Lila Sciences
资本形成能力突出,自主科学基础设施野心很大,但商业牵引力和可重复科研产出的公开证据仍落后于估值。
Lila 是市场上资本最充足的 AI-for-science 初创公司之一,但当前估值已经预设了公开记录尚未完全证明的科学和商业结果。
封面要素
公司概况
Lila Sciences 是 Flagship Pioneering 孵化的创业公司,成立于 2023 年,并在 2025 年 3 月公开亮相,目标是打造其所谓的科学超级智能。其平台把 Lila Iris、科学软件、自动化实验、机器人和 AI Science Factories 组合起来,加速疗法、生物技术、化学和材料领域的发现。公开证据还显示,公司早期融资基础异常庞大:先有 $200M 种子轮,再有 $350M Series A,累计资本 $550M,并在 Cambridge 拿下重要实验室空间。公开记录同样缺失的内容也很关键:已命名付费客户、收入、利润率,以及科研产出主张能否在规模化下重复的独立验证。
- 成立时间
- 2023-01-01
- 创始人
- Geoffrey von Maltzahn
- 创立地点
- Cambridge, Massachusetts
- 总部
- Cambridge, Massachusetts
- 产品
- Lila 向合作伙伴出售闭环 AI 科学平台的访问权。该平台可生成假设、设计并运行实验,并借助其科学模型和 AI Science Factory 实验室基础设施返回已验证数据、资产或技术路线图。
- 客户
- 生物制药、生物技术、化学品、材料、能源、半导体,以及其他研究密集型组织;这些机构需要更快的发现周期,但不想自行搭建完整的 AI 加实验室栈。
- 商业模式
- 平台访问与科学服务混合模式:Catalyst 提供 Lila Iris、科学专家和 Lab-as-a-Service 产能;Creation 运行端到端合作伙伴项目,目标是产出已验证资产、数据包、IP 或新公司。
- 阶段
- Series A / pre-commercial
- 融资情况
- 2025 年 3 月发布时宣布 $200M 种子轮,2025 年完成 $350M Series A,使已披露融资总额达到 $550M,最新已披露估值超过 $1.3B。
执行摘要
主要优势
- 早期资本底座和投资人质量都很突出,包括 Flagship、Braidwell、Collective Global 以及 Nvidia 背书的参投方。
- 差异化的端到端 AI Science Factory 逻辑,把模型、机器人和自动化实验组合起来,而不是只做软件发现工具。
- 创始人与科学团队水准高,兼具深厚 Flagship 和前沿科学可信度。
- 如果闭环平台证明可重复,疗法、生物技术、化学和材料都可能打开大上行空间。
主要风险
- 目前没有公开收入、毛利率、定价、利用率或具名付费客户数据支撑传统承销。
- 公开科学证据还撑不起其雄心和估值叙事,repeatability 和 transferability 仍是开放问题。
- Series A 阶段估值超过 $1.3B,执行失误的余地很小。
- 模型看起来资本开支重,因为需要实验室基础设施、机器人、前沿 AI 人才和大量算力。
- 生命科学和材料工作流中,监管、生物安全、IP 和下游商业化交接风险仍然显著。
未决问题
- Lila 所称 discovery outcomes、benchmark wins 和 throughput economics 的独立验证仍有限。
- 具名客户 reference、定价、合同结构和 recurring revenue quality 仍未披露。
- 烧钱速度、现金跑道、毛利率和股权结构条款未公开。
- AI Science Factories 能否在多个领域无需定制支持就可重复扩张,证据仍不完整。
目录
01公司概况
1.1 身份、使命和商业模式
Lila Sciences 把自己定义为一家科学超级智能公司,而不是通用 AI 实验室或单一产品型生物技术公司。官方材料反复描述的平台,是一个 AI 系统:它能生成假设、设计实验、通过 AI Science Factory 仪器运行实验,并从生命科学、化学和材料问题产生的数据中实时学习。这一点对尽调很重要,因为其商业形态看起来是平台访问:Lila 正在为外部科学项目搭建自动化实验室和企业软件,而不是传统的内部疗法管线。Reuters 也强化了这一定位,报道称管理层希望 Lila 平台上的合作伙伴和创业公司把分子、材料或能源突破推进到下游开发。因此,第一章的运营结论是:Lila 同时像前沿模型公司、机器人实验室运营商和发现基础设施供应商。公开材料有力支持其使命和架构,但尚未给出量化收入、已命名客户或员工数披露;尽管融资故事规模很大,核心商业化指标仍部分不透明。[CO001, CO002, CO003, CO004, CO024, CO025]
| 指标 | 数值 / 状态 | 日期 | 置信度 | 备注 / 缺口 |
|---|---|---|---|---|
| 成立 | 2023 | 2023 | 高 | 成立于 Flagship 实验室;公开亮相发生在 2025 年 3 月。 |
| 总部 | Cambridge, Massachusetts | 2025-10 | 高 | Reuters、AGBI、Economic Times 以及 CNBC 专访均提供支持。 |
| Flagship 设施 | 235,500 sq ft 租赁 | 2025-10 | 高 | Alewife Park 占地被用作旗舰 AI Science Factory 规模标记。 |
| 其他已披露枢纽 | San Francisco;London | 2025-10 至 2026-05 | 中 | 官方和独立报道都显示公司在多城市扩张。 |
| 种子轮融资 | $200M 已承诺 | 2025-03-10 | 高 | Lila 和 PR Newswire 披露了启动融资。 |
| Series A 总额 | $350M | 2025-10-14 | 高 | 包括 10 月 $115M 扩展轮。 |
| 累计融资 | $550M | 2025-10-14 | 高 | 种子轮加完整 Series A。 |
| 最新估值 | >$1.3B | 2025-10-14 | 高 | Reuters、Goodwin、CNBC 和转载报道相互印证。 |
| 具名客户 / 客户数 | 未公开披露 | 2026-06-02 | 中 | 提到首批客户队列,但没有公开名称或数量。 |
| 收入 / 运行率 | 未公开披露 | 2026-06-02 | 中 | 已审阅来源没有提供收入或 ARR。 |
| 员工数 | 未公开披露 | 2026-06-02 | 低 | 招聘和扩张表述暗示在招人,但没有给出人数。 |
这里的 null 式运营指标表示公开证据包没有披露该数字,不代表指标为零或无关紧要。
[CO005, CO007, CO019, CO020, CO021, CO027]Lila 如何把科学问题、AI 模型、自动化实验室、自有数据和合作伙伴商业化串起来。
[CO002, CO003, CO024, CO025, CO026, CO032]1.2 创立故事、领导层和治理
Lila 的创立叙事与 Flagship Pioneering 绑定得异常紧。PR Newswire 和 Flagship 称,Lila 于 2023 年在 Flagship 实验室成立,并在多年孵化后于 2025 年 3 月公开发布;公司自己的发布说明称,它在 Flagship 内部幕后搭建了约三年。Geoffrey von Maltzahn 是故事核心:Lila 和 Flagship 都把他描述为连续公司创建者,履历覆盖 Generate:Biomedicines、Tessera、Indigo、Sana、Seres 及相关企业。治理也仍然高度依赖赞助方,因为 Noubar Afeyan 既是 Flagship 创始人 / CEO,也是 Lila 联合创始人 / 董事长。围绕 Geoffrey,公开领导团队强于典型刚亮相的平台公司:Andrew Beam 支撑 AI 科学可信度,Jawad Ahsan 带来规模化财务纪律,Chris Fussell 带来组织和国家安全运营经验,Julie Shah 增加机器人深度,Rafael Gómez-Bombarelli 强化化学 / 材料覆盖。主要尽调风险不是缺少资深领导者,而是 Flagship 还保留多少实际控制权,以及这种治理结构在后续融资或商业化阶段能维持多久。[CO005, CO006, CO009, CO010, CO011, CO012]
| 人物 | 职务 | 背景 | 覆盖范围 / 创始人市场匹配 | 关键人物依赖 |
|---|---|---|---|---|
| Geoffrey von Maltzahn(首席执行官) | 联合创始人兼 CEO | Flagship 普通合伙人;曾任 Generate、Tessera、Indigo、Sana、Seres 及相关公司的创始 CEO 或联合创始人 | 设定使命、融资叙事、公司创建路径,并在 AI 与生物技术投资人中建立外部可信度 | 高 — 公司故事和投资人信心都与他紧密绑定 |
| Noubar Afeyan | 联合创始人兼董事长 | Flagship Pioneering 创始人兼 CEO;Moderna 时代的公司创建者 | 嵌入发起方治理、资本通道和战略监督 | 高 — 对发起方连续性和董事会影响力至关重要 |
| Andrew Beam | 首席技术官 | Generate:Biomedicines 联合创始人;前 Flagship 高级研究员;Harvard 流行病学教员 | 掌管 AI 科学架构和技术可信度 | 高 — 对差异化模型质量至关重要 |
| Jawad Ahsan | COO 与 CFO | 前 Axon CFO;前 Numerator/Market Track CFO;GE 财务老将 | 补入规模化运营财务、规划和资本市场纪律 | 中高 — 对烧钱管理和基础设施扩张很重要 |
| Chris Fussell | 运营总裁 | 前 Navy SEAL 军官、前 McChrystal Group 总裁 | 带来组织设计、跨职能执行和政府相邻能力 | 中 — 对执行和国家安全叙事有意义 |
| Julie Shah | 首席机器人官 | MIT 机器人学领军人物、AeroAstro 系主任 | 加深实验室自动化和人机系统能力 | 中 — 支撑物理实验室投资逻辑 |
