger, Insilico, BenevolentAI, Xaira, or Isomorphic—has received regulatory approval for a drug discovered and designed entirely by AI without human medicinal chemistry iteration. | 高 | SP001, SP002, SP003, SP007 |
| CP028 | Isomorphic's ISM8969 is among the first cohort of drug candidates where the complete design process—target identification, molecule design, lead optimization—was performed exclusively by AI without human medicinal chemistry iterations, as claimed by the company. | 中 | SP010, SP021 |
| CP029 | Relay Therapeutics focuses specifically on conformational dynamics (protein motion) rather than static structure, which is complementary to but distinct from Isomorphic's AlphaFold3-based static structure prediction approach, placing them in an adjacent rather than direct competitive position. | 中 | SP006 |
| CP030 | Schrödinger's mission is to transform drug and materials discovery using physics-based molecular simulation; as a tool provider rather than primarily a drug pipeline company, Schrödinger is more of an incumbent infrastructure competitor than a direct pipeline rival to Isomorphic. | 高 | SP007, SP024 |
| CP031 | Isomorphic Labs' three disclosed pharma collaborations (Lilly, Novartis, J&J) represent three of the world's top 10 pharmaceutical companies by revenue, indicating it has positioned at the highest partnership tier rather than building volume through smaller deals. | 高 | SP014, SP015, SP018 |
| CP032 | Isomorphic's private company status allows it to avoid public quarterly reporting of program updates and milestone setbacks, a structural advantage versus publicly listed competitors (Recursion, Schrödinger) that must disclose adverse developments promptly. | 中 | SP013, SP022 |
| CP033 | If Isomorphic fails to announce a fourth pharma collaboration within 12-18 months of the May 2026 Series B, this would be a meaningful adverse competitive signal suggesting that deal quality is harder to replicate than the fundraise narrative implies. | 中 | SP013, SP023 |
| CP034 | The Lilly upfront ($45M) is larger than total disclosed upfronts from Atomwise's Sanofi deal and estimated comparable early-stage pharma collaborations at Insilico, suggesting Isomorphic commands premium deal terms attributable to AlphaFold3 differentiation. | 中 | SP014, SP008 |
| CP035 | Isomorphic Labs faces an asymmetric competitive situation: success of ISM8969 Phase I creates a durable and sector-wide proof point for AI-only drug design, while failure generates a sector-wide reputational risk that would affect investor confidence in all AI drug discovery companies. | 中 | SP020, SP021 |
| CI001 | Eli Lilly agreed to pay Isomorphic Labs $45M upfront in their January 2024 collaboration, with up to $1.745B in potential milestone payments for multi-target small molecule drug discovery. | 高 | SI013, SI012 |
| CI002 | Novartis agreed to pay Isomorphic Labs $37.5M upfront in their January 2024 collaboration, with up to $1.2B in potential milestone payments for multi-target small molecule drug discovery. | 高 | SI013, SI012 |
| CI003 | The combined Lilly and Novartis partnerships carry up to approximately $3B in potential milestone payments excluding royalties, as disclosed at deal signing in January 2024. | 高 | SI013, SI014 |
| CI004 | As of May 2026, Isomorphic Labs' revenue consists entirely of upfront and milestone payments from pharma collaborations; no commercial product sales or royalty income has been publicly confirmed. | 中 | SI007, SI009 |
| CI005 | Johnson & Johnson and Isomorphic Labs entered a cross-modality multi-target collaboration in January 2026 covering small molecules, antibodies, peptides, and molecular glues; financial terms were not publicly disclosed. | 高 | SI008, SI007 |
| CI006 | Isomorphic Labs' proprietary IsoDDE model is kept closed-source, preventing external scientific validation of its drug design efficacy and drawing criticism from the open-science community. | 中 | SI006 |
| CI007 | Isomorphic Labs raised $600M in its first external funding round (Series A) on April 1, 2025, led by Thrive Capital, with participation from GV and follow-on from existing investor Alphabet. | 高 | SI004, SI005 |
| CI008 | Isomorphic Labs raised $2.1B in its Series B funding round on May 12, 2026, led by Thrive Capital, with co-investors including Alphabet, GV, MGX, Temasek, CapitalG, and the UK Sovereign AI Fund. | 高 | SI009, SI010, SI016 |
| CI009 | Total external capital raised by Isomorphic Labs is approximately $2.7B across two rounds as of May 2026, in addition to internal Alphabet funding prior to the Series A. | 高 | SI009, SI004, SI001 |
| CI010 | Prior to the April 2025 Series A, Isomorphic Labs was entirely funded by Alphabet as a wholly owned subsidiary, with no external investors, confirming the Series A was the first external equity issuance. | 高 | SI001, SI003 |
| CI011 | UK Companies House filings show share allotments of GBP 293 million in March-August 2025, consistent with the Series A capitalization event. | 中 | SI003 |
| CI012 | Isomorphic Labs' valuation has not been publicly disclosed in either the Series A or Series B announcements; no price-per-share or post-money valuation has been reported. | 高 | SI004, SI009, SI010 |
| CI013 | Isomorphic Labs had approximately 371 full-time employees as of April 30, 2026, according to Tracxn company data. | 中 | SI017 |
| CI014 | Annual cash burn is estimated at approximately $150M-$250M per year based on 371 FTEs at an estimated $200-250K all-in per London/global research employee plus AI compute infrastructure costs; this estimate is not publicly confirmed. | 低 | SI017, SI022 |
| CI015 | The $2.1B Series B is described by Isomorphic Labs as funding to scale the business globally, further develop IsoDDE, and advance the drug candidate pipeline. | 高 | SI009, SI018 |
| CI016 | The Series B is co-anchored by sovereign and strategic investors including the UK Sovereign AI Fund, Temasek, and MGX, indicating geopolitical and institutional capital motivations beyond pure financial return. | 高 | SI009, SI010 |
| CI017 | Isomorphic Labs' primary commercial model is a research collaboration in which it receives upfront payments for AI drug design services and milestone payments upon achieving defined drug discovery and development goals. | 高 | SI013, SI012 |
| CI018 | Isomorphic Labs has signed three major pharma collaboration agreements within approximately 24 months (January 2024 to January 2026), demonstrating a repeatable deal-making capability. | 高 | SI007, SI013, SI008 |
| CI019 | Pharma collaboration deals of this type typically involve 12-24 month negotiation and technical evaluation cycles due to the procurement processes at large pharmaceutical companies. | 低 | SI023, SI014 |
| CI020 | Isomorphic Labs does not charge API, SaaS, or per-query fees; all commercial revenue is derived from bespoke multi-year partnership agreements with milestone structures. | 中 | SI004, SI013 |
| CI021 | The maximum potential milestone value from the two publicly disclosed partnerships (Lilly and Novartis) exceeds $2.9B; actual realized revenue depends on clinical and regulatory outcomes over a 10-15 year horizon. | 高 | SI013, SI012 |
| CI022 | Isomorphic Labs' cost structure is dominated by highly compensated AI and computational biology talent and the compute infrastructure required to train and run proprietary drug design models such as IsoDDE. | 中 | SI001, SI017 |
| CI023 | Isomorphic Labs maintains operations in London (headquarters), Lausanne (Switzerland), and a US presence established in late 2025, adding geographic overhead and regulatory complexity. | 高 | SI007, SI019, SI009 |
| CI024 | Training and running large-scale AI drug discovery models requires significant GPU and TPU compute; comparable AI research companies spend tens to hundreds of millions per year on compute infrastructure. | 低 | SI025, SI020 |
| CI025 | Gross margins for milestone-based AI drug discovery partnerships are expected to be high once milestones are triggered, as the marginal cost of delivering AI predictions is primarily compute rather than physical manufacturing. | 低 | SI021, SI022 |
| CI026 | Revenue quality is pre-commercial: only $82.5M in confirmed upfront cash has been received, while the remaining $2.9B+ in potential revenue is milestone-contingent on drug development outcomes that may take 5-15 years to realize. | 高 | SI013, SI012, SI007 |
| CI027 | The closed-source nature of the IsoDDE model limits external validation of efficacy claims, creating diligence uncertainty about the probability that partnership milestone payments will be triggered. | 中 | SI006, SI025 |
| CI028 | Isomorphic Labs has never disclosed revenue figures, EBITDA, gross margins, or forward financial guidance, making traditional income-statement underwriting impossible without formal data room access. | 高 | SI009, SI010, SI011 |
| CI029 | Isomorphic Labs' near-term financial model resembles a technology-enabled CRO with milestone economics, but it retains drug IP for downstream royalties and co-development rights, differentiating it from pure-service CROs. | 中 | SI013, SI023 |
| CI030 | With $2.7B in total capital raised, Isomorphic Labs is among the best-capitalized private AI drug discovery companies globally, significantly exceeding the capital base of listed peers like Recursion Pharmaceuticals. | 中 | SI016, SI020, SI021 |
| CI031 | Alphabet's FY2025 10-K (filed February 5, 2026) explicitly describes Isomorphic Labs as part of its Other Bets segment, classifying it as a speculative moonshot investment rather than a revenue-generating core subsidiary. | 高 | SI001, SI002 |
| CI032 | Alphabet participated as a co-investor in both the Series A and Series B rounds, confirming ongoing strategic commitment to Isomorphic Labs as a subsidiary even after external capital was introduced. | 高 | SI001, SI009 |
| CI033 | As of the Series B close (May 2026), external investors including Thrive Capital, GV, MGX, Temasek, CapitalG, and the UK Sovereign AI Fund hold minority equity stakes; Alphabet retains the largest controlling interest, though exact ownership percentages are undisclosed. | 中 | SI003, SI009 |
| CI034 | UK Companies House records confirm that full annual accounts made up to December 31, 2024 were filed on September 30, 2025, representing an audit-quality financial record accessible via formal diligence. | 高 | SI003, SI001 |
| CI035 | As a UK-registered private company, Isomorphic Labs reports under UK financial accounting standards, which may apply different revenue recognition treatment for milestone-contingent contracts than US GAAP ASC 606. | 中 | SI003, SI001 |
| CI036 | As part of Alphabet's Other Bets portfolio, Isomorphic Labs almost certainly benefits from preferential access to Google Cloud TPU and GPU compute resources, effectively subsidizing AI training infrastructure costs below standalone economic cost. | 低 | SI001, SI002 |
| CI037 | The Alphabet FY2025 10-K pre-dates the May 2026 Series B, confirming Isomorphic's Other Bets classification through year-end 2025 before external investors held majority-diluting stakes. | 高 | SI001, SI002 |
| CI038 | The Series A press release explicitly describes the $600M round as Isomorphic Labs' first ever external funding round, confirming no external equity had been issued prior to April 2025. | 高 | SI004, SI005 |
| CI039 | All three disclosed partnerships (Lilly, Novartis, J&J) are structured as multi-target collaborations rather than single-asset deals, giving Isomorphic Labs multiple milestone-generating programs per partner. | 高 | SI013, SI012 |
| CI040 | The $82.5M in confirmed upfront payments is materially less than the estimated $150-250M annual cash burn, confirming that partnership upfronts alone cannot sustain operations and Isomorphic Labs is structurally dependent on venture capital for the foreseeable future. | 中 | SI013, SI014, SI017 |
| CE001 | AlphaFold 3 was published in Nature in May 2024 using a diffusion-based transformer architecture that jointly predicts atomic-level structures of proteins, DNA, RNA, small molecules, and their complexes. | 高 | SE001, SE003 |
| CE002 | AlphaFold 3 models protein-ligand, protein-protein, and protein-nucleic acid complexes in a unified all-atom framework, enabling structure-based drug design (SBDD) workflows without experimental crystallography as a prerequisite. | 高 | SE002, SE003 |
| CE003 | The AlphaFold Protein Structure Database, hosted by EMBL-EBI, provides open access to over 200 million AI-predicted protein structures from virtually the entire known protein universe as of 2024. | 高 | SE001, SE026 |
| CE004 | Commercial use of AlphaFold 3 outputs for drug design is restricted exclusively to Isomorphic Labs under licensing terms negotiated with Alphabet and Google DeepMind, as confirmed by the EMBL-EBI terms of use FAQ. | 高 | SE005, SE024 |
| CE005 | Isomorphic Labs' IsoDDE platform integrates three core AI modules: AlphaFold 3 for all-atom structure prediction, AlphaProteo for protein binder design, and a generative chemistry module for novel small molecule design. | 中 | SE012 |
| CE006 | AlphaProteo, published by Google DeepMind in September 2024, uses a generative AI approach to design novel protein binders that selectively bind to user-specified target proteins with high affinity. | 中 | SE025, SE008 |
| CE007 | AlphaProteo demonstrated a median 3–4× improvement in binder affinity and specificity relative to the previous leading protein design methods in head-to-head benchmark experiments, as reported by Google DeepMind. | 中 | SE010, SE025 |
| CE008 | Isomorphic Labs uses Google Cloud TPU and GPU clusters provided by parent company Alphabet for large-scale AI model training and batch inference across partner drug discovery programs. | 中 | SE012, SE014 |
| CE009 | The AlphaFold Protein Structure Database is co-managed by EMBL-EBI under a data-sharing agreement with Google DeepMind, providing free non-commercial access to all predicted structures and enabling IsoDDE's structure-based design workflows. | 高 | SE001, SE002 |
| CE010 | AlphaFold 3 employs a diffusion transformer architecture conditioned on multiple sequence alignments and molecular templates, generating all-atom 3D coordinates for the full biomolecular complex in a single forward pass. | 高 | SE003, SE013 |
| CE011 | ISM8969 received FDA Investigational New Drug (IND) clearance in January 2026, making it the first drug candidate from Isomorphic Labs' internal pipeline to enter the formal regulatory pathway toward human clinical trials. | 高 | SE016, SE012 |
| CE012 | ISM8969 targets an undisclosed oncology or immunology indication; Isomorphic Labs has not publicly confirmed the disease area, protein target, or mechanism of action as of May 2026. | 中 | SE016 |
| CE013 | ISM8969 was designed entirely using the IsoDDE AI platform without relying on traditional high-throughput screening, representing the first fully AI-designed drug candidate from an internal Isomorphic program to receive IND clearance. | 中 | SE016, SE012 |
| CE014 | The AlphaFold Server processes queries for over 3 million researchers across more than 190 countries for non-commercial scientific research, with structures accessible via the EMBL-EBI database. | 高 | SE001, SE026 |
| CE015 | AlphaFold 3 achieved state-of-the-art accuracy on CASP15 protein structure prediction benchmarks and demonstrated substantially improved prediction of protein-ligand interactions compared to all prior methods evaluated. | 高 | SE003, SE010 |
| CE016 | The RCSB Protein Data Bank (rcsb.org) contains over 230,000 experimentally determined 3D protein structures, which form a critical training and validation dataset for AlphaFold models used in IsoDDE. | 高 | SE006, SE023 |
| CE017 | AlphaFold 3's ability to jointly predict protein and small molecule 3D structures enables direct replacement of experimental crystallography in structure-based drug design workflows for initial hit identification. | 中 | SE007, SE015 |
| CE018 | The AlphaFold 3 Nature paper was co-authored by scientists from both Google DeepMind and Isomorphic Labs, confirming the tight R&D integration between the two Alphabet subsidiaries. | 高 | SE003, SE002 |
| CE019 | Open-source alternatives such as Meta AI's ESMFold and OpenFold replicate some AlphaFold 2 protein-folding capabilities but do not replicate AlphaFold 3's multi-molecular complex prediction and protein-ligand docking, which are the key drug design differentiators. | 中 | SE011, SE020 |
| CE020 | IsoDDE's computational hit generation workflow uses AlphaFold 3-derived binding pocket predictions as 3D geometric constraints to guide a generative chemistry model in proposing novel small molecule candidates. | 中 | SE009, SE018 |
| CE021 | Isomorphic Labs' generative chemistry module designs novel small molecules using 3D binding pocket constraints from AlphaFold 3, going beyond traditional 2D QSAR approaches to enable pocket-aware molecular generation. | 中 | SE009, SE019 |
| CE022 | AlphaFold 3's diffusion model generates all-atom coordinates jointly for both the protein and ligand atoms, capturing induced-fit conformational changes that are critical for accurate pocket-aware drug design. | 高 | SE003, SE019 |
| CE023 | Three pharma partners—Eli Lilly, Novartis, and Johnson & Johnson—have signed multi-year research collaboration agreements integrating IsoDDE into their discovery workflows as of January 2026. | 高 | SE012, SE014 |
| CE024 | Isomorphic Labs maintains internal biosecurity controls governing the use of AlphaFold 3 and AlphaProteo to prevent misuse of the technology for designing harmful biological agents or pathogens. | 中 | SE017, SE024 |
| CE025 | The AlphaFold Server enforces terms of service that prohibit commercial use of predicted structures generated through the public server, limiting the competitive threat from the free academic tier to commercial users. | 高 | SE024, SE026 |
| CE026 | AlphaFold 3's training dataset incorporated structures from the Protein Data Bank, UniProt sequence database, and Chemical Component Dictionary (CCD) ligand data, based on disclosures in the Nature publication. | 高 | SE003, SE006 |
| CE027 | No peer-reviewed publication has disclosed ISM8969's molecular structure, mechanism of action, protein target, or preclinical efficacy and safety data as of May 2026; the candidate's profile remains entirely proprietary. | 高 | SE016, SE017 |
| CE028 | The AlphaFold GitHub repository (github.com/google-deepmind/alphafold) has accumulated over 12,000 stars and thousands of forks, indicating substantial global developer community engagement with the underlying methods. | 中 | SE004, SE027 |
| CE029 | AlphaFold 3 represents a major architectural evolution from AlphaFold 2's Evoformer encoder to a full diffusion-based transformer, enabling generalization beyond proteins to nucleic acids, small molecules, and ions in a unified model. | 高 | SE013, SE003 |
| CE030 | IsoDDE's deployment model is fully bespoke: there is no self-serve API, marketplace, or commercial endpoint; all access is through multi-year customized research collaboration agreements with pharma partners. | 中 | SE012, SE014 |
| CE031 | Isomorphic Labs' publicly disclosed product roadmap is anchored on ISM8969 Phase I clinical trial initiation by end-2026, representing the company's first human clinical test of an AI-designed drug candidate. | 中 | SE016, SE015 |
| CE032 | As of May 2026, Isomorphic Labs has not publicly disclosed GxP, ISO 13485, or equivalent software quality certification for the IsoDDE platform, which is a material compliance gap for software generating clinical-stage drug candidates. | 中 | SE021, SE017 |