| Rafael Gómez-Bombarelli(首席科学官) | 联合创始人兼 CSO,物理科学 | MIT 材料科学家、机器学习化学先行者 | 锚定生命科学之外的化学 / 材料扩张 | 中高 — 对非生物技术可信度很重要 |
该部分表格聚焦对公司创建、机器人、科学和财务尽调最关键的高管与发起方,而非完整组织架构。
[CO009, CO010, CO011, CO012, CO013, CO014]1.3 资本基础、投资者和运营足迹
Lila 从隐身状态很快变成大额资本故事。2025 年 3 月发布时带着 $200 million 种子轮,9 月 Series A 首次关闭新增 $235 million,10 月扩展轮把 Series A 推至 $350 million,累计资本达到 $550 million。Reuters、CNBC、Goodwin 及转载报道均把扩展后的估值置于 $1.3 billion 以上。辛迪加的重要性不只在规模:Flagship 仍处核心位置,Braidwell 和 Collective Global 领投首次关闭,扩展轮又加入 NVentures、IQT、Analog Devices、Catalio 等投资者,拓宽了公司的 AI、国防和工业邻近性。实体足迹同样激进。Reuters 及相关报道称,Lila 签下 235,500 平方英尺的 Cambridge 租约,被描述为 2025 年 Boston 最大实验室租约之一;Flagship 和后续媒体还指向 San Francisco 与 London 的进一步扩张。实体建设很关键,因为 Lila 的投资逻辑依赖拥有自动化实验产能,而不只是训练更大的软件模型。[CO018, CO019, CO020, CO021, CO022, CO023]
| 利益相关方 | 角色 | 控制 / 经济重要性 | 证据 | 尽调问题 |
|---|---|---|---|---|
| Flagship Pioneering | 创始方、孵化器和持续投资人 | 发起 Lila,并通过 Noubar Afeyan 以及种子轮和 Series A 的持续参与保持绑定 | Flagship 公司页面、Geoffrey 简介、发布新闻稿、Series A 公告 | 索取当前持股、董事会权利,以及与 Flagship 关联方的任何平台服务协议 |
| Braidwell | Series A 联合领投方 | 锚定首个 $235M 交割,并可能为新资金治理设定价格 / 参照点 | CafePharma 和 Robotics & Automation News 对 Series A 首次交割的报道 | 确认董事会席位、按比例认购权和扩展轮参与情况 |
| Collective Global | Series A 联合领投方 | 与 Braidwell 共同领投首次交割 | 同一批首次交割报道和后续官方回顾 | 确认持股比例,以及权利是否与 Braidwell 匹配 |
| NVentures | 扩展轮投资人 / AI 战略连接 | 增加 NVIDIA 相邻性,并帮助把估值推到 $1.3B 以上 | 官方扩展轮公告以及 Reuters / Fierce | 澄清投资是否也附带算力、进入市场或技术合作钩子 |
| General Catalyst | 种子轮和 Series A 投资人 | 重复参与使其成为持久的跨阶段支持者,而非一次性名字 | 种子轮公告和 Series A 合作伙伴名单 | 核查储备策略、信息权和后续成长期融资意愿 |
| ADIA 子公司 | 种子轮和 Series A 投资人 | 在多轮融资中提供主权资本存在感 | 种子轮发布说明加 Series A 合作伙伴名单 | 确认持股集中度、时间周期和任何附函经济条款 |
| IQHQ / Alewife Park 房东 | 基础设施利益相关方 | 235,500 sq ft Cambridge 租赁支撑物理实验室扩张故事 | Bisnow 租赁报道和 Reuters 引用 | 审查入驻时间、租户装修和与实验室建设绑定的最低支出 |
该部分利益相关方地图聚焦最可能影响治理、规模或后续融资的资本提供方和基础设施关系。
[CO008, CO016, CO018, CO022, CO023, CO027]目前支撑公司概况叙事的资本、估值和场地指标。
[CO005, CO007, CO019, CO020, CO021, CO027]1.4 里程碑、商业化和怀疑信号
里程碑记录显示,公司正试图把戏剧性的资本故事转化为可信的平台业务。2025 年 3 月发布确立了种子轮融资和公开使命。9 月 Series A 首次关闭和 10 月扩展轮随后抬高估值叙事;管理层在官方融资公告中称,Lila 正在欢迎首批客户并向外部合作伙伴开放平台。Reuters 补充称,能源、半导体和药物开发公司已经表现出兴趣,但也表示 Lila 不计划亲自把产品一路推进到临床开发或大规模工业部署。相较于完全整合的生物技术公司,这让模型更轻资本;但证明也取决于合作伙伴转化和已验证案例研究。两个怀疑来源加深了风险。Fierce Biotech 指出,公司尚未公开发布支持多项发现主张的数据;CNBC 称,热度可能跑在现实前面,因为许多 AI 平台仍难以稳定跑赢传统研究模型。因此,近期尽调问题在于:在估值叙事进一步跑赢证据之前,Lila 能否把科学超级智能话术转化为外部可审计结果。[CO007, CO018, CO019, CO024, CO025, CO026]
| 日期 | 事件 | 类型 | 金额 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2023 | Lila 在 Flagship Pioneering 实验室内部成立 | 创立 | 公司组建 | 发起方:Geoffrey von Maltzahn;Noubar Afeyan;Flagship | 确立发起方搭建的起源故事 |
| 2025-03-10 | 结束隐身并公开发布 | 创立 | 多年孵化后公开亮相 | Lila;Flagship | 公司从内部搭建转向公开招聘和合作伙伴拓展 |
| 2025-03-10 | 宣布种子轮融资 | 融资 | $200M 已承诺 | Flagship;General Catalyst;March Capital;ADIA 子公司;其他 | 为平台和实验室基础设施的首次公开建设提供资金 |
| 2025-03-10 | 披露发布时领导团队 | 治理 | 具名资深 AI、科学和运营团队 | Geoffrey von Maltzahn;Andrew Beam;George Church;Chris Fussell;其他 | 对一家新亮相平台公司而言,显示出异常资深的创始团队 |
| 2025-09-15 | 宣布 Series A 首次交割 | 融资 | $235M,估值 $1B+ / 约 $1.2B 区间 | Braidwell;Collective Global;Flagship;既有投资人 | 显示发布后迅速获得后续融资动能 |
| 2025-09-15 | 强调新增 AI Science Factory 枢纽 | 扩张 | Boston、San Francisco 和 London 扩张计划 | Lila 领导层 | 把平台投资逻辑推进为多地点建设计划 |
| 2025-10-14 | 宣布 Series A 扩展轮 | 融资 | +$115M;Series A 达到 $350M;累计资本 $550M | NVentures;Analog Devices;IQT;Catalio;Pennant;其他 | 将估值抬至 $1.3B 以上,并扩大投资人基础 |
| 2025-10-14 | 宣布开放商业合作伙伴 | 合作 | 迎来首批客户队列 | Lila;潜在合作伙伴和初创公司 | 开始面向外部的平台访问模式 |
| 2025-10 | 签署 Cambridge 租赁 | 扩张 | 设施:Alewife Park 235,500 sq ft | Lila;IQHQ | 为 AI Science Factories 建立旗舰物理占地 |
| 2025-10 | Fierce 标记证据缺口 | 反向 | 尚无公开数据支持若干重大主张 | Fierce Biotech;Lila | 公开证据仍落后于公司的雄心叙事 |
| 2026-05-19 | CNBC Disruptor 50 专访增加怀疑 | 反向 | 排名 | CNBC;Lila | 估值和可见度上升后,外部审视增强 |
时间线优先呈现定义公司概况视角的创立、融资、扩张、合作、治理和反向信号事件;已审阅材料中没有看到公开监管里程碑。
[CO005, CO007, CO018, CO019, CO020, CO024]Lila 从 Flagship 孵化走向大额融资、并受到更多外部审视的路径时间线。
[CO005, CO006, CO007, CO018, CO019, CO021]1.5 图表
02市场分析
2.1 市场边界、纳入支出和替代栈
不能用单一通用的“AI for science”总可用市场(TAM)来承销 Lila Sciences。公开证据把相关支出拆成相邻层。实验室自动化报告聚焦药物发现、基因组学和诊断中使用的机器人系统、自动化工作站、液体处理、筛选和工作流软件。实验室信息学报告覆盖数据与控制骨干——LIMS、ELN、LES、云交付和合规工具。AI 药物发现报告描述用于靶点识别、筛选、老药新用、从头设计和临床前优先级排序的软件与服务。自驱动实验室文献又描述了更窄、更早期的编排层,连接自动化仪器、AI 决策引擎和数据系统,形成闭环实验。对 Lila 而言,纳入支出是这些层被一起采购、用于加速发现或流程优化的重叠部分。排除支出应包括常规诊断运营、通用企业 AI、完整临床开发或 CRO 服务收入,以及不处在实验闭环内的广义工业自动化。实践中,现状替代方案通常是一套碎片化栈:仪器、信息学、内部脚本、CRO 工作和人工驱动的实验规划,而不是某个直接在位产品。[CM001, CM002, CM003, CM004, CM005, CM006]
| 细分市场 / 类别 | 纳入支出 | 排除支出 | 买方 / 付款方 | 相关性 |
|---|---|---|---|---|
| 实验室自动化硬件和工作流系统 | 发现实验室使用的机械臂、液体处理、工作站、筛选工作流和工作流软件 | 常规诊断运营、广义医院自动化和通用制造机器人 | 平台研发、筛选或实验室运营负责人;由中央 R&D 或实验室 capex/opex 预算支付 | 构成 AI 科学工厂的物理执行层,也是规模最容易估算的相邻类别。 |
| 实验室信息化和数据骨干 | LIMS、ELN、LES、云交付、审计轨迹、数据采集、工作流配置和互操作层 | 通用企业数据湖、无关 ERP/CRM 系统和非实验室分析栈 | 实验室信息化、质量或数字实验室负责人;由 R&D 软件和合规预算支付 | 关键在于 Lila 需要结构化数据和编排,而不只是机器人。 |
| AI 药物发现软件和服务 | 靶点识别、分子筛选、再利用、从头设计和临床前优先级排序工具 | 临床试验软件、商业分析和无关医疗 AI | 发现信息化、转化科学或计算化学负责人;由发现项目预算支付 | 是衡量客户是否愿意为模型驱动科学加速付费的最佳代理。 |
| 自主 / 自驱动实验室编排 | 跨仪器和数据系统的闭环实验规划、执行、分析和再规划 | 没有学习循环或编排层的单一用途仪器控制 | 自动化工程或平台科学领导层;付款方通常是中央 R&D | 这是 Lila 最差异化的一层,也是公开报告中独立规模估算最少的一层。 |
| 材料和化学发现自动化 | 电池、催化剂、聚合物、特种化学品和应用材料中的高循环实验项目 | R&D 阶段之后的广义工业流程自动化、仅常规 QC 的支出 | 先进材料、配方或应用研究负责人;由创新预算支付 | 这是战略上重要的相邻领域,自驱动实验室文献在这里最强。 |
| 现状替代栈 | 内部脚本、碎片化仪器、CRO 工作、人工实验设计和点状解决方案 | N/A | 科学团队和实验室经理通过人员和供应商碎片化间接吸收支出 | 这是实际替代目标;Lila 很少替代一个单一的庞大既有厂商。 |
纳入支出要求存在重复实验循环,模型、数据和自动化执行被一起购买;排除支出位于该闭环发现工作流之外。
[CM001, CM002, CM003, CM004, CM005, CM006]2.2 规模测算视角、矛盾估计和受证据约束的范围
公开市场证据足以说明类别重要性,但不足以支撑一个头条 TAM。仅实验室自动化一项,2026 年规模估计就从 FMI 的约 US$2.7 billion 到 Business Research Insights 的 US$12.12 billion 不等,MarketsandMarkets 为 US$6.60 billion,Precedence 为 US$8.91 billion。实验室信息学更窄但同样不一致:Mordor 估计 2026 年 US$4.05 billion,Business Research Insights 为 US$5.4 billion;Grand View 则把 2025 年该类别定在 US$4.1 billion,并预计到 2033 年 CAGR 较慢,为 4.9%。AI 药物发现按当前收入看是最小的相邻类别,但增长最快:Mordor 认为 2026 年市场为 US$3.25 billion,到 2031 年 CAGR 为 25.94%;Global Market Insights 称 2025 年市场已超过 US$3.1 billion,并将在 2035 年前每年增长 30.5%。这些数字支持一个低十亿美元级别的广义相邻市场包络,但不能干净相加成 TAM,因为同一买家可以同时购买三层。更可信的结论是,Lila 追逐的是几个已获资金支持类别中的快速增长集成问题,而最内层的自主实验室控制层在公开资料中仍未被量化。[CM007, CM008, CM009, CM010, CM011, CM012]
| 发布方 | 年份 | 地理范围 | 数值 | CAGR | 方法论 | 置信度 | 局限 |
|---|---|---|---|---|---|---|---|
| MarketsandMarkets | 2026 | 全球 | 6.6 | 6.6% (2026-2031) | 跨硬件、软件、应用和终端用户的实验室自动化市场摘要 | 中 | 有用的基准,但仍只是 Lila 栈中的一层。 |
| Precedence Research | 2026 | 全球 | 8.91 | 6.55% (2025-2034) | 实验室自动化市场公开执行摘要 | 中 | 高于 MarketsandMarkets,显示边界差异。 |
| Future Market Insights(市场来源) | 2026 | 全球 | 2.7 | 9.7% (2026-2036) | 带终端用户分割的实验室自动化市场摘要 | 中 | 相比其他发布方非常保守。 |
| Business Research Insights | 2026 | 全球 | 12.12 | 8.47% (2026-2035) | 实验室自动化市场公开摘要 | 低 | 外壳估计激进,方法透明度较低。 |
| Mordor Intelligence | 2026 | 全球 | 4.05 | 8.46% (2026-2031) | 实验室信息化市场估计和细分 | 中 | 软件数据层,不是完整自动化栈。 |
| Business Research Insights | 2026 | 全球 | 5.4 | 9.11% (2026-2035) | 实验室信息化市场摘要 | 低 | 高于 Mordor,方法透明度更低。 |
| Mordor Intelligence | 2026 | 全球 | 3.25 | 25.94% (2026-2031) | AI 药物发现市场估计和细分 | 中 | 高增长类别,但仍是软件 / 服务视角。 |
| Global Market Insights(市场来源) | 2025 基准 | 全球 | 3.1 | 30.5% (2026-2035) | 引用 2025 基准和远期 CAGR 的 AI 药物发现市场摘要 | 中 | 以 2025 基准发布,而不是直接给出 2026 点估计。 |
这些行是相邻市场视角,不是一个干净 TAM 的可相加细分。它们支持规模和增长方向,但 Lila 的实际可服务市场仍取决于客户组合和部署模式。
[CM007, CM008, CM009, CM010, CM011, CM012]对 Lila 最站得住脚的市场口径,应从几个相邻融资类别收窄到尚未定量的自主实验室控制层。
这个金字塔是范围观察口径,不是收入汇总。它显示公开类别证据在哪里最强,以及市场判断从哪里开始主观化。
[CM001, CM005, CM017, CM018, CM019, CM043]已发布的相邻类别估计有方向性价值,但差异足够大,Lila 不能只靠一个头部 TAM 叙事定价。
中点值只是展示辅助,不是权威出版方数字。图表用同一单位比较相邻类别价值区间,目的是显示差异,而不是拼出一个可直接相加的 TAM。
[CM011, CM012, CM013, CM015, CM016, CM017]2.3 买家、用户、付款方和初始可服务市场
最清晰的商业买家群体是制药和生物技术研发。Mordor 称,制药和生物技术公司在 2025 年占实验室信息学支出的 53.14%;Thermo Fisher 的 2024 年收入结构显示,57% 收入来自制药和生物技术客户。CRO 是下一个最相关的细分,因为它们明确出现在实验室自动化终端用户名单中,也是增长较快的信息学群体之一。学术和政府实验室对技术验证、方法开发和参考账户很重要——NIH 表示,其近 US$48 billion 预算支持超过 2,500 家机构的近 50,000 项竞争性资助——但该采购基础分散,通常不太可能支持完整工厂式企业合同。材料、化学和工业研发具有战略意义,因为自驱动实验室文献在这些领域最强,Thermo 和 Agilent 也都强调先进材料和应用实验室工作流;但公开市场报告没有清晰拆分这些项目。在集成部署中,实际买家通常是平台研发、药物化学、筛选、自动化工程或实验室运营负责人;用户则是台架科学家、自动化工程师和计算科学家。付款方通常是中央研发预算负责人,他们能用吞吐量、周期压缩或可重复性来为自动化支出辩护,而不是只讲 IT 现代化。[CM020, CM021, CM022, CM023, CM024, CM025]
| 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 |
|---|---|---|---|---|---|---|
| 大型制药研发 | 发现平台、药物化学或转化科学负责人 | 实验科学家、自动化工程师、计算化学家 | 中央研发预算 | 靶点识别、筛选、先导优化、DMTA 循环 | 研发高级副总裁或平台负责人 | 研发周期长,团队需要提高产出、通量和项目筛选质量 |
| 新兴生物技术公司 | 研究副总裁、CSO,或平台生物 / 化学负责人 | 小型跨学科实验室团队 | 项目预算或公司级研发预算 | 人手有限,需要更快生成假设 | CSO 或研究副总裁 | 每位科学家需要跑更多实验,并压缩里程碑周期 |
| CRO / CDMO 发现服务 | 发现业务的站点或业务单元负责人 | 检测团队、自动化人员、项目经理 | 与客户项目绑定的运营预算 | 高通量检测执行和外包筛选 | 总经理或运营负责人 | 在守住利润率的同时,提高通量和利用率 |
| 学术与政府研究 | PI、核心设施主任或中心负责人 | 研究生、博士后、核心设施员工 | 资助经费或机构资本预算 | 方法开发、筛选、转化研究 | PI、研究所主任或共享仪器委员会 | 资助项目需要新能力或更高可复现性 |
| 材料 / 化学 / 工业研发 | 先进材料、配方或应用研究负责人 | 科学家、机器人专家、数据科学家 | 创新预算或业务单元研发拨款 | 催化剂、聚合物、电池或配方优化 | CTO、创新副总裁或应用研究负责人 | 需要缩短从发现到规模化的周期,并提高可复现性 |
| 诊断 / 应用实验室 | 实验室主任或运营负责人 | 技术人员和工作流经理 | 实验室运营或质量预算 | 样本处理、数据完整性和受监管工作流 | 实验室主任或质量负责人 | 需要提高通量、降低错误率,但可能购买比 Lila 全栈更窄的系统 |
不同组织的买方头衔不一样;稳定模式是商业赞助人贴近实验通量,IT 更多是赋能方,而不是唯一预算负责人。
[CM020, CM021, CM022, CM023, CM024, CM025]最好的初始 SAM,是预算集中、数据准备真实、且 ROI 能绑定项目产出而非分散科研资助的细分市场。
这个矩阵是基于留存买方、预算和工作流证据得出的优先级观察口径;它不同于 TM003 中按角色拆分的运营地图。
[CM020, CM021, CM022, CM024, CM026, CM043]2.4 增长驱动因素、采用约束和碎片化竞争