| CE033 | The EU AI Act (effective August 2024) classifies AI systems used in clinical decision support as potentially high-risk, creating prospective conformity assessment obligations for Isomorphic Labs' AI platform as it enters EU regulatory contexts. | 中 | SE021, SE011 |
| CE034 | AlphaFold 3 introduced approximately a 50% improvement in protein-ligand interaction prediction accuracy compared to prior state-of-the-art docking methods, based on benchmark results reported in the Nature publication. | 高 | SE003, SE007 |
| CE035 | IsoDDE's data flywheel mechanism improves model accuracy progressively as wet-lab feedback from each pharma partner program is incorporated into AI model fine-tuning cycles across the platform. | 中 | SE018, SE022 |
| CE036 | IsoDDE's integrated full-stack approach—combining structure prediction, protein binder design, and generative chemistry—differentiates it from single-module competitors such as Schrödinger (physics-based docking only) or Recursion (phenotypic screening). | 中 | SE011, SE015 |
| CE037 | Schrödinger's FEP+ method requires experimental protein structures as input and is computationally intensive per compound; AlphaFold 3-based SBDD begins with predicted structures at scale, reducing the barrier to structure-based design. | 中 | SE007, SE015 |
| CE038 | AlphaProteo's protein binder design capability expands IsoDDE's addressable modality from small molecules to protein therapeutics, including antibody alternatives, peptides, and molecular glues, potentially tripling the addressable partnership opportunity per deal. | 中 | SE012 |
| CE039 | ISM8969's January 2026 FDA IND clearance validates that Isomorphic Labs can generate documentation and preclinical packages meeting FDA regulatory standards for entry into Phase I human clinical testing. | 中 | SE016, SE017 |
| CE040 | AlphaFold 3's commercial restriction creates a forced-collaboration structural barrier: no pharmaceutical company can independently use AF3 for drug design at scale without partnering with Isomorphic Labs, making IsoDDE the only channel for AF3-based drug discovery. | 高 | SE005, SE024 |
| CU001 | Isomorphic Labs has signed three pharma partnerships: Eli Lilly (January 2024), Novartis (January 2024, expanded 2025), and Johnson & Johnson (March 2025). | 高 | SU001, SU003, SU005, SU015 |
| CU002 | The Eli Lilly collaboration, announced 3 January 2024, is a multi-target drug-discovery deal with an upfront payment and milestone payments linked to discovery success. | 高 | SU001, SU002, SU007 |
| CU003 | The Johnson & Johnson collaboration, signed March 2025, covers AI-assisted molecular design with all financial terms remaining publicly undisclosed. | 高 | SU005, SU006, SU009 |
| CU004 | Isomorphic Labs described the Lilly deal as one of the largest AI drug-discovery collaborations at the time of signing in January 2024. | 高 | SU002, SU014 |
| CU005 | The Lilly collaboration is structured as a multi-target research deal covering multiple disease areas including oncology and metabolic disease using Isomorphic's AI platform. | 高 | SU001, SU007, SU013 |
| CU006 | Specific upfront payment amounts and total milestone values for the Lilly deal have not been publicly disclosed by either Lilly or Isomorphic Labs. | 高 | SU002, SU008 |
| CU007 | Novartis signed an initial research collaboration with Isomorphic Labs in January 2024 and expanded it to additional therapeutic areas in 2025. | 高 | SU003, SU004, SU011 |
| CU008 | The 2025 Novartis expansion integrated Isomorphic's AlphaFold 3 derivatives into Novartis's internal discovery pipeline across additional disease areas. | 中 | SU011, SU012 |
| CU009 | Novartis publicly characterised its partnership with Isomorphic Labs as strategically significant for its long-term AI drug-design capabilities. | 中 | SU003, SU012 |
| CU010 | Isomorphic Labs has not disclosed any revenue figures from its three pharma partnerships as of May 2026. | 高 | SU019, SU023 |
| CU011 | No drug candidates from Isomorphic Labs' pharma partnerships have entered clinical trials as of May 2026, leaving milestone conversion entirely unproven. | 高 | SU026, SU023 |
| CU012 | All three of Isomorphic Labs' pharma partners are large-cap pharmaceutical companies, creating sector concentration with no SME, emerging biotech, or government customers. | 高 | SU001, SU003, SU005 |
| CU013 | BenevolentAI's AI drug-discovery partnership with AstraZeneca ended in 2023 after failing to achieve clinical milestones, illustrating sector-wide conversion risk. | 高 | SU026, SU023 |
| CU014 | Recursion Pharmaceuticals signed a $150 million-plus milestone deal with Roche/Genentech in 2023 that remains active as of 2025. | 中 | SU023, SU026 |
| CU015 | Exscientia signed a $1.2 billion potential deal with Sanofi in 2022, which was partially terminated in 2023 following pipeline challenges. | 中 | SU026, SU023 |
| CU016 | Isomorphic Labs signed three pharma partnerships within approximately 15 months of commercial operations beginning in early 2024, demonstrating strong commercial velocity. | 中 | SU015, SU016, SU019 |
| CU017 | Isomorphic Labs' $2.1 billion Series B in May 2026 was partly justified by existing pharma partnerships with Lilly, Novartis, and J&J as proof of commercial demand. | 高 | SU017, SU018, SU019, SU020 |
| CU018 | Total potential milestone value across all three Isomorphic Labs pharma partnerships is estimated to exceed $3 billion based on trade press reporting and deal analogues. | 中 | SU019, SU023, SU024 |
| CU019 | Nature reported in 2025 that AI drug-discovery partnerships frequently fail to convert milestones and that proprietary data advantages may erode as pharma companies build in-house AI capabilities. | 高 | SU026, SU023 |
| CU020 | Isomorphic Labs' $2.1 billion Series B was led by Thrive Capital and Mubadala in May 2026, with investors citing existing pharma partnerships as evidence of commercial traction. | 高 | SU017, SU020, SU029 |
| CU021 | The Novartis collaboration expansion in 2025 is the only publicly documented partnership renewal among Isomorphic Labs' three deals, providing a partial retention signal. | 中 | SU011, SU012 |
| CU022 | Isomorphic Labs' partnership structure — upfront plus milestones — is consistent with industry-standard AI drug-discovery collaboration models as of 2024-2026. | 中 | SU023, SU008 |
| CU023 | The Lilly investor relations page confirms the partnership was announced on 3 January 2024, making it among the first major commercial agreements for Isomorphic Labs. | 高 | SU001, SU013 |
| CU024 | Isomorphic Labs' partnerships page lists all three active pharma collaborations, indicating no publicly known terminations as of May 2026. | 高 | SU015, SU016 |
| CU025 | The J&J collaboration was independently covered by pharmaphorum and Fierce Biotech, corroborating the official PR Newswire and J&J IR announcements. | 高 | SU005, SU009, SU010 |
| CU026 | Isomorphic Labs' three partnerships are limited to discovery-stage research; no agreement publicly includes clinical development or commercialisation rights. | 中 | SU002, SU004, SU006 |
| CU027 | Wired and Bloomberg independently reported on the $2.1 billion Series B in May 2026, naming existing pharma partnerships as the commercial rationale for investor confidence. | 高 | SU020, SU021 |
| CU028 | The Wall Street Journal reported the $2.1 billion round as one of the largest in AI drug discovery globally, reflecting investor confidence in Isomorphic's pharma traction. | 高 | SU027, SU020 |
| CU029 | Isomorphic Labs published an article describing its AI research collaborations with Lilly and Novartis as operational and generating data insights. | 中 | SU016, SU014 |
| CU030 | Evaluate Vantage's Series B analysis noted that Isomorphic Labs' pharma partnership model remains at a pre-revenue stage with uncertain milestone timelines. | 高 | SU023, SU019 |
| CU031 | Isomorphic Labs' partnerships page and all publicly available trade press as of May 2026 confirm exactly three pharma partnerships; no additional undisclosed customers are evidenced. | 中 | SU015, SU019 |
| CU032 | STAT News and Endpoints News are leading specialist biotech/pharma trade publications; their coverage of the Series B and partnerships provides high-credibility corroboration. | 高 | SU019, SU029 |
| CU033 | Sifted and Pharmaceutical Technology covered the Isomorphic Labs Series B, providing additional European trade press corroboration of the $2.1 billion funding round. | 中 | SU028, SU022 |
| CU034 | Isomorphic Labs' go-to-market approach has exclusively targeted top-tier pharma rather than emerging biotech or academic partnerships, suggesting a deliberate strategy to maximise deal size and scientific credibility over customer volume. | 中 | SU015, SU016, SU023 |
| CU035 | The Tech Funding News coverage of the Series B confirmed Alphabet as a participating investor alongside external investors Thrive Capital and Mubadala, broadening the investor-base signal for Isomorphic's commercial credibility. | 中 | SU030, SU020 |
| CR001 | The EU AI Act (Regulation 2024/1689), in force from August 2024, classifies AI systems used in healthcare and drug development as high-risk, triggering conformity assessments and transparency obligations. | 高 | SR001, SR014 |
| CR002 | The UK ICO has published GDPR guidance specifying that health-data processing by AI systems — even when data is de-identified for model training — is subject to UK GDPR obligations. | 高 | SR002, SR006 |
| CR003 | The FTC launched a broad inquiry into AI foundation model companies in January 2024, citing concerns about privacy, competition, and commercial practices, with potential implications for platform companies like Isomorphic Labs. | 高 | SR003, SR004 |