Lila 所处市场的采用逻辑建立在生产率压力上。高通量筛选、劳动力短缺和减少人工错误的需求,持续把实验室推向自动化工作流。在信息学领域,合规、可审计性和云原生数据处理正在迫使实验室升级记录系统。在 AI 药物发现领域,买家受发现本身的成本和时间负担驱动:Mordor 强调压缩多年发现周期的压力,并引用商业化一个分子的平均 US$2.6 billion 成本。但约束同样清晰可见。市场报告和自驱动实验室综述反复指向遗留系统集成、碎片化仪器资产、高前期成本、实施负担和互操作性弱。Bruker/Chemspeed 发布也从供给侧强化了同一点:异构实验室仍受困于孤立工具和集成缺口。市场还面临可信度风险。STAT 2024 年报道引用 Insitro 的 Daphne Koller 警告称,人们期待“明天”就有突破,提醒投资者热情可能跑在实际部署之前。竞争格局不是赢家通吃,而是碎片化:实验室自动化由 Thermo、Danaher、Agilent、Tecan 和 Roche 等在位者主导;AI 发现有自己的软件群体;自驱动实验室创业公司仍在由仪器、信息学和服务定义的更广泛栈内竞争。[CM027, CM028, CM029, CM030, CM031, CM032]
| 驱动因素 / 约束 | 方向 | 时间 | 含义 | 尽调问题 |
|---|---|---|---|---|
| 高通量筛选和实验量 | 顺风因素 | 当前 | 支撑工作流自动化和集成执行层预算 | 询问哪些客户工作流能把通量扩到足以支撑工厂式部署 |
| 制药和生物技术发现生产率压力 | 顺风因素 | 当前 | 压缩周期、优先排序实验具备经济价值 | 要求证明 Lila 能减少迭代周期或提高候选物质量 |
| 云原生数据底座和合规现代化 | 顺风因素 | 当前 | 拉动可支撑编排和模型训练的信息学层需求 | 检查 Lila 如何接入既有 LIMS/ELN 和受监管数据环境 |
| AI 辅助靶点识别和设计 | 顺风因素 | 2026-2031 | 快速增长的软件预算可能把需求拉向集成湿实验室执行 | 衡量 Lila 是卖进既有 AI 发现预算,还是需要新增预算科目 |
| 遗留系统集成和异构仪器 | 逆风因素 | 当前 | 推高真实实验室的实施成本,并拖慢见效时间 | 梳理 Lila 哪些仪器和数据系统开箱即支持 |
| 前期投入高、ROI 不清 | 逆风因素 | 当前 | 没有清晰回报时,小型实验室和部分工业项目可能推迟采用 | 索取回本周期、利用率指标和部署人工要求 |
| 数据质量、安全和监管信任 | 逆风因素 | 当前 | 治理薄弱时,即使试点看起来不错,也会挡住生产部署 | 审查审计轨迹、QA 工作流和模型治理控制 |
| 炒作风险和漫长企业销售周期 | 逆风因素 | 当前 | 市场预期可能比真实生产采用更快膨胀 | 收集已从试点走向规模化重复使用的客户证明 |
含义综合了分析师市场总结、技术评审和行业报道;它们有助于安排尽调优先级,但不能替代 Lila 自身部署证据。
[CM027, CM028, CM029, CM030, CM031, CM032]企业采用需要实验闭环、数据底座,以及足够的 ROI 证据,才能把试点接到规模化部署。
这条流程有证据支撑,但仍是定性描述。它描述的是反复出现的采购和部署路径,而不是数值漏斗。
[CM027, CM029, CM030, CM032, CM034, CM035]2.5 对估值真正重要的规模和采用尽调缺口
核心承销问题不是市场是否存在,而是 Lila 是否在货币化正确的切片。公开证据不支持自主实验室或 AI science factories 的干净、独立 TAM。它支持一些相邻且已有资金支持的类别,可被拼接成商业化投资逻辑;其中制药和生物技术是最可辩护的第一切入点,材料发现则是战略上重要但更难量化的第二切入点。这意味着估值工作需要来自 Lila 的自下而上商业证据,而不是更多自上而下市场报告。关键问题很直接:按客户类型划分的当前 ACV,软件、自动化与服务的组合,实施周期,续约行为,以及客户是否真的能从一个工作流扩展到更广泛工厂模式的证据。没有这些数据,广义 TAM 可以解释对该类别的兴趣,但不能支撑对 Lila 份额获取或利润率结构的确信。[CM018, CM019, CM039, CM043, CM044, CM045]
2.6 图表
03竞争格局
3.1 直接挑战 AI science factory 投资逻辑的集成型对手
即便放在 AI for science 内部,Lila 的公开叙事也异常激进。公司称自己正在为科学构建一个通用操作系统,可自主生成假设、设计实验、运行实验,并在生命、化学和材料科学中从结果学习。这让 Recursion 加 Exscientia、Insilico Medicine 和 Isomorphic Labs 成为最接近的直接竞争组,但原因各不相同。Recursion 已经把大型自有生物学和化学数据集、自动化湿实验室和模型驱动设计结合起来,Exscientia 交易又加入精准化学和自动化合成,使其最接近全栈小分子药物发现对手。Insilico 在疗法领域也明确端到端,但其公开表述仍是 Pharma.ai 和从 A 到 Z 的管线创建,而不是通用科学操作系统。Isomorphic 同样重前沿模型,并通过制药合作拥有良好分发,但其公开叙事仍是数字生物学和分子设计,而不是跨领域实验自主性。因此,直接对手图谱是真实的,但仍窄于 Lila 的主张:大多数直接同行销售 AI 赋能的疗法发现,而 Lila 主张跨多个科学领域的自主科学工厂。[CP001, CP002, CP003, CP005, CP006, CP007]
| 竞争对手 | 类别 | 规模 / 融资信号 | 目标细分市场 | 差异化 | 局限 |
|---|---|---|---|---|---|
| Lila Sciences | 参照公司 / AI 科学工厂 | 启动时承诺种子轮 $200M;公开愿景覆盖生命、化学和材料科学 | 寻求一套自主发现栈的研究人员、制药公司和科学项目 | 通用自主科学平台横跨多个领域,从假设生成一直覆盖到实验执行 | 公开材料未披露具名客户、通量指标或商业定价 |
| Recursion / Exscientia | 直接集成型 TechBio 对手 | 2024 年二季度合计现金约 $850M;上市公司平台;预计 18 个月内约 10 项临床读数 | 重视 AI 赋能小分子发现且需要湿实验室规模的生物制药团队 | 规模化生物学探索,加上 Exscientia 精准化学和自动化合成 | 公开叙事仍集中在小分子疗法,而不是更广的科学工厂领域 |
| Insilico Medicine | 直接 AI 药物发现对手 | 平台覆盖靶点识别到 II 期;按 2021 年销售额计,与前 20 大药企中的 10 家合作 | 需要 AI 发现疗法和可合作管线资产的生物制药团队 | 明确的 A-to-Z AI 药物发现管线,配有自动化和合作验证 | 公开范围偏疗法,比 Lila 跨领域自主主张更窄 |
| Isomorphic Labs | 前沿模型药物设计对手 | Lilly 预付款 $45M,里程碑最高 $1.7B;2025 年新闻页列出 $600M 外部投资轮 | 希望借合作获得 AI 优先分子设计能力的大型制药发现团队 | 源自 AlphaFold 的数字生物学栈,以及顶级药企伙伴准入 | 合作驱动的商业模式清晰,但公开材料更强调药物设计,而不是自主湿实验室执行 |
| Benchling | 邻近 / 基础设施替代品 | 获 1,200+ 家生物技术组织信任;声称有数千项实施 | 将发现、临床前和工艺开发工作流数字化的研发组织 | 深度嵌入的信息学层,带 AI 工具、集成和端到端工作流支持 | 不声称自主运行科学方法,也不拥有完整湿实验室闭环 |
| Arcadia Science | 开放科学替代品 | 2021 年成立,拥有专门的研究、软件和实验室运营团队 | 被开放工具、协议和社区导向研究资产吸引的科学家 | 重新思考研究周期,同时把工具和管线回馈社区 | 开放科学姿态不等于工业化端到端自主执行 |
| OpenBioML + Opentrons | 开放 / 模块化栈替代品 | OpenBioML 有工业级算力支持;Opentrons 销售可重构自动化硬件 | 选择开放模型加灵活自动化、而不是单一封闭供应商栈的实验室 | 开放协作、公开代码库和无锁定的模块化自动化 | 需要集成工作,也没有呈现统一发现 P&L 或已验证的跨领域工厂 |
| 制药公司内部 AI 项目 | 现状 / 内部自建替代品 | Genentech 提到数十年的实验室和临床数据,以及 NVIDIA 支持的生成式 AI;AstraZeneca 引文显示自动化建立在中立基础设施上 | 更愿意把发现能力留在内部的大型制药研发组织 | 数据、科学家、预算和分发能力已经嵌在买方组织内部 | 资本和集成负担高,且不同药企公开细节不均衡 |
规模单元格只使用已保留抓取来源中的公开证据。客户牵引、定价或通量不公开时,表格直接标出可见性缺口,而不做估算。
[CP001, CP002, CP004, CP008, CP009, CP010]按纵向整合度与买方触达 / 分销能力,对最相关竞争者类别做顺序型定位。
坐标轴是分析师基于留存公开证据给出的顺序评分,依据是整合度、自动化、商业模式和买方触达,而不是已发布的基准数据集。
[CP001, CP005, CP008, CP014, CP017, CP019]3.2 模块化软件、自动化和开放科学替代方案
更危险的替代集合不只是直接 AI 药物发现玩家。Benchling、Opentrons、OpenBioML 和 Arcadia 展示了 Lila 集成投资逻辑之外的模块化替代路径。Benchling 提供企业研发软件、端到端流程跟踪、AI 工具、集成和实施规模,但不声称自主运行科学方法。Opentrons 同样营销可重配置自动化和摆脱封闭系统的明确自由,使其成为湿实验室层替代品,而非 science-factory 所有者。OpenBioML 在模型和社区层扩展了替代图谱:其开放研究实验室定位、公开代码库和有计算支持的合作显示,生物 AI 的一部分可以在开放生态中搭建,而不必放进自有垂直栈。Arcadia 则从另一个侧翼发力:把工具、协议和软件管线发布回社区,同时试图重构研究周期。这些努力单独都不能复现 Lila 的完整主张,但合在一起描述了一条可行的自组装路径:买家组合数据基础设施、自动化硬件,以及开放或伙伴驱动模型,而不是采用一个封闭工厂。[CP020, CP021, CP022, CP023, CP025, CP026]
| 购买标准 | Lila | Recursion / Exscientia | Insilico | Isomorphic | Benchling | 开放 / 模块化栈 | 制药公司内部自建 | 备注 |
|---|---|---|---|---|---|---|---|---|
| 跨领域科学范围 | 强 | 中 | 低 | 低 | 低 | 中 | 中 | Lila 明确覆盖生命、化学和材料科学,而多数直接对手首先销售的是疗法发现 |
| 自动化湿实验室反馈闭环 | 强(声称) | 强 | 中 | 部分 / 未公开 | 低 | 中 | 强 | Recursion 和 Genentech 给出具体的实验室在环描述;Isomorphic 公开材料更聚焦模型和合作 |
| 小分子药物设计深度 | 中 | 强 | 强 | 强 | 低 | 低 | 强 | Recursion-Exscientia、Insilico 和 Isomorphic 的公开小分子定位证据都强于 Lila |
| 企业信息学和集成层 | Unknown | 中 | 中 | 低 | 强 | 中 | 强 | Benchling 在工作流、数据模型和集成上最强;制药公司内部自建可以把这一层与内部系统结合 |
| 开放 / 可扩展工具姿态 | 低 | 低 | 低 | 低 | 中 | 强 | 中 | OpenBioML 和 Opentrons 是最清晰的反锁定替代品 |
| 制药分发 / 买方准入 | Unknown | 强 | 强 | 很强 | 强 | 低 | 很强 | Isomorphic、Recursion 和制药公司内部项目拥有最清晰的大型药企准入信号 |
| 商业可见度 | 低 | 中 | 中 | 中 | 中 | 开放性高,集成商业责任低 | 内部可见度高 | Lila 在准入模式、客户和通量上的公开信息最不清晰 |
单元格比较的是公开证据质量,不是绝对技术事实。「未知」表示本来源集没有浮现足够直接的公开证据,无法有把握地给该标准打分。
[CP001, CP005, CP007, CP010, CP013, CP016]紧凑热力图,显示哪些竞争者类别能替代 Lila 投资逻辑中的模型、湿实验室、信息化、开放性和制药公司触达层。
标签按能力层概括留存公开证据,而不是供应商验证过的基准分数。“Unknown” 表示公开证据缺失,不代表没有能力。
[CP020, CP021, CP029, CP030, CP031, CP037]3.3 分发能力和制药内部自建路径
Lila 最难的竞争战场可能是分发和访问,而不是原始技术野心。Isomorphic 的公开证据显示,它通过 Novartis、Lilly 和 Johnson & Johnson 走合作伙伴主导路线,并包含很大的里程碑经济权益。Recursion 和 Exscientia 也强调与主要交易对手组成的制药合作组合及里程碑潜力,这意味着大型生物制药买家可以通过成熟联盟模式获得 AI 赋能发现,而不必采用新的通用平台。Benchling 的客户证据又提供另一条路:大型研发组织可以在中立软件和自动化层上升级内部科学运营,而不把控制权交给单一 science-factory 供应商。Genentech 自己的 lab-in-a-loop 叙事让替代类别更尖锐。如果大型制药公司能组合自有数据、内部科学家、湿实验室基础设施,以及外部计算或软件伙伴,那么分发优势就在现有研发组织内部的嵌入式项目。放在这个背景下,Lila 的公开材料在商业化、外部客户和吞吐量上仍相对不透明。这并不否定技术故事,但确实让上市路径比周边那些重合作伙伴和内部自建的替代方案更不清晰。[CP004, CP014, CP017, CP018, CP019, CP022]
| 竞争者类别 | 公开准入或定价姿态 | 合同 / 包装模式 | 包含能力 | 未知项或折扣模式 | 含义 |
|---|---|---|---|---|---|
| Lila Sciences | 未保留公开标价 | 可能通过企业、伙伴或项目方式接入科学工厂 | 通用自主科学平台,加自有实验室基础设施 | 已保留来源未公开具名客户、计价单位和合同结构 | 商业就绪度比技术故事更不清晰 |
| Recursion / Exscientia | 未保留公开软件式价目表 | 上市公司平台,加合作项目和里程碑经济 | 规模化生物学、精准化学、自动化合成、转化和管线资产 | 经济条款主要通过并购和合作披露可见,而非标价 | 以平台加项目公司竞争,不是透明的基础设施软件 |
| Insilico Medicine | 未保留稳定公开标价 | 平台、管线和合作 / 授权模式 | AI 靶点发现、分子设计、自动化和治疗项目 | 公开来源强调管线阶段和合作,而不是标准席位或使用费 | 最适合作为疗法引擎比较,而不是 SaaS 费用项 |
| Isomorphic Labs | 未保留开放平台定价 | 与大型药企研究合作,包含预付款和里程碑经济 | 围绕伙伴选定项目开展 AI 优先分子设计和靶点工作 | Lilly 交易经济条款公开,但更广商业条款是定制的 | 分发能力强,但准入集中在伙伴关系中 |
| Benchling | 报价驱动的企业软件 | 实施驱动的信息学订阅 / 平台模式 | 电子实验记录本、数据模型、工作流自动化、样本和流程管理 | 已保留来源显示范围和客户证明,但没有稳定标价 | 买方想要基础设施、而不是外包科学时,Benchling 是最清晰的模块化替代品 |
| 开放 / 模块化栈 | 开放或按组件定价 | 开源模型 / 社区,加硬件和软件采购 | 公开代码库、开放协作、模块化自动化硬件和工作流软件 | 集成成本由买方承担,且整个栈没有标准化 | 锁定效应更低,但集成负担高得多 |
| 制药公司内部自建 | 内部预算科目,不是外部标价 | 既有研发预算内的资本开支、算力、软件和科学家时间 | 实验室在环 AI、内部数据、科学家,以及中立软件或算力伙伴 | 公开支出细节稀少,ROI 取决于内部采用和治理 | 药企有规模自建时,这是独立外部工厂最危险的替代品 |
本表比较准入模式和经济包装,因为已保留公开来源没有给出多数竞争对手的稳定标价。未知项明确列出,而不是估算。
[CP004, CP009, CP013, CP014, CP017, CP018]3.4 护城河耐久性和反向证据
公开反向证据不支持把自主科学类别视为已经定局。SLAS 2026 市场图谱描述至少 15 家公司在争夺实验室操作系统层,这意味着编排、集成和闭环自动化正在许多供应商之间碎片化。Royal Society 综述更直接:自驱动实验室在某些场景下几乎能自动化完整科学方法,但真正完全自主的 Level-5 AI 科学家尚未实现。UChicago 的 AI-advisor 表述认为,领先实践者仍希望人类共享控制,而不是从闭环中消失。Northwestern 的 megalibrary 工作展示了材料领域的另一项特定挑战:对某些发现问题,大规模并行筛选可跑赢迭代式自驱动实验室路径,因此 Lila 的材料投资逻辑可能面对不使用同一工厂模型、但数据丰富的替代方案。含义是,Lila 的护城河不能只靠声称自己垂直整合或自主。它必须证明,跨领域闭环执行能创造比更窄疗法栈、模块化实验室系统或制药内部项目更好的经济性或更好的发现。在公开客户、吞吐量和结果数据出现之前,护城河耐久性更像战略论点,而不是已证明的市场锁定。[CP032, CP033, CP034, CP035, CP036, CP041]
| 护城河主张 | 威胁 | 严重性 | 证据支持的理由 | 缓释措施 / 尽调问题 |
|---|---|---|---|---|
| 一个通用 AI 科学工厂 | Recursion / Exscientia 已经像一个全栈小分子对手 | 高 | Recursion 的自动化生物学加上 Exscientia 的自动化化学,是公开市场中最接近的全栈疗法类比 | 要求证明 Lila 的跨领域栈比仅做疗法的垂直方案带来更好结果 |
| 跨领域范围具有独特价值 | 买方可能更偏好药物发现、信息学或材料领域中更窄但已验证的栈 | 高 | 直接对手范围更窄但更清晰,模块化替代品也让买方只为需要的层付费 | 按买方类型和领域索取赢单 / 输单数据,证明跨领域宽度在哪里具备商业意义 |
| 自主执行创造持久锁定 | 领先公开文献和研究人员仍偏好人在环自主 | 中 | Royal Society 称 Level-5 全自主系统尚未实现,UChicago 提出共享控制 | 索取无人值守循环、错误率,以及人类必须介入时点的证据 |
| 封闭系统优于开放工具 | Benchling、Opentrons 和 OpenBioML 提供反锁定替代方案 | 高 | 开放集成、模块化自动化和开放模型社区,会削弱“必须由一家供应商拥有全栈”的论证 | 量化相对模块化栈的切换成本和集成收益 |
| 数据飞轮难以复制 | 药企内部项目已经沉淀了数十年的实验室和临床数据 | 高 | Genentech 的实验室闭环和 NVIDIA 支持的平台表明,有规模的买方可以把数据和分发留在内部 | 要求证明外部客户能从池化学习中受益,而这种收益是其内部无法复制的 |
| 材料科学自主化是清晰切入口 | 在部分材料工作流中,巨型库这类并行搜索平台可能优于迭代式自驱实验室 | 中 | Northwestern 认为,针对某些材料问题,巨型库能比自驱实验室更快生成数据和候选物 | 将 Lila 的材料工作流与巨型库或高通量并行筛选替代方案做基准对比 |
| Lab OS 所有权会走向整合 | SLAS 2026 证据显示,编排层竞争拥挤,供应商众多 | 高 | Drug Discovery Trends 梳理出至少 15 家公司在争夺 AI 赋能实验室操作系统层 | 说明 Lila 拥有哪些差异化层,能在编排商品化后继续存在 |
严重性估计的是 Lila 竞争地位可能承受的压力,而不是损失的确定性。若干威胁属于替代或分发风险,并非一对一功能替换风险。
[CP008, CP009, CP032, CP033, CP034, CP035]公开记分卡,跟踪最可能决定 Lila 科学工厂定位能否变成持久护城河的维度。
[CP008, CP039, CP041, CP042, CP043, CP044]3.5 图表
04财务情况
4.1 收入模式、定价和商业准备度
公开证据显示,Lila 的货币化更可能是软件访问、科学项目收入和付费工厂产能的混合,而非纯 SaaS 供应商。官方材料称,新资本将用于把平台带给客户,并且首批商业伙伴正在被接入。Reuters 补充称,Lila 计划出售面向企业的软件访问权,使客户使用其 AI 模型和自动化实验室;Sacra 则描述了一种以项目制发现项目为核心、未来叠加 lab-as-a-service 或按使用量计费层的商业模式。对一家在生命科学、材料、能源和半导体领域运营机器人湿实验室的公司而言,这种组合在经济上说得通,但收入质量尚未得到证明。已审阅的官方或市场数据来源没有披露标价、标准合同条款、ACV、收入组合或已命名付费客户。结果是一条可信的商业化路径,但公开货币化披露很弱:投资者能看到 Lila 想卖什么,却还看不到可用于承销经常性收入质量的条款、客户证据或可重复性。[CI002, CI003, CI004, CI009, CI010, CI011]