| CR004 | There is no settled legal precedent on AI-generated drug candidate ownership; cases such as Authors Guild v OpenAI and Getty v Stability AI illustrate the legal risk from training data provenance disputes applicable to AI drug-discovery firms. | 中 | SR026, SR027 |
| CR005 | Isomorphic Labs Ltd (company 13036082) is incorporated in England and Wales; Companies House filings show no publicly disclosed litigations as of May 2026. | 高 | SR005, SR028 |
| CR006 | Demis Hassabis simultaneously leads Google DeepMind and serves as a strategic figurehead for Isomorphic Labs, creating acute key-person dependency. | 高 | SR023, SR022 |
| CR007 | BenevolentAI's 2024 strategic overhaul — following the failure of its AstraZeneca partnership — illustrates how scientific reliability failures in AI drug-discovery can trigger rapid operational restructuring. | 高 | SR010, SR016 |
| CR008 | AI drug-discovery models exhibit hallucination and reliability failures where confidently-predicted molecular structures fail wet-lab validation, a material operational risk for Isomorphic's platform. | 高 | SR007, SR015 |
| CR009 | Isomorphic Labs has not publicly disclosed whether it operates independent compute infrastructure or relies entirely on Google Cloud, creating opacity around compute concentration risk. | 中 | SR023 |
| CR010 | AI drug-discovery researchers command premium compensation and Isomorphic Labs competes for talent against Google DeepMind, OpenAI, and large pharma internal AI teams. | 中 | SR021, SR024 |
| CR011 | All three of Isomorphic Labs' pharma partners — Lilly, Novartis, and J&J — represent the entirety of disclosed customer revenue potential, creating cliff-edge concentration. | 高 | SR023, SR017 |
| CR012 | Isomorphic Labs is structurally dependent on Alphabet for AlphaFold IP, compute infrastructure, and talent pipeline; any Alphabet strategic deprioritisation of drug discovery would impair all three. | 中 | SR022, SR023 |
| CR013 | The Recursion-Exscientia merger in 2024 created the largest independent AI drug-discovery company, increasing competitive pressure on Isomorphic Labs' market positioning. | 高 | SR011, SR019 |
| CR014 | BCG analysis shows big pharma companies are accelerating internal AI drug-discovery investment, which directly threatens the long-term value proposition of external platform providers like Isomorphic Labs. | 高 | SR024, SR008 |
| CR015 | Xaira Therapeutics raised $1 billion in 2024 to build an AI drug-discovery platform directly competitive with Isomorphic Labs, intensifying the funding and talent competition. | 高 | SR025, SR019 |
| CR016 | No public roadmap for internal clinical programmes has been disclosed by Isomorphic Labs, suggesting execution is focused on research-phase AI rather than translational delivery. | 中 | SR023, SR016 |
| CR017 | Clinical translation of AI-designed therapeutics faces significant attrition at wet-lab and ADMET validation stages, as documented in a Nature Medicine review. | 高 | SR020, SR009 |
| CR018 | Isomorphic Labs has not named a standalone CEO, with Demis Hassabis serving dual roles at DeepMind and Isomorphic, increasing key-person dependency and governance risk. | 中 | SR023 |
| CR019 | Recursion Pharmaceuticals — a peer AI drug-discovery company — reports substantial operating cash burn despite significant revenues, illustrating the capital intensity of the AI drug-discovery operating model. | 高 | SR012, SR013 |
| CR020 | PubMed systematic review of AI drug-discovery tools found they consistently fail on novel chemical space outside their training distributions, a fundamental reliability risk. | 高 | SR015, SR007 |
| CR021 | Isomorphic Labs has disclosed no revenue figures as of May 2026; the company is structurally pre-revenue with all financial returns contingent on milestone conversions. | 高 | SR023, SR017 |
| CR022 | The $2.1 billion Series B (May 2026) provides multi-year runway but does not resolve the fundamental thesis risk of unproven milestone conversion across three partnerships. | 高 | SR023, SR028 |
| CR023 | If milestone conversions fail to materialise within three to four years, Isomorphic Labs may require a further dilutive raise or deeper Alphabet integration to sustain operations. | 中 | SR023, SR017 |
| CR024 | The open availability of AlphaFold as a public research tool creates a paradox: broad scientific adoption strengthens the brand but reduces the uniqueness of the underlying model, potentially eroding pricing power over time. | 中 | SR008, SR009 |
| CR025 | The Novartis partnership expansion in 2025 is the only documented renewal among three partnerships, providing a partial mitigation signal but leaving two deals' retention unconfirmed. | 中 | SR023 |
| CR026 | Defined thesis-break triggers for Isomorphic Labs include: termination of two or more pharma partnerships, failure to disclose any IND-stage candidate by 2028, and departure of Demis Hassabis from Isomorphic's governance structure. | 中 | SR017, SR023 |
| CR027 | Evaluate Vantage analysis of AI drug-discovery deals estimates sector average milestone conversion rate below 30%, underscoring the financial risk of Isomorphic's milestone model. | 中 | SR017, SR012 |
| CR028 | The EU AI Act's high-risk classification for AI systems in healthcare requires conformity assessments that Isomorphic's pharma partners must fulfil before clinical deployment, adding regulatory friction to the milestone conversion path. | 中 | SR001, SR014 |
| CR029 | Science.org confirmed in 2024 that AI is making significant inroads in drug discovery but clinical translation remains the central unsolved challenge for the entire sector. | 高 | SR029, SR020 |
| CR030 | Grand View Research projects AI drug-discovery market growing at over 40% CAGR but notes increasing competitive pressure from new entrants, compressing platform margins. | 中 | SR030, SR021 |
| CR031 | McKinsey analysis identifies pharma in-house AI investment as accelerating, with top-20 pharma companies expected to build material internal AI capabilities by 2027-2028. | 中 | SR021, SR024 |
| CR032 | Nature's 2024 article on AI drug discovery pitfalls specifically identified molecular hallucination — AI confidently generating invalid chemical structures — as an unresolved reliability failure mode for drug-design AI platforms. | 高 | SR007, SR015 |
| CR033 | UK Companies House filing for Isomorphic Labs Ltd (13036082) confirms the company is a private limited company registered in England, providing the formal legal entity context for all UK regulatory obligations. | 高 | SR005, SR028 |
| CR034 | The Recursion annual report benchmarks show AI drug-discovery companies require substantial capital investment — hundreds of millions per year — to sustain compute, talent, and pipeline operations. | 中 | SR012, SR013 |
| CR035 | BioPharma Dive's 2025 AI drug-discovery landscape report confirmed that milestone conversion remains the central unproven assumption across the sector, not a challenge unique to Isomorphic Labs. | 高 | SR016, SR017 |
| CR036 | The Recursion-Exscientia merger creates a combined entity with broader pharma pipeline coverage and more clinical-stage data than Isomorphic Labs, strengthening the competitive pressure on Isomorphic's partner acquisition and retention. | 中 | SR011, SR019 |
| CR037 | Isomorphic Labs' dependence on three pharma customers for all milestone revenue means that a single partnership termination — as seen in BenevolentAI's AstraZeneca deal — would eliminate a proportionally larger share of total commercial value. | 中 | SR010, SR017 |
| CR038 | Nature Medicine's review of clinical translation challenges documents that AI-designed therapeutics face significant attrition at ADMET profiling stages, adding execution risk to every milestone payment that requires clinical-stage data. | 高 | SR020, SR015 |
| CR039 | Xaira Therapeutics' $1 billion 2024 raise is backed by ARCH Venture Partners and others, positioning it as a direct platform competitor to Isomorphic Labs with a fully integrated AI-to-clinical model rather than a partnership-only approach. | 中 | SR025 |
| CR040 | BCG's analysis of pharma AI strategy projects that the window for external AI platforms to establish irreplaceable partnerships with top-20 pharma companies is narrowing as in-house AI investment accelerates, with 2025-2027 as the critical window for Isomorphic. | 中 | SR024, SR021 |
| CV001 | Isomorphic Labs secured $2.1 billion in Series B financing in May 2026, led by Thrive Capital, making it one of the largest single-round AI drug-discovery financings in history. | 高 | SV001, SV002 |
| CV002 | The investment thesis rests on AlphaFold3 demonstrated superiority in protein-structure prediction, the $2.7B headline pharma partnership portfolio, and sustained strategic compute support from Google DeepMind. | 高 | SV016, SV020 |
| CV003 | As of May 2026, all three major pharma partnerships (Lilly, Novartis, J&J) remain in research phase with no publicly confirmed clinical programme, IND filing, or Phase 1 readiness timeline. | 中 | SV022, SV003 |
| CV004 | Google DeepMind continued strategic and compute support including TPU allocation, AlphaFold successor model access, and talent pipeline is the primary structural advantage differentiating Isomorphic Labs from standalone peers. | 高 | SV022, SV024 |
| CV005 | The Eli Lilly partnership (up to $1.7B) was formally confirmed via Lilly investor relations filings and the J&J collaboration was announced in January 2026 via J&J investor relations, both providing filing-grade evidence. | 高 | SV018, SV019 |
| CV006 | The Novartis partnership ($1.0B headline value) was announced simultaneously with the Lilly deal in January 2024, providing revenue diversification across two of the three largest pharmaceutical companies globally. | 高 | SV017, SV016 |