| 收入流 | 机制 | 单位 | 当前价值 / 状态 | 收入质量 | 尽调问题 |
|---|---|---|---|---|---|
| 面向合作伙伴 R&D 的发现项目 | 客户带来科学问题,Lila 围绕该问题运行 AI 指导的实验项目 | 按项目 / 里程碑 | Sacra 称其为当前最清晰的商业模式;官方来源暗示有合作伙伴项目,但未披露合同 | 潜在意义较大,但在可重复性得到证明前,经济属性更接近高价值服务 | 要求提供付费项目数量、ACV、里程碑组合,以及续约 / 扩张模式 |
| 企业软件访问 | Reuters 称 Lila 计划通过企业软件提供其 AI 模型和自动化实验室访问 | 按组织 / 席位 / 平台合同 | 商业计划已披露;没有公开定价或客户名称 | 如果能与工厂工作拆开,可能支撑经常性收入,但捆绑风险未知 | 要求提供 SKU 结构、部署模型、合同最低额,以及纯软件收入占比 |
| AI Science Factory 产能 | 将自动化实验室吞吐作为实验产能或托管访问出售 | 按运行 / 时段 / 使用量块 | 根据官方工厂建设和 Sacra 的实验室即服务描述推断出的计划模型 | 如果标准化且利用率高,可能很好地变现固定资产;若高度定制则较差 | 要求提供计费单位、吞吐假设、利用率目标,以及各工厂贡献利润率 |
| 首批商业合作伙伴 | 官方文章称 Lila 正在迎接第一批客户 | 试点 / 付费试点 / 早期合同 | 商业化已经启动,但没有公开具名客户或参考账户 | 在试点转为可重复的付费使用并产生可衡量 ROI 前,质量偏低 | 要求提供具名客户、试点转付费转化率,以及参考账户经济性 |
| 跨领域科学合作 | 平台面向生命科学、化学、材料、能源、半导体和创业公司营销 | 联合项目 / 企业协议 | 目标行业公开;各行业已签收入未公开 | 更大的 TAM 可分散需求,但每个行业可能需要不同销售动作和支持 | 要求按行业和合同类型提供管线、胜率和收入结构 |
各行捕捉截至 2026-06-02 公开可见的变现路径;它们描述机制和披露状态,而非已实现收入结构。
[CI009, CI010, CI011, CI012, CI013, CI014]| 界面 | 价格 / 单位 / 合同 | 标价 vs. 实际成交价 | 折扣 / 未知项 | 来源 |
|---|---|---|---|---|
| 官方客户入驻措辞 | 未公布公开价格 | 缺少标价 | 未公开最低额、试点、补差结算或续约条款 | Lila 官方文章和主页 |
| 企业软件访问 | 计划提供软件访问;定价未披露 | 缺少实际定价 | 未知是否按席位、组织、工作流、模型使用量定价,或与实验室访问捆绑 | Reuters 经 Yahoo Finance |
| 发现项目工作 | 已描述项目制模式,但没有公开费率表 | 未发现标价 | 里程碑安排、范围蔓延和科学成功经济性均未知 | Sacra |
| 实验室即服务 / 使用访问 | 描述了未来订阅或使用量基础,但未公布价格表 | 分析师描述,并非官方费率卡 | 计费单位、最低承诺和利用率传导均未知 | Sacra |
| 首批客户合同 | 商业发布措辞暗示已有合同,但条款未披露 | 未披露 | 具名客户、定价、期限和 ROI 证据均缺失 | 官方文章和 Reuters |
本表将可见商业化界面与缺失商业条款分开;空白经济数据应被视为披露缺口,而非零值。
[CI010, CI011, CI012, CI014, CI029]公开来源显示,Lila 正从隐身研发走向混合模式,可能结合项目收入、软件访问和工厂产能;但定价和具名客户证据仍让这座桥断开。
该图为定性图,因为审阅的公开来源没有披露客户数量、ACV 或按产品拆分的已实现收入。
[CI009, CI010, CI011, CI012, CI013, CI014]4.2 成本结构和单位经济代理指标
尽管没有 P&L,可能的成本结构已经可见。Lila 正在多个地理区域建设 AI Science Factories,Reuters 称其签下 235,500 平方英尺 Cambridge 实验室租约,当前招聘材料显示多站点设施领导、大额资本规划、重型设备搬迁、工艺气体、水和空气系统、废水处理以及合规工作。招聘网站同时显示,公司在前沿 AI、实验室运营、产品、合作伙伴和企业销售上激进招聘。Flagship 与 AWS 的合作也指向湿实验室资本开支和科学劳动力之外的可观云与计算支出。合在一起,该模型比典型软件创业公司更依赖资本和利用率。已审阅公开来源没有披露毛利率、CAC、回本周期、留存、利用率或收入成本细节,因此投资者无法判断公司会获得类软件贡献利润率,还是落入需要更高固定成本吸收的服务加产能业务。公开可用的主要代理指标是方向性的:如果工厂利用率、重跑率和标准化不能快速改善,租约、设备、计算和专业人员组合可能会重压利润率。[CI016, CI017, CI018, CI019, CI020, CI021]
| 指标 | 数值 / 状态 | 置信度 | 重要性 | 尽调问题 |
|---|---|---|---|---|
| 收入 / ARR | 低 | 判断估值支撑和商业速度所必需。 | 要求提供月收入、ARR、积压订单,以及按软件、项目和产能销售划分的收入结构 | |
| 毛利率 | 低 | 决定工厂经济性是否有可能接近软件利润率。 | 要求拆分收入成本:实验室运营、云 / 计算、试剂、支持和折旧 | |
| CAC / 回本期 | 低 | 需要理解企业软件和合作伙伴驱动销售能否高效扩张。 | 要求按客户细分提供全口径 CAC、回本期和销售周期数据 | |
| 产能利用率 | 低 | 高固定成本实验室需要吞吐来吸收租赁、设备和人员开销。 | 要求提供各工厂利用率、重跑率、闲置时间和吞吐量 | |
| 固定成本基础代理指标 | 剑桥大型实验室租约,加上多地点设施预算和扩张岗位 | 中 | 表明公司在收入公开前已经背负有意义的实体运营足迹。 | 要求提供年度租赁费用、资本开支计划,以及按地点划分的设施运营开支 |
| 计算 / 云强度代理指标 | AWS 支持加上大量 ML / AI 招聘,显示基础设施支出有意义 | 中 | 科学 AI 经济性同时取决于湿实验室吞吐和计算效率。 | 要求提供云支出、模型训练预算和每个项目的推理成本 |
| 运营复杂度代理指标 | 岗位描述提到气体、水、空气系统、废水、重型设备、装卸码头和合规 | 中 | 这些公用工程和安全要求会抬高维护、停机和合规成本。 | 要求按工厂提供公用工程支出、停机率和维护预算 |
空值代表公开财务披露不可得;代理行只是方向性运营信号,并非公司报告的 KPI。
[CI017, CI018, CI019, CI020, CI021, CI023]Lila 的融资缓冲很厚,但在公开收入可衡量之前,现金消耗很可能被工厂建设、多地点设施、科学人才和云 / AI 基础设施拉动。
只有融资节点带有公开数值;下游成本节点均为定性,因为公司未按类别披露烧钱速度、资本开支或运营开支。
[CI001, CI004, CI017, CI018, CI019, CI020]从科学突破到持久收入的路径,仍取决于具名客户、标准产品、工厂利用率,以及尚未公开的利润率披露。
所有节点均为定性,因为审阅的公开记录没有披露收入、利用率、CAC、回本周期或毛利率数值。
[CI024, CI025, CI026, CI027, CI028, CI029]4.3 资本充足性和融资依赖
资本充足性是公开记录中最强的一块。Flagship 在 2025 年 3 月带着 $200M 已承诺种子资本发布 Lila;公司随后披露 9 月 $235M Series A 首次关闭,并在 10 月进一步扩展 $115M,使该轮达到 $350M,已披露总资本达到 $550M。Bloomberg 将 9 月估值锚定在约 $1.23B,Reuters 称扩展轮把 Lila 推至 $1.3B 以上,Forge 后来展示了 $1.42B 的 Series A 估值快照。该融资栈大幅降低近期救援融资风险,也给 Lila 建实验室、招聘和测试商业化留出空间。但它并未消除融资依赖。由于公开来源未披露收入、现金、烧钱速度、毛利率或客户集中度,投资者无法判断公司正以多快速度把资本转化为耐久运营基础。SEC 和 NASAA 针对 AVSF - Lila Sciences 2025, LLC 的 Form D 也显示,2025 年融资过程中至少涉及一个 feeder 或辛迪加工具,强化了该轮融资广泛且结构化,而不是简单双边风险投资。已审阅公开来源没有披露债务设施或项目融资义务。[CI001, CI002, CI003, CI004, CI005, CI006]
| 指标 | 公开数值 / 状态 | 置信度 | 重要性 | 尽调问题 |
|---|---|---|---|---|
| 种子融资 | 2025 年 3 月承诺的 $200M 种子资本 | 高 | 在商业化前形成异常庞大的资本基础。 | 确认总募资额、交割时间,以及是否按地点或项目指定用途 |
| Series A 首次交割 | 2025 年 9 月 $235M,由 Braidwell 和 Collective Global 共同领投 | 高 | 建立外部投资者验证和初始独角兽估值。 | 确认一级资本、交割日期和董事会 / 治理条款 |
| Series A 延展 | 2025 年 10 月 $115M,包括 NVentures / Nvidia | 高 | 引入战略资本,并进一步延长扩张预算。 | 确认一级 vs 二级拆分,以及延展投资者附带的任何战略权利 |
| 总募资额 | 种子轮和 Series A 合计 $550M | 高 | 显著降低近期救援融资压力。 | 确认扣费后新增净现金和当前无限制现金余额 |
| 估值锚点 | 9 月约 $1.23B,10 月 >$1.3B,2026 年 Forge 上为 $1.42B | 中 | 为未来商业证明设定投资者预期。 | 要求提供官方投后估值、股数、清算顺位,以及当前 409A / 优先股标记 |
| 备案 / 辛迪加结构 | SEC / NASAA Form D 显示 AVSF - Lila Sciences 2025, LLC 有一笔 $817,500 集合基金发行 | 高 | 暗示至少一个联接基金或辛迪加载体参与了融资流程。 | 澄清哪些投资者通过 SPV 或联接基金进入,以及权利是否不同于直接持有人 |