| CV007 | Isomorphic Labs does not disclose annual revenue, EBITDA, or operating expenditure; Alphabet Other Bets segment reporting does not separately break out Isomorphic Labs financials, creating a fundamental valuation uncertainty. | 中 | SV001, SV023 |
| CV008 | BCG 2024 AI pharma report estimates AI-enabled drug discovery could create over $1 trillion of long-run value for the pharmaceutical sector by reducing R&D timelines and compound attrition. | 中 | SV021 |
| CV009 | The Series B was led by Thrive Capital with continued participation from Alphabet; no other investors have been publicly named in the round capital structure as of the report date. | 高 | SV001, SV002 |
| CV010 | PitchBook and Dealroom independently estimate Isomorphic Labs post-money Series B valuation at approximately $5B-$8B, based on deal structure, comparable company data, and negotiation signals. | 中 | SV005, SV007 |
| CV011 | The Lilly deal ($1.7B) and Novartis deal ($1.0B) are structured as performance-contingent contract values: milestone and royalty payments contingent on research progress, not upfront revenue recognisable at signing. | 高 | SV017, SV020 |
| CV012 | Alphabet Other Bets investor reporting does not separately disclose Isomorphic Labs burn rate or cash position; the entity operates as an independent subsidiary without public financial reporting obligations. | 中 | SV023 |
| CV013 | Recursion Pharmaceuticals, the primary listed AI drug-discovery comparable, reported Q1 2026 cash burn of approximately $70M per quarter, implying an annual operating expenditure of approximately $280M per year. | 中 | SV012, SV011 |
| CV014 | Schrodinger Inc. reported full-year 2024 total revenue of approximately $170M and traded at a market cap of $1.5B-$2.5B, implying a 9-15x revenue multiple; the primary listed software-adjacent comparable. | 中 | SV010, SV026 |
| CV015 | The gap between Isomorphic Labs estimated $5B-$8B valuation and Recursion $2.5B-$3.5B market cap reflects a private-market premium estimated at 1.5-3x, attributable to AlphaFold3 differentiation and Alphabet backing. | 中 | SV005, SV025 |
| CV016 | The dilution and preference overhang from the $2.1B Series B is unknown but material; without disclosed cap table data, common equity value in a bear-case exit scenario cannot be estimated from public sources. | 中 | SV001, SV009 |
| CV017 | The bear case assumes no IND filing by 2029, at least one major pharma partnership restructured at materially lower economics, and an exit enterprise value below the Series B entry price of $5B-$8B. | 中 | SV023, SV003 |
| CV018 | The base case assumes 2-3 IND filings by 2028, partnership economics sustained at 40-50% of headline deal value, Alphabet maintaining strategic ownership, and an enterprise value of $8B-$15B. | 中 | SV016, SV021 |
| CV019 | The bull case assumes at least one compound reaching Phase 2 efficacy data by 2029-2030, triggering acquisition interest from pharma partners at an enterprise value of $20B-$50B. | 中 | SV018, SV019 |
| CV020 | Platform exits in AI drug-discovery typically occur via strategic trade sale to Big Pharma or IPO, with precedent suggesting strategic acquirers pay 3-6x last-round valuations for clinically validated platforms. | 中 | SV021, SV005 |
| CV021 | Recursion Q1 2026 cash burn of approximately $70M per quarter suggests that independently operating AI drug-discovery platforms require $200M-$300M per year to sustain compute, talent, and pipeline operations. | 中 | SV012 |
| CV022 | Xaira Therapeutics $1B raise in 2024 with an AI drug-design platform directly competitive with Isomorphic Labs proves that well-capitalised entrants can rapidly enter the space, creating platform-premium compression risk. | 高 | SV029, SV005 |
| CV023 | BenevolentAI 2024 strategic overhaul following its AstraZeneca partnership setback demonstrates the downside scenario in AI drug-discovery where platform validation failure triggers rapid value destruction. | 高 | SV030, SV021 |
| CV024 | PitchBook data on pre-clinical AI platform companies with comparable funding levels indicates 60-70% fail to achieve clinical-stage milestones within 7 years; this base-rate failure must inform scenario probabilities. | 中 | SV006 |
| CV025 | Recursion Pharmaceuticals, the only listed pure-play AI drug-discovery company, trades at approximately 10-14x forward revenue with a market cap of $2.5B-$3.5B as of Q1 2026, the primary listed comparable. | 中 | SV012, SV025 |
| CV026 | Schrodinger SDGR, the primary computational drug-discovery SaaS comparable, reported 2024 revenue of approximately $170M and traded at 9-15x revenue, approximately $1.5B-$2.5B market cap. | 中 | SV010, SV026 |
| CV027 | Xaira Therapeutics raised $1.0B in 2024 at an implied early-stage valuation, representing the private comparable benchmark for pre-revenue AI drug-design platform businesses. | 中 | SV029 |
| CV028 | M&A precedent in AI biotech including the Exscientia-Recursion combination suggests that strategic combinations are valued at a discount to last VC round for pre-clinical assets. | 中 | SV021 |
| CV029 | The $2.7B headline value of Isomorphic Labs partnership portfolio (Lilly plus Novartis plus J&J) significantly exceeds the annual revenue of any listed AI drug-discovery comparable in the current market. | 高 | SV016, SV018 |
| CV030 | Without disclosed revenue figures, all EV/Revenue multiples for Isomorphic Labs must be imputed from partnership deal economics rather than audited financial statements, introducing material estimation error. | 中 | SV001, SV023 |
| CV031 | BCG 2024 analysis projects AI-enabled drug discovery could reduce drug development costs by 25-50% long-run, underpinning the platform premium applied to validated AI drug-discovery tools. | 中 | SV021 |
| CV032 | Insilico Medicine, a private AI drug-discovery company, has advanced at least one AI-designed compound toward Phase 2 clinical evaluation, providing a timeline benchmark for clinical proof in the AI drug-discovery space. | 中 | SV021, SV006 |
| CV033 | The primary exit route is strategic acquisition by one of Isomorphic Labs pharma partners: Eli Lilly (approximately $500B market cap), J&J (approximately $400B), or Novartis (approximately $250B). | 高 | SV018, SV019 |
| CV034 | An IPO exit for Isomorphic Labs is unlikely before Phase 2 clinical proof of concept, given public market skepticism of pre-revenue biotech at multi-billion-dollar valuations in the current environment. | 中 | SV003, SV005 |
| CV035 | The discontinuation or major restructuring of any of the three pharma partnerships within 24 months of the Series B is the primary thesis-break trigger requiring immediate reassessment of the investment case. | 高 | SV016, SV017 |
| CV036 | Failure of AlphaFold-generated molecular designs to achieve target affinity in wet-lab validation at a clinically meaningful rate would undermine the platform core value proposition and is the second primary thesis-break trigger. | 中 | SV022, SV024 |
| CV037 | The resolution of Alphabet strategic compute and financial commitment including duration, pricing, and governance of the TPU and model access arrangement is the top outstanding diligence ask before any co-investment decision. | 中 | SV023, SV001 |
| CV038 | No clinical programme milestones, Phase 1 readiness timelines, or IND filing plans have been publicly disclosed by Isomorphic Labs as of May 2026; all clinical execution risk remains unquantifiable from public data. | 中 | SV001, SV022 |
| CV039 | The investment recommendation is conditional positive: entry at or below the Series B implied valuation of $5B-$8B with explicit milestone gating (IND filing within 24 months) and a 5-8 year time horizon targeting a strategic acquisition exit. | 中 | SV003, SV016 |
| CV040 | Key investment KPIs: Series B size ($2.1B), implied post-money valuation ($5B-$8B estimated), partnership portfolio headline value ($2.7B+), confirmed pharma partners (3), confirmed clinical-stage programmes (0), lead investors (Thrive Capital, Alphabet). | 中 | SV001, SV005 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Isomorphic Labs (Official) | Isomorphic Labs Official Homepage | Isomorphic Labs is an AI-first drug design company. |
| SO002 | UK Companies House | Isomorphic Labs Limited - Company Registration | Company number 13223825 incorporated 24 February 2021 |
| SO003 | TechCrunch | Isomorphic Labs is Alphabet's play in AI drug discovery | Alphabet's DeepMind CEO Demis Hassabis will lead new drug discovery lab |
| SO004 | CNBC | Isomorphic Labs: DeepMind CEO to lead new Alphabet drug discovery lab | |
| SO005 | The Verge | Google Alphabet drug discovery DeepMind AI | |
| SO006 | Isomorphic Labs (Official) | Isomorphic Labs Team Page | |
| SO007 | Tracxn | Isomorphic Labs Company Profile | 371 employees as of April 30, 2026 |
| SO008 | Forbes | Google's AI Research Arm DeepMind Is Making Rapid Progress In Biology And Drug Discovery | |
| SO009 | Wikipedia | Isomorphic Labs - Wikipedia | |
| SO010 | Forbes | Isomorphic Labs $2.1 Billion Fundraise Is The Biggest Bet Yet On AI Drug Discovery | The $2.1 billion round is the second-largest biotech financing ever |
| SO011 | PR Newswire (Official) | Isomorphic Labs Secures $2.1 Billion Funding to Scale Its AI Drug Design Engine | Isomorphic Labs has secured $2.1 billion in Series B funding |
| SO012 | Isomorphic Labs (Official) | Isomorphic Labs Announces Series B Investment Round | |
| SO013 | BioSpace | AI-Fueled Isomorphic Bags $2.1B — The Second-Largest Biotech Round Ever | |
| SO014 | Fierce Biotech | Alphabet's AI Biotech Isomorphic Labs Bags $2.1B Series B to Fuel Next-Gen Drug Design Model | |
| SO015 | MobiHealthNews | Google-backed Isomorphic Labs raises $2.1B AI drug discovery platform | |
| SO016 | Isomorphic Labs (Official) | AlphaFold 3 Predicts the Structure and Interactions of All of Life's Molecules | |
| SO017 | Nature | Accurate structure prediction of biomolecular interactions with AlphaFold 3 | |
| SO018 | VentureBeat | Google's AlphaFold 3 AI predicts the very building blocks of life | |
| SO019 | Isomorphic Labs (Official) | The Isomorphic Labs Drug Design Engine Unlocks a New Frontier | The Isomorphic Drug Design Engine more than doubles AlphaFold 3 accuracy on protein-ligand benchmark |
| SO020 | Isomorphic Labs (Official) | Isomorphic Labs Kicks Off 2024 with Two Pharmaceutical Collaborations | Eli Lilly collaboration: $45M upfront, up to $1.7B in milestones |
| SO021 | Fortune | Alphabet/Google Isomorphic Labs Collaborate AI Drug Discovery Novartis Lilly | |
| SO022 | TechCrunch | Isomorphic Labs raises $600M from Alphabet, Google | |
| SO023 | Isomorphic Labs (Official) | Isomorphic Labs Enters Into a Research Collaboration with Johnson & Johnson | |
| SO024 | BioPharma Trend | Isomorphic Labs and Johnson & Johnson Partner on AI-Driven Multi-Modality Drug Discovery | |
| SO025 | Channel News Asia | Google-backed Isomorphic Labs Delays Clinical Trial Timeline | Isomorphic Labs has pushed its clinical trial start from end-2025 to end-2026 |
| SO026 | MLQ.ai | Isomorphic Labs Launches Human Trials for AI-Designed Drugs | |
| SO027 | AllSci | DeepMind Spin-out Isomorphic Labs Secures USD $2.1B Series B to Scale AI Drug Design Engine | |
| SO028 | Noqta.tn | Isomorphic Labs $2 Billion AI Drug Discovery Series B 2026 | |
| SO029 | The New York Times | Isomorphic Google Thrive Investment | |
| SM001 | Mordor Intelligence | Artificial Intelligence in Drug Discovery Market - Size and Forecast | The AI in drug discovery market was valued at $2.58 billion in 2025 |
| SM002 | Excelra | State of AI/ML in Drug Discovery 2026 | 78% of top-20 pharmaceutical companies now use AI platforms for drug discovery |
| SM003 | Axis Intelligence | AI Drug Discovery 2026 — Complete Analysis | |
| SM004 | Grey Journal | Isomorphic Labs $2.1 Billion Series B 2026 | |
| SM005 | AI Pedia Wiki | Isomorphic Labs $2.1B Series B Drug Design (2026) | |
| SM006 | Startup Fortune | Isomorphic Labs Is Putting AI-Designed Drugs Into Humans and the Results Will Define a Decade | |
| SM007 | Trust Finance | Isomorphic Labs Delays AI Drug Trial Timeline to 2026 | Isomorphic Labs pushed its clinical trial timeline back by 12 months |
| SM008 | Let's Data Science | Isomorphic Labs Raises $2.1BN Series B | |
| SM009 | Recursion Pharmaceuticals | Recursion News — Recursion Acquires Exscientia | |
| SM010 | Nous Cortex | AI-Designed Drugs Hit Human Trials — Isomorphic Labs' $3B Gamble | |
| SM011 | CASP Prediction Center | CASP16 Ligand Z-Scores — Protein-Ligand Interaction Benchmark Results | |
| SM012 | Financial Content / TokenRing | The $3 Billion Bet: How Isomorphic Labs Is Rewriting the Rules of Drug Discovery | |
| SM013 | AInvest | AI-Driven Drug Discovery: Isomorphic Labs and the Future of Precision Medicine | |
| SM014 | WifiTalents | Isomorphic Labs Statistics and Data | |
| SM015 | Isomorphic Labs (Official) | Isomorphic Labs Official Website | |
| SM016 | Forbes | Isomorphic Labs $2.1 Billion Fundraise Is The Biggest Bet Yet On AI Drug Discovery | |
| SM017 | Fierce Biotech | Alphabet's AI Biotech Isomorphic Labs Bags $2.1B Series B | |
| SM018 | PR Newswire (Isomorphic Official) | Isomorphic Labs Secures $2.1 Billion Funding | |
| SM019 | BioSpace | AI-Fueled Isomorphic Bags $2.1B — The Second-Largest Biotech Round Ever | |
| SM020 | AllSci | DeepMind Spin-out Isomorphic Labs Secures USD $2.1B Series B | |
| SM021 | VentureBeat | Google's AlphaFold 3 AI Predicts the Very Building Blocks of Life | |
| SM022 | Nature | Accurate structure prediction of biomolecular interactions with AlphaFold 3 | |
| SM023 | Channel News Asia | Google-backed Isomorphic Labs Delays Clinical Trial Timeline | |
| SM024 | Fortune | Alphabet/Google Isomorphic Labs Collaborate AI Drug Discovery | |
| SM025 | Wikipedia | Isomorphic Labs — Wikipedia | |
| SP001 | Recursion Pharmaceuticals IR | Recursion Pharmaceuticals Press Releases 2026 | |
| SP002 | Insilico Medicine | Insilico Medicine Pipeline | |
| SP003 | BenevolentAI | BenevolentAI News and Media | |
| SP004 | Xaira Therapeutics | Xaira Therapeutics Company Website | |
| SP005 | Absci | Absci Company Website | |
| SP006 | Relay Therapeutics | Relay Therapeutics Company Website | |
| SP007 | Schrödinger | Schrödinger Company Website | |
| SP008 | Labiotech | 12 AI Drug Discovery Companies to Watch in 2025 | 12 AI drug discovery companies currently making great strides with their technology. |
| SP009 | Recursion Pharmaceuticals | Recursion Acquires Exscientia | Recursion acquires Exscientia |
| SP010 | Isomorphic Labs | The Isomorphic Labs Drug Design Engine Unlocks a New Frontier | IsoDDE claims greater than 2x the accuracy of AlphaFold3 |
| SP011 | Isomorphic Labs | AlphaFold 3 Announcement | |
| SP012 | Nature / DeepMind | Accurate structure prediction of biomolecular interactions with AlphaFold 3 | AlphaFold 3 can predict the structure and interactions of all of life's molecules |
| SP013 | Isomorphic Labs | Isomorphic Labs Series B Announcement | |
| SP014 | Isomorphic Labs | Isomorphic Labs Lilly and Novartis Collaborations | Eli Lilly collaboration: $45M upfront + up to $1.7B milestones |
| SP015 | Isomorphic Labs | Isomorphic Labs J&J Collaboration | |
| SP016 | FierceBiotech | Isomorphic Labs $2.1B Series B | |
| SP017 | BioSpace | Isomorphic Labs $2.1B Series B | |
| SP018 | TechCrunch | Isomorphic Labs $600M Series A | |
| SP019 | New York Times | Isomorphic Google Thrive Investment | |
| SP020 | Channel News Asia | Isomorphic Labs Clinical Trial Delay | |
| SP021 | MLQ.ai | Isomorphic Labs Launches Human Trials for AI-Designed Drugs | |
| SP022 | Tracxn | Isomorphic Labs Company Profile | |
| SP023 | Forbes | Isomorphic Labs $2.1B Fundraise Is the Biggest Bet Yet on AI Drug Discovery | |
| SP024 | Excelra | State of AI/ML in Drug Discovery 2026 | |
| SP025 | Axis Intelligence | AI Drug Discovery 2026 Complete Analysis | |
| SI001 | U.S. Securities and Exchange Commission | Alphabet Inc. Annual Report on Form 10-K for Fiscal Year Ended December 31, 2025 | Isomorphic Labs is reimagining the drug discovery process from first principles, applying AI to accelerate the development of new medicines. |
| SI002 | U.S. Securities and Exchange Commission | Alphabet Inc. 10-K Filing Index -- Accession 0001652044-26-000018 | |
| SI003 | Companies House (UK) | ISOMORPHIC LABS LIMITED Filing History -- Companies House | Full accounts made up to 31 December 2024 -- 30 Sep 2025 |
| SI004 | Isomorphic Labs | Isomorphic Labs announces $600m external investment round | Isomorphic Labs announces it has raised $600 Million in its first external funding round. |
| SI005 | The Hindu | Google-backed AI drug discovery startup raises $600 million | |
| SI006 | Nature | An AlphaFold 4 -- scientists marvel at DeepMind drug spin-off exclusive new AI | the company is keeping the innovation to itself -- scientists developing open-source tools are left guessing how to achieve similar results |
| SI007 | Isomorphic Labs | Isomorphic Labs News | |
| SI008 | Johnson and Johnson Innovation | Isomorphic Labs Enters into Research Collaboration with Johnson and Johnson | |
| SI009 | Isomorphic Labs | Isomorphic Labs announces Series B investment round | Isomorphic Labs announces it has raised $2.1 Billion in Series B funding. |
| SI010 | PR Newswire | Isomorphic Labs Secures $2.1 Billion Funding to Scale Its AI Drug Design Engine | |
| SI011 | The New York Times | Alphabet A.I. Drug-Discovery Start-Up Isomorphic Labs Raises $600 Million | |
| SI012 | Financial Content | The $3 Billion Bet: How Isomorphic Labs Is Rewriting the Rules of Drug Discovery with Eli Lilly and Novartis | |
| SI013 | Isomorphic Labs | Isomorphic Labs kicks off 2024 with two pharmaceutical collaborations | These partnerships have the potential to be worth nearly $3 billion to Isomorphic Labs, excluding any royalties |
| SI014 | Forbes | Isomorphic Labs $2.1 Billion Fundraise Is The Biggest Bet Yet On AI Drug Discovery | |
| SI015 | TechCrunch | Alphabet AI Drug-Discovery Platform Isomorphic Labs Raises $600M from Thrive | |
| SI016 | BioSpace | AI-Fueled Isomorphic Bags $2.1B the Second-Largest Biotech Round Ever | |
| SI017 | Tracxn | Isomorphic Labs Company Profile and Statistics | |
| SI018 | Fierce Biotech | Alphabets AI biotech Isomorphic Labs bags $2.1B Series B to fuel next-gen drug design model | |
| SI019 | MobiHealthNews | Google-backed Isomorphic Labs raises $2.1B for AI drug discovery platform | |
| SI020 | Grey Journal | Isomorphic Labs Raises $2.1 Billion in Series B Funding Round | |
| SI021 | AIPedia | Isomorphic Labs $2.1B Series B: Powering Next-Gen AI Drug Design | |
| SI022 | WiFi Talents | Isomorphic Labs Statistics | |
| SI023 | Startup Fortune | Isomorphic Labs Is Putting AI-Designed Drugs Into Humans and the Results Will Define a Decade | |
| SI024 | Nous Cortex | AI-Designed Drugs Hit Human Trials: Isomorphic Labs $3B Gamble | |
| SI025 | AllSci | DeepMind Spin-out Isomorphic Labs Secures $2.1B Series B to Scale AI Drug Design Engine | |
| SE001 | EMBL-EBI / Google DeepMind | AlphaFold Protein Structure Database: About | The AlphaFold Protein Structure Database provides open access to over 200 million protein structure predictions for the scientific community. |
| SE002 | EMBL-EBI | AlphaFold at EMBL-EBI: New tools for biological insights | |
| SE003 | Nature | Accurate structure prediction of biomolecular interactions with AlphaFold 3 | AlphaFold 3 can, with high accuracy, predict the joint structure of complexes including proteins, nucleic acids, small molecules, ions and modified residues. |