| 债务 / 项目融资义务 | 未有经审阅公开来源披露债务融资或项目融资义务 | 低 | 没有披露债务会简化可见资本结构,但公开沉默并不证明没有债务。 | 要求提供债务明细、设备融资、租赁负债和表外承诺 |
资本事实反映截至 2026-06-02 的公开融资公告和市场数据快照;估值标记是锚点,不是经审计公允价值。
[CI001, CI002, CI003, CI004, CI005, CI006]Lila 的公开财务锚点在融资和估值上很丰富,在经营表现上几乎缺位,凸显融资披露领先于业务披露。
估值数值来自新闻和二级市场平台的公开锚点,并非审计后的公允价值标记;估值项使用低 / 中 / 高区间来显示来源之间的分散度。
[CI001, CI002, CI003, CI004, CI005, CI006]4.4 财务结论和尽调阻断点
从财务上看,Lila 筛选结果是资金异常充足、战略野心很大,但在最关键运营指标上仍处于证据前阶段。上行情景很容易理解:罕见投资者辛迪加已经为建设买单,公司正从生物技术扩展到能源、半导体和材料,并终于从隐身转向首批客户商业化。下行情景同样清晰。Fierce 指出,Lila 尚未公开发布支持若干突破主张的数据;怀疑分析认为,只有当工厂吞吐量、标准化和客户转化可被衡量时,该模型才会在经济上成立。由于公开记录缺少收入、实际定价、毛利率、烧钱速度、利用率和参考账户,下一个承销里程碑不是另一则融资公告,而是证明首批客户能转化为可重复付费项目,或带有可接受单位经济的“软件加产能”合同。在证据出现之前,正确的财务立场是看好资本充足性,但对收入质量、利润率路径,以及 science-factory 叙事变成真实业务的速度保持谨慎。[CI009, CI011, CI012, CI024, CI025, CI026]
| 缺失的私有指标 | 重要性 | 公开替代项 | 对判断的影响 | 精确尽调路径 |
|---|---|---|---|---|
| 具名客户和合同价值 | 决定首批需求是否真实、付费且可重复 | 只有首批客户措辞和未具名行业兴趣公开 | 没有参考账户,商业化仍停留在前景判断 | 要求提供客户名单、ACV、试点转付费转化率和三通参考电话 |
| 收入 / ARR / 收入结构 | 需要检验估值是否由商业牵引支撑 | 公开来源披露融资和估值,而非运营收入 | 无法严谨分析市销率或烧钱倍数 | 要求按软件、项目和产能收入提供月度收入桥 |
| 毛利率和收入成本 | 需要区分可扩展软件经济性与定制实验室服务 | 只有运营代理指标:租赁、公用工程、设备和云需求 | 使利润率路径和长期盈利能力仍属推测 | 要求提供 COGS 拆分、折旧政策、支持成本,以及按产品划分的利润率 |
| 现金余额、烧钱和现金跑道 | 需要判断 $550M 资本栈按当前扩张速度能撑多久 | 已募资本公开;现金部署不公开 | 无法精确承销现金跑道和下一轮融资时间 | 要求提供当前现金、月度烧钱、计划资本开支和情景现金跑道 |
| 产能利用率和吞吐 | 利用率决定工厂能否吸收固定成本 | 行业评论和招聘暗示有大型固定资产,但没有公开运营指标 | 工厂经济性只能依赖叙事,而非实测产出 | 要求提供利用率、实验吞吐、重跑率、周期时间和积压订单 |
| 续约、留存和集中度 | 如果试点转为长期企业或平台合同,这些指标很重要 | 未公开续约、NRR、扩张或客户集中度指标 | 即便首批项目签约,也无法判断收入质量的耐久性 | 要求提供续约批次、扩张率、集中度和流失原因 |
这些是真正的尽调阻塞项,而非格式遗漏;每个缺失指标都会实质性改变估值和融资判断。
[CI014, CI015, CI024, CI025, CI027, CI028]05产品与技术
5.1 闭环科学引擎和产品表面
对一家年轻科学平台而言,Lila 的公开产品故事异常具体。公司把 Lila Iris 描述为一个在实验生成 token 上训练的科学推理模型,再用验证器、科学工具、计算和能产生真实世界奖励信号的 AI Science Factories 围绕该模型。换句话说,公开架构不是给科学家的聊天机器人;它是迭代发现的控制平面,让假设、实验设计、执行、解释和策略优化彼此反馈。面向买家的层把该引擎转译成两个商业动作。Catalyst 是平台访问产品:团队通过 Lab-as-a-Service 模式直接访问 Lila Iris、工厂产能和科学专家,把固定实验室资本开支转成按需吞吐量。Creation 是结果导向产品:Lila 运行项目,生成已验证资产、协议和数据包,并产出 IP 与去风险后的路线图。这个组合支持一种尽调判断:Lila 既是软件供应商,也是发现产能供应商,而不是纯模型供应商。[CE001, CE002, CE003, CE004, CE006, CE007]
| 模块 / 资产 | 主要用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| Lila Iris 科学推理模型 | 内部科学家和合作伙伴团队 | 核心叙事 / 公开描述 | 实验生成的科学 token,加上验证器和工具,而非仅用互联网数据训练 | 模型架构、训练数据量和基准方法不公开。 |
| AI Science Factories | Lila 运营人员和合作伙伴项目 | 核心建设 / 公开描述 | AI 控制下的可扩展仪器网络,提供真实世界奖励信号 | 具体仪器清单、检测类别、正常运行时间和设施利用率不公开。 |
| Catalyst | 企业 R&D 团队 | 商业访问模型 / 在线页面 | 直接平台访问,加上科学专家和实验室即服务经济性 | 具名客户、定价和运营 SLA 不公开。 |
| Creation | 战略合作伙伴和投资者 | 商业解决方案模型 / 在线页面 | 以结果为导向的活动,交付已验证资产、实验方案和数据包 | 项目经济性、收入分成和复购客户证据不公开。 |
| 治疗药物工作流 | 生物制药发现团队 | 活跃解决方案领域 / 明确工作流覆盖 | 在载荷、递送、安全性和可制造性之间做全栈优化 | 没有具名治疗药物客户或已发布项目结果。 |
| 生物技术工作流 | 生物工艺、试剂和检测团队 | 活跃解决方案领域 / 明确工作流覆盖 | 将设计智能体接到制造约束下的高通量执行 | 没有按用例公开附加率或部署数据。 |
| 化学工作流 | 化学和工业 R&D 团队 | 活跃解决方案领域 / 明确工作流覆盖 | 结合分子设计、模拟、高通量实验和反应器选择 | 没有用于催化剂胜出或周期时间改善的独立基准集。 |
| 材料与能源工作流 | 材料、能源和先进制造团队 | 活跃解决方案领域 / 明确工作流覆盖 | 覆盖涂层、吸附剂、无稀土磁体、电催化剂和其他硬资产问题 | 相比公开路线图的广度,独立证明仍然偏薄。 |
成熟度标签反映公开证据深度,而非内部收入结构或内部就绪度评审。
[CE001, CE006, CE009, CE010, CE011, CE012]Lila 平台的公开栈视图,从面向合作伙伴的交付模式,到科学推理、工具和仪器化工厂。
该技术栈根据技术、解决方案和团队页面重构,并非官方工程架构图。
[CE003, CE004, CE005, CE006, CE007, CE009]Lila 公开商业和技术页面隐含的闭环运营流程。
公开页面没有发布 BPMN 式流程图,因此该流程把反复出现的假设—设计—实验—分析表述转写为面向用户的运营序列。
[CE002, CE008, CE009, CE011, CE012, CE017]5.2 横跨生命科学、化学和材料的领域项目
领域图谱很宽,但围绕生命科学、化学和材料保持连贯。Lila 的疗法页面聚焦可编程遗传药物、递送载体、抗体和配体工程,以及对效力、特异性、持久性、安全性和可制造性的协同优化。生物技术页面把该逻辑延伸到构建体、宿主系统、表达平台、配方、试剂、测定和具有商业相关性的生产工作流。化学和能源页面把同一运营模型推进到催化剂发现、反应器选择、电催化剂、吸附剂、无稀土磁体、燃料和化学信息驱动分离。先进材料页面又加入薄膜涂层和基础设施组件,强化了同一平台意图在生物、化学和物理科学问题之间迁移。官方页面持续强调真实实验、已验证或经人类验证的数据,以及在制造或商业对齐条件下运行,显示 Lila 目标是产出可部署资产,而不是止步于虚拟筛选。公开证据在工作流类别和技术方向上最强,在每个垂直内的已命名客户项目或第三方基准上较弱。[CE015, CE017, CE018, CE019, CE020, CE021]
| 用户任务 | 当前工作流 | Lila 解决方案 | 可衡量收益 | 限制 |
|---|---|---|---|---|
| 设计下一代遗传药物 | 围绕载荷和递送变量顺序设计检测并进行湿实验室迭代 | 治疗药物工作流,同时优化载荷、配方、靶向和可制造性 | 官方页面称每次迭代都有已验证真实世界数据,并跨关键开发变量优化 | 没有具名客户项目或试验阶段产出公开。 |
| 工程化抗体或配体候选物 | 蛋白发现通常在建模和手工台架验证之间来回切换 | 自主 AI 设计加实验测试,覆盖结合、特异性和可开发性 | 官方页面称该工作流协同优化稳定性、溶解度、聚集风险和表达 | 没有公开与既有发现栈对比的基准。 |
| 改进生物工艺或检测工作流 | 构建体和工艺调优通常需要多轮手工循环 | 生物技术工作流,优化构建体、宿主系统、表达平台、配方和方法 | 官方页面称周期可从数月压缩到数周 | 没有按检测类别或可重复性指标公开拆分。 |
| 发现催化剂或分离材料 | 化学团队通常缓慢筛选大空间,且设备测试稀疏 | 化学工作流结合分子设计、预测建模、高通量实验和贴近设备的测试 | 公开材料声称速度更快,且比试错筛选更贴近商业化 | 公开案例研究和客户经济性未披露。 |
| 开发涂层或基础设施材料 | 材料 R&D 通常需要漫长的设计—构建—测试循环 | 面向薄膜、涂层和其他基础设施组件的先进材料工作流 | 官方页面将该平台定位为更快获得尚不存在材料的路径 | 没有公开资产级成熟度或认证路径。 |
| 快速启动合作伙伴发现项目 | 搭建定制机器人产能需要资本开支和专业运营人员 | Catalyst 和 Creation 提供按需平台访问或以结果为导向的活动 | 官方页面承诺用更少时间做更多实验,并为下游管线提供已验证资产 | 定价、合同结构和客户参考仍属私有。 |
收益仅限有来源支持的工作流主张,不应解读为经独立审计的绩效结果。
[CE010, CE011, CE017, CE018, CE019, CE020]基于公开证据深度,而非私人路线图审查的定性成熟度矩阵。
评分只概括保留公开证据的强弱;不能替代内部 QA、客户使用或财务表现数据。
[CE010, CE013, CE021, CE028, CE033, CE039]5.3 机器人、多模态科学和关键依赖