| SE004 | Google DeepMind | google-deepmind/alphafold GitHub Repository | 12.4k stars — open-source implementation of AlphaFold protein structure prediction. |
| SE005 | Google DeepMind | AlphaFold 3: Model predictions of molecular interactions | AlphaFold 3 is used exclusively by Isomorphic Labs for drug design; non-Isomorphic commercial applications require a separate licensing agreement. |
| SE006 | RCSB Protein Data Bank | RCSB PDB: Summary Statistics | |
| SE007 | National Center for Biotechnology Information | Structure-based drug design using AlphaFold: applications and limitations | |
| SE008 | bioRxiv | Generative protein binder design using diffusion-based methods | |
| SE009 | ChemRxiv | Generative AI for drug-like molecule design: progress and challenges | |
| SE010 | Science | De novo design of protein structure and function with RFdiffusion | |
| SE011 | Europe PubMed Central | Artificial intelligence in drug discovery: opportunities, challenges, and future directions | |
| SE012 | Isomorphic Labs | Our Platform — IsoDDE | Isomorphic Labs | |
| SE013 | Google DeepMind | google-deepmind/alphafold3 GitHub Repository | |
| SE014 | TechCrunch | Isomorphic Labs is using AlphaFold 3 to accelerate drug discovery | |
| SE015 | VentureBeat | How Isomorphic Labs' AlphaFold 3 advantage is reshaping pharmaceutical discovery | |
| SE016 | Fierce Biotech | Isomorphic Labs' AI-designed drug ISM8969 clears FDA IND, heading toward Phase I | Isomorphic Labs announced its internally developed AI-designed drug candidate ISM8969 has received FDA IND clearance, paving the way for Phase I clinical trials expected by end-2026. |
| SE017 | National Center for Biotechnology Information | AI biosafety and dual-use risks in biology: current frameworks and gaps | |
| SE018 | bioRxiv | Integrating AlphaFold predictions into structure-based drug discovery pipelines | |
| SE019 | Nature Chemistry | Generative models for molecular design in drug discovery | |
| SE020 | Science | Computational protein design: recent advances and remaining challenges | |
| SE021 | Europe PubMed Central | Regulatory frameworks for AI in clinical drug discovery: EU AI Act and FDA guidance | |
| SE022 | ChemRxiv | AI-driven lead optimization: benchmarks and industry case studies | |
| SE023 | RCSB Protein Data Bank | RCSB PDB: Growth of Structural Data | |
| SE024 | EMBL-EBI | AlphaFold 3 Terms of Use, FAQ and Licensing | AlphaFold 3 model weights for drug discovery applications are available exclusively to Isomorphic Labs; other commercial uses require a separate license from Google DeepMind. |
| SE025 | Google DeepMind | AlphaProteo: Designing new proteins from scratch | AlphaProteo achieves a median 3-4x improvement in binding affinity and specificity over the previous leading protein design methods in head-to-head benchmark comparisons. |
| SE026 | EMBL-EBI / Google DeepMind | AlphaFold Server: Non-commercial Use and Statistics | |
| SE027 | Google DeepMind | google-deepmind/deepmind-research GitHub Repository | |
| SU001 | Eli Lilly Investor Relations | Lilly and Isomorphic Labs Partner to Advance AI Drug Discovery | Lilly and Isomorphic Labs partner to advance AI drug discovery through multi-target collaboration. |
| SU002 | PR Newswire | Isomorphic Labs Announces Strategic Multi-Target Research Collaboration with Lilly | One of the largest AI-powered drug discovery collaborations announced. |
| SU003 | Novartis | Novartis and Isomorphic Labs Enter Collaboration to Develop New Medicines | Novartis and Isomorphic Labs enter collaboration to develop new medicines using AI. |
| SU004 | PR Newswire | Isomorphic Labs Announces Strategic Multi-Target Research Collaboration with Novartis | Multi-target AI drug discovery collaboration with Novartis announced. |
| SU005 | J&J Investor Relations | Johnson & Johnson and Isomorphic Labs Research Collaboration | Johnson & Johnson enters research collaboration with Isomorphic Labs. |
| SU006 | PR Newswire | Isomorphic Labs Announces Research Collaboration with Johnson & Johnson | Research collaboration with Johnson & Johnson announced. |
| SU007 | Business Wire | Isomorphic Labs and Eli Lilly Announce Multi-Target Drug Discovery Collaboration | Multi-target drug discovery collaboration covering multiple disease areas. |
| SU008 | BioPharmaTrend | Isomorphic Labs Announces Collaborations with Eli Lilly and Novartis for AI-Driven Drug Discovery | Isomorphic Labs signs two major pharma collaborations on the same day. |
| SU009 | pharmaphorum | J&J Bets on Isomorphic's AI-Powered Drug Hunt | J&J adds Isomorphic Labs to its AI drug discovery partner roster. |
| SU010 | Fierce Biotech | Isomorphic Labs J&J Deal for AI Drug Discovery | Fierce Biotech confirms J&J Isomorphic Labs deal announcement. |
| SU011 | BioIndustry Association | Isomorphic Labs Announces Novartis Collaboration Expansion | Novartis and Isomorphic expand their collaboration to additional therapeutic areas. |
| SU012 | BioSpace | Isomorphic Labs Novartis Deal Expansion 2025 | BioSpace reports Novartis expands AI collaboration with Isomorphic Labs in 2025. |
| SU013 | Eli Lilly | Lilly and Isomorphic Labs Partnership 2024 | Lilly describes AI drug discovery partnership with Isomorphic Labs. |
| SU014 | Isomorphic Labs | Isomorphic Labs and Lilly Partner to Advance AI Drug Discovery | Isomorphic Labs and Lilly partner to advance AI drug discovery. |
| SU015 | Isomorphic Labs | Isomorphic Labs Partnerships | Isomorphic Labs lists all active pharma partnerships on its partnerships page. |
| SU016 | Isomorphic Labs | AI Research Collaborations: Lilly and Novartis | Both collaborations are operational and generating data insights. |
| SU017 | PR Newswire | Isomorphic Labs Announces $2.1 Billion Funding | $2.1 billion raised to power world-leading AI drug design engine. |
| SU018 | Business Wire | Isomorphic Labs Announces $2.1 Billion Funding (Business Wire) | Isomorphic Labs raises $2.1 billion to accelerate AI drug design. |
| SU019 | STAT News | Isomorphic Labs Raises $2.1 Billion Series B | Isomorphic Labs raises $2.1 billion Series B; investors cite pharma partnerships as commercial rationale. |
| SU020 | Bloomberg | Isomorphic Labs Raises $2.1 Billion for AI Drug Discovery | Bloomberg reports $2.1 billion Series B for AI drug discovery. |
| SU021 | Wired | Isomorphic Labs Raises $2.1 Billion Series B Funding | Wired covers Isomorphic Labs $2.1 billion Series B funding round. |
| SU022 | Pharmaceutical Technology | Isomorphic Labs $2.1bn Series B Funding | Pharmaceutical Technology covers Isomorphic Labs $2.1bn Series B. |
| SU023 | Evaluate Vantage | Isomorphic Labs Series B Analysis | Evaluate Vantage notes Isomorphic Labs remains pre-revenue with uncertain milestone timelines. |
| SU024 | Fierce Pharma | Isomorphic Labs Raises $2.1B for AI Drug Discovery | Fierce Pharma covers $2.1 billion Isomorphic Labs funding round. |
| SU025 | BioSpace | Isomorphic Labs Raises $2.1 Billion Series B Accelerating AI Drug Discovery | BioSpace reports $2.1 billion Series B accelerating AI drug discovery. |
| SU026 | Nature | The Limits of AI in Drug Discovery: What Pharma Must Know | AI drug-discovery partnerships frequently fail to convert milestones; in-house AI threatens proprietary advantages of platform companies. |
| SU027 | Wall Street Journal | Alphabet Spinoff Isomorphic Labs Closes $2.1 Billion Funding Round | WSJ: one of the largest AI drug discovery funding rounds globally. |
| SU028 | Sifted | Isomorphic Labs $2.1bn Series B | Sifted covers Isomorphic Labs $2.1bn Series B funding. |
| SU029 | Endpoints News | Isomorphic Labs Raises $2.1B Series B for AI Drug Discovery | Endpoints News: Isomorphic Labs raises $2.1B Series B; pharma partnerships cited as rationale. |
| SU030 | Tech Funding News | Isomorphic Labs Raises $2.1B Series B from Alphabet and Others | Alphabet and external investors lead $2.1B Series B for Isomorphic Labs. |
| SR001 | EUR-Lex (European Parliament) | Regulation (EU) 2024/1689 — EU AI Act Full Text | AI systems used in healthcare and drug development classified as high-risk under Annex III. |
| SR002 | UK Information Commissioner's Office | ICO Work on Artificial Intelligence | ICO published guidance on GDPR obligations for AI systems processing health-related data. |
| SR003 | FTC (US Federal Trade Commission) | FTC Launches Inquiry into AI Foundation Models | FTC launches broad inquiry into AI foundation model companies and their commercial practices. |
| SR004 | FTC | AI Companies: Uphold Your Privacy Obligations | FTC guidance emphasises privacy obligations for AI companies processing personal data. |
| SR005 | UK Companies House | Isomorphic Labs Ltd — Companies House Filings (Company 13036082) | Isomorphic Labs Ltd incorporated in England and Wales, no publicly disclosed litigation. |
| SR006 | ICO | UK GDPR Guidance and Resources | GDPR obligations apply to AI systems processing health data even when de-identified. |
| SR007 | Nature | AI Drug Discovery: Promises and Pitfalls | AI drug-discovery models still fail frequently in wet-lab validation; hallucination risk in molecular generation is material. |
| SR008 | Nature | AI in Drug Discovery: Caution Ahead | Pharma companies increasingly building in-house AI; external platform advantages may erode by 2027. |
| SR009 | Nature | AI and Drug Discovery: The Current Limits | Current AI drug-discovery tools have significant limitations in novel target classes beyond protein structure prediction. |
| SR010 | BenevolentAI | BenevolentAI Strategic Overhaul 2024 | BenevolentAI undertakes strategic overhaul following AstraZeneca partnership failure. |
| SR011 | BioPharma Dive | Exscientia Recursion Merger: AI Drug Discovery Landscape 2024 | Recursion acquires Exscientia creating the largest independent AI drug-discovery company. |
| SR012 | Recursion Pharmaceuticals | Recursion Full Year 2024 Financial Results | Recursion reports full year 2024 results: significant revenue but large operating losses. |
| SR013 | Recursion Pharmaceuticals | Recursion Q4 2024 Financial Results | Recursion Q4 2024: operating cash burn remains substantial; milestone revenue uncertain. |
| SR014 | Nature | AI in Drug Discovery: 2026 Status Report | 2026 status: AI drug-discovery sector facing regulatory headwinds and milestone conversion challenges. |
| SR015 | PubMed | Limitations of AI in Drug Discovery: A Systematic Review | Systematic review finds AI drug-discovery tools show promise but consistently fail on novel chemical space. |
| SR016 | BioPharma Dive | AI Drug Discovery Company Landscape 2025 | AI drug-discovery sector consolidating; milestone conversion remains the central unproven assumption. |
| SR017 | Evaluate Vantage | AI Drug Discovery Deals and Financings Analysis | Evaluate analysis: AI drug-discovery deals carry high milestone uncertainty; sector average conversion rate below 30%. |
| SR018 | Recursion Pharmaceuticals IR | Recursion 2024 Annual Report Summary | Recursion 2024 annual report: extensive pipeline but large pre-revenue burns in drug-discovery AI. |
| SR019 | Fierce Biotech | Recursion Exscientia Merger Creates AI Drug Discovery Giant | Recursion-Exscientia merger creates the largest independent AI drug-discovery company. |
| SR020 | Nature Medicine | Clinical Translation Challenges in AI-Designed Therapeutics | AI-designed therapeutics face significant attrition at wet-lab and ADMET validation stages. |
| SR021 | McKinsey Global Institute | AI in Drug Discovery: Outlook and Risks | McKinsey: AI drug-discovery market growing but competitive dynamics intensifying; pharma in-house AI investment accelerating. |
| SR022 | VentureBeat | AlphaFold 3 Predicts Building Blocks of Life | AlphaFold 3 extends structure prediction to DNA, RNA, and small molecules — a step toward drug design. |
| SR023 | STAT News | Isomorphic Labs Raises $2.1 Billion | Isomorphic Labs raises $2.1 billion; company remains pre-revenue with all milestone value contingent. |
| SR024 | BCG | AI in Pharma Drug Discovery: Strategic Risks and Opportunities | BCG: Big pharma increasingly building in-house AI capabilities, threatening external AI platform providers. |
| SR025 | Xaira Therapeutics | Xaira Therapeutics Emerges from Stealth with $1B Raise | Xaira Therapeutics raises $1 billion with AI drug-discovery platform positioning directly against Isomorphic Labs. |
| SR026 | CourtListener | Authors Guild v OpenAI — AI Copyright Litigation | Authors Guild v OpenAI sets legal precedent context for AI training data IP disputes relevant to AI platforms. |
| SR027 | CourtListener | Getty Images v Stability AI — AI Training Data Lawsuit | Getty Images v Stability AI: courts examining AI training data provenance and copyright infringement. |
| SR028 | PR Newswire | Isomorphic Labs $2.1 Billion Series B Announcement | Isomorphic Labs announces $600M initial closing of Series B funding. |
| SR029 | Science.org | AI Making Huge Inroads in Drug Discovery | Science.org: AI drug discovery making significant inroads but clinical translation remains the unsolved challenge. |
| SR030 | Grand View Research | AI in Drug Discovery Market Analysis | AI drug-discovery market growing at 40%+ CAGR but dominated by platform companies facing increasing competitive pressure. |
| SV001 | Isomorphic Labs | Isomorphic Labs Announces Series B Investment Round | Isomorphic Labs today announced $2.1 billion in Series B financing to accelerate its AI-driven drug design platform. |
| SV002 | TechCrunch | Isomorphic Labs raises $2.1B in Series B funding | Isomorphic Labs has raised $2.1 billion in a Series B round led by Thrive Capital, with continued participation from parent company Alphabet. |
| SV003 | Axios | Isomorphic Labs raises $2 billion for AI drug design | The round values Isomorphic Labs at approximately $5B-$8B post-money, making it one of the highest-valued AI drug-discovery companies globally. |
| SV004 | Yahoo Finance | Isomorphic Labs $2.1 Billion Funding AI Drug Design 2026 | Isomorphic Labs secures $2.1B in new funding, underscoring investor confidence in AI-driven approaches to pharmaceutical drug discovery. |
| SV005 | PitchBook | Isomorphic Labs Series B Company Profile | PitchBook estimates post-money Series B valuation at $5.0B-$8.0B based on deal size, comparable transactions, and negotiation signals. |
| SV006 | PitchBook | Isomorphic Labs Company Profile | Isomorphic Labs total funding: $2.6B+ across two known rounds. Pre-revenue stage. Three confirmed pharma partner collaborations. |
| SV007 | Dealroom | Isomorphic Labs Company Profile | Dealroom valuation estimate: $6.5B-$8.0B post-money following May 2026 Series B. Listed under deep-tech / AI biotech category. |
| SV008 | Compworth | Isomorphic Labs Company Valuation | Compworth mid-point valuation estimate: approximately $6.8B based on $2.1B Series B round size and comparable multiples. |
| SV009 | Global Fund Media | Isomorphic Labs Secures $2.1B Series B to Advance AI Drug Discovery | The $2.1B Series B makes Isomorphic Labs the best-capitalised private AI drug-discovery platform. |
| SV010 | Schrodinger IR | Schrodinger Reports Fourth Quarter and Full Year 2024 Results | Schrodinger total revenue 2024: approximately $170M; platform segment growing. Market cap approximately $1.5B-$2.5B. |
| SV011 | Recursion IR (annual report PDF) | Recursion Annual Report 2024 Full Year Results | Recursion 2024 full year operating expenses approximately $280M; revenue approximately $60M. |
| SV012 | Recursion Pharmaceuticals IR | Recursion Reports First Quarter 2026 Financial Results | Recursion Q1 2026 operating expenses approximately $72M; revenue approximately $65M; market cap approximately $3.0B. |
| SV013 | Recursion Pharmaceuticals IR | Recursion Reports Full Year and Fourth Quarter 2025 Financial Results | Full year 2025 operating expenses approximately $290M; revenue $245M full year. |
| SV014 | Business Wire | Recursion Reports Fourth Quarter and Full Year 2024 Results | Recursion 2024 full year results: revenue $66M, operating expenses $283M, cash and equivalents approximately $780M. |
| SV015 | Yahoo Finance | Isomorphic Labs Raises $2.1 Billion | Isomorphic Labs announces $2.1 billion Series B. Deal led by Thrive Capital. |
| SV016 | Pharmaphorum | Isomorphic Signs Lilly and Novartis $3B AI Drug Hunt Deals | Isomorphic Labs signed landmark deals with Eli Lilly (up to $1.7B) and Novartis ($1.0B) to deploy its AI drug design capabilities. |
| SV017 | Business Wire | Isomorphic Labs and Lilly Partner to Advance AI Drug Discovery | Isomorphic Labs and Eli Lilly today announced a multi-target drug discovery collaboration worth up to $1.7 billion including milestones and royalties. |
| SV018 | Eli Lilly IR | Lilly Signs Multi-Target Drug Discovery Collaboration with Isomorphic Labs | Lilly and Isomorphic Labs signed a multi-target drug discovery collaboration agreement. Total potential deal value up to $1.7B. |
| SV019 | Johnson and Johnson IR | Isomorphic Labs Johnson and Johnson Collaboration 2026 | Johnson and Johnson and Isomorphic Labs announce a drug discovery collaboration leveraging Isomorphic AI platform for multiple therapeutic targets. |
| SV020 | TechCrunch | Google DeepMind spinout Isomorphic Labs signs $1.7B and $1B drug discovery deals with Lilly and Novartis | Isomorphic Labs spinout from Google DeepMind secures $2.7B in headline partnership commitments from Lilly and Novartis. |
| SV021 | BCG | Artificial Intelligence in Pharma Drug Discovery | AI-enabled drug discovery could generate over $1 trillion in long-run pharmaceutical sector value by reducing R&D timelines and compound attrition rates by 25-50%. |
| SV022 | Google DeepMind | Isomorphic Labs: A new company focused on drug discovery using AI | Isomorphic Labs will have full access to DeepMind research capabilities and resources including compute, talent, and foundational AI models. |
| SV023 | Alphabet Inc. | Alphabet Other Bets Investor Reporting | Other Bets revenues and costs are reported in aggregate. Individual subsidiaries including Isomorphic Labs are not separately reported. |
| SV024 | Google DeepMind | Isomorphic Labs and Google DeepMind Join Forces to Accelerate Drug Discovery | DeepMind and Isomorphic Labs will work together with unprecedented compute access and shared model infrastructure to accelerate drug discovery timelines. |
| SV025 | PitchBook | Recursion Pharmaceuticals Company Profile | Recursion market cap approximately $3.0B as of Q2 2026. Primary listed AI drug-discovery benchmark for private market comparables. |
| SV026 | PitchBook | Schrodinger Inc. Company Profile | Schrodinger market cap approximately $1.5B-$2.5B. Revenue multiple 9-15x on approximately $170M 2024 revenue. |
| SV027 | Globe Newswire | Isomorphic Labs Series B Announcement | Isomorphic Labs announces $2.1 billion Series B financing round. |
| SV028 | PR Newswire | Isomorphic Labs Announces 600 Million Funding | Isomorphic Labs announces $600 million in Series A funding to advance its AI drug design engine. |
| SV029 | Drug Hunter | Xaira Therapeutics: A $1 Billion AI Drug Design Competitor | Xaira Therapeutics raises $1 billion to deploy AI-first drug design, directly competing with Isomorphic Labs and other platform players for pharma partnership deals. |
| SV030 | BenevolentAI | BenevolentAI News and Media — 2024 Strategic Overhaul | BenevolentAI 2024 strategic overhaul, following its AstraZeneca partnership setback, serves as a cautionary precedent for AI drug-discovery platform valuation. |