Lila 的运营模型依赖严肃的机器人和科学计算栈,公开记录也给出了公司正在为此配备人员的可信证据。Julie Shah 领导机器人;Milad Abolhasani 把自驱动实验室、多模态分析和机器人专长带入化学项目;Rafael Gómez-Bombarelli 支撑化学和材料中的实验加物理基础 AI;Kenneth Stanley 覆盖开放式发现方法。招聘信号强化了这套领导班底:Greenhouse 岗位覆盖生命科学基础模型、前沿能力、AI 安全、蛋白工程和 AI 数据;CareersInRobotics 招聘又加入 simulation-to-real、MoveIt、LiDAR、SLAM、灵巧操作,以及 NVIDIA Isaac Sim 和 Omniverse。这足以推断其有定制实验室编排和仿真环境,但不足以重构确切硬件 BOM 或设施拓扑。因此,产品差异来自科学推理、机器人和领域专长的组合,同时也依赖稀缺仪器、计算和安全运营控制。NVentures 的支持和 Lila 声明的技术合作议程进一步强化了平台生态故事。[CE005, CE006, CE007, CE029, CE030, CE031]
| 层 / 组件 | 角色 | 依赖 | 风险 |
|---|---|---|---|
| 科学推理模型(Lila Iris) | 跨多个科学模态生成假设、规划实验并解读结果 | 依赖实验生成 token 的持续流入,以及足够的前沿计算 | 架构和基准细节不公开,模型质量难以从外部审计。 |
| 验证器和科学工具 | 用奖励信号和领域专用计算给智能体落地 | 依赖可访问的模拟器、结构预测器、量子化学求解器、编辑器和其他专用工具 | 工具链脆弱或验证偏弱,可能降低真实世界学习质量。 |
| 自主设计和工作流编排 | 将科学目标转成可执行的多步骤计划 | 依赖编排软件、规划逻辑和可靠的实验室排程 | 工作流复杂度放大后,可能埋下隐性失效模式。 |
| AI Science Factory 仪器层 | 执行实体实验,并为飞轮返回已验证数据 | 依赖机器人、仪器、传感器、耗材和可靠的自动化基础设施 | 仪器配置和正常运行时间披露很少,设施成熟度需要更重的尽调。 |
| 仿真与物理科学栈 | 借助基于物理的仿真和多尺度建模,把平台延伸到化学和材料领域 | 依赖领域模型、实验数据,以及物理科学方向的科研带头人 | 若没有公开基准细节,仿真到实验的迁移风险仍然重要。 |
| 机器人与仿真环境 | 支撑感知、操作、路径规划和 sim-to-real 迭代 | 依赖机器人人才、仿真软件,以及与实体设备的集成 | 定制硬件集成可能资本开支很重,也难以跨站点复制。 |
| 商业与安全层 | 平台向客户开放后,支撑合作伙伴访问、隐私控制和 AI 安全项目 | 依赖访问控制、加密、监控和组织层面的安全流程 | 公开保证材料仍比平台的技术野心薄。 |
本表根据公开产品页面、管理层履历、招聘信号和独立报道重建运营架构;它不是内部系统图。
[CE003, CE004, CE006, CE007, CE029, CE030]从依赖视角拆解 Lila 的公开平台叙事,突出其对机器人、算力、招聘和合作伙伴资本的依赖。
依赖关系只反映官方技术、招聘和融资材料中可见的关系;不推断隐藏供应商或云依赖。
[CE029, CE030, CE031, CE033, CE034, CE035]5.4 信任表面、路线图和尽调缺口
公开信任信号存在,但仍比产品叙事的雄心薄得多。公司材料称,Lila 受安全、人类影响和科学严谨性指引;当前招聘计划也包括横跨生物和物理科学的 AI 安全岗位。候选人隐私通知比营销页面更具体,列出基于角色的权限、传输中和静态加密、异常监控,以及对第三方招聘工具的定期安全审查;一般隐私政策也提到技术和组织保障措施,以及按需访问控制。与此同时,已审阅公开材料没有列出产品级认证、受监管质量体系、公开正常运行时间目标,或 AI Science Factories 的公开状态页。路线图证据在资本和建设上更强:公司带着用于建设首批工厂的种子资金发布,后来加入一笔规模可观且有 NVentures 支持的 Series A,称将把更多仪器置于 AI 控制之下,并向初始商业客户群开放平台。结果是一则技术上差异化的故事,但生产保障和客户证据仍需要实质尽调。[CE016, CE033, CE036, CE037, CE038, CE039]
| 控制 / 信号 | 状态 | 范围 | 公开证据 | 缺口 |
|---|---|---|---|---|
| 已验证数据闭环 | 公开宣称的运营原则 | 治疗和生物技术发现工作流 | 官方页面强调每轮迭代都结合真实实验和已验证或经人工验证的数据 | 没有公开的可复现性基准集或外部验证报告。 |
| AI 安全工作流 | 专门招聘信号 | 前沿能力以及生物和物理科学 | Greenhouse 列出 AI 安全和技术缓解方向的科学家、研究工程师岗位 | 方法、评估套件和生产治理没有公开。 |
| 网站隐私控制 | 公开文档披露 | 网站访客数据 | 隐私政策提到物理、技术和组织措施,以及按需授权的访问控制 | 没有公开说明网站控制如何映射到产品或实验室基础设施控制。 |
| 招聘数据安全控制 | 公开文档披露 | 候选人与招聘数据 | 候选人隐私通知提到基于角色的权限、传输和静态加密、监控,以及第三方安全审查 | 这些控制针对招聘系统,而不是 AI Science Factory 运营。 |
| 产品保证材料 | 公开表面有限 | 商业平台和自主实验室运营 | Series A 页面提到世界级 AI 安全;关于页面强调安全和严谨 | 保留来源中没有点名公开 SOC、ISO、GxP、正常运行时间目标或状态页。 |
信任相关行区分了已经明确公开的内容和仍需私下尽调的内容;这里没有点名某项材料,不应理解为该控制本身不存在。
[CE016, CE017, CE033, CE051, CE052, CE053]| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 含义 | 来源 |
|---|---|---|---|---|
| 2023 | 公司在 Flagship 实验室内部成立 | 历史 / 已完成 | 平台和自主实验室论点的起点早于公开发布 | Flagship 新闻稿 |
| 2025-03 | 公开亮相,并获得 $200M 种子轮,用于建设首批 AI Science Factories | 历史 / 已完成 | 为平台和工厂建设提供资本基础 | Flagship 新闻稿;PR Newswire |
| 2025 年在线网站 | Catalyst 和 Creation 商业页面发布 | 当前 / 已上线 | 显示两个产品化商业动作,而不是一个泛化落地页 | Lila Catalyst 和 Lila Creation 页面 |
| 2025 年在线网站 | 治疗、生物技术、化学、材料和能源行业页面发布 | 当前 / 已上线 | 显示横跨生命科学和物理科学的跨领域应用策略 | 官方行业页面 |
| 2025 年 Series A | 总融资达到 $550M,并获得 NVentures 支持和技术合作表述 | 近期 / 已完成 | 提升扩展算力、仪器和商业化工作的能力 | Lila Series A 公告;Industry Examiner |
| 2025-2026 年招聘潮 | 发布基础模型、AI 安全、机器人、仿真和细胞生物学岗位 | 当前 / 活跃 | 说明核心科学和自动化栈仍在积极扩建 | Greenhouse;CareersInRobotics |
| 2025 年外部报道 | 工厂扩张和首批客户商业化被公开讨论 | 近期 / 进行中 | 暗示公司正从隐身平台建设转向面向客户的部署 | Industry Examiner;BioPharmaTrend |
日期和状态标签概括公开里程碑及当前公开表面;它们不是大规模客户采用的证据。
[CE009, CE011, CE013, CE033, CE034, CE036]5.5 图表
06客户情况
6.1 客户图谱和细分:ICP 很宽,但尚无广泛安装基础
截至 2026-06-02 运行日期,Lila 的公开客户故事仍主要是目标买家图谱,而不是已验证账户名单。公司把自己营销为科学操作系统,可服务“你的项目、你的科学家和你最重要的发现挑战”;解决方案页面把这一承诺包装成两个面向合作伙伴的动作:Catalyst 负责平台访问和 Lab-as-a-Service,Creation 负责端到端项目,生成已验证资产甚至新公司。这一表述指向企业研发领导者、首席研究员、风险创建者和科学团队这些真实买家与用户。它不指向自助式产品,也不指向已经大规模部署的安装基础。最可能的早期用户是 Flagship 关联的内部项目和少数定制化外部团队,尤其因为 Biopharma Dive 称 Lila 计划与其他 Flagship 初创公司和外部生物技术公司合作,而 Reuters 后来称公司才刚开始向商业客户开放平台。[CU001, CU002, CU003, CU004, CU005, CU006]
| 客群 | 买方 / 用户 / 付款方 | 用例 | 规模 | 收入 / 战略价值 | 主要缺口 |
|---|---|---|---|---|---|
| Flagship 关联内部项目和组合公司 | 买方 / 付款方:Flagship 孵化的创业构建者或关联项目负责人;用户:内部科学团队 | 用 Lila 加速早期发现、资产生成和新公司孵化 | 最可能是最早使用场景;没有公开数量 | 对早期吞吐和工作成果证明重要,但不等同于多元化第三方收入 | 没有已点名组合公司公开确认使用或付款 |
| 外部治疗和生物技术 R&D 团队 | 买方:R&D 或平台负责人;用户:发现科学家;付款方:生物技术或药企项目预算 | 加速遗传医学、抗体、小分子、生物工艺、试剂或检测项目 | ICP 营销指向清晰;未披露已点名客户 | Lila 已经用上游发现的语言说话,因此最可能最接近首批外部收入 | 没有已点名账户、部署指标或结果案例研究 |
| 材料、化学和能源企业 | 买方:工业 R&D 或先进工程负责人;用户:材料、化学和工艺团队;付款方:企业 R&D 预算 | 催化剂发现、涂层、吸附剂、磁体,以及贴近商业目标的材料测试 | 已报道兴趣;没有公开账户名单 | 如果客户转化,可能让收入跳出生物技术并缩短反馈闭环 | 只披露行业兴趣;没有已签约参考客户 |
| 使用 Creation 的战略伙伴或投资者 | 买方 / 付款方:战略伙伴、投资者或创业工作室;用户:Lila 加合作伙伴团队 | 提出一个问题空间或论点,获得已验证资产、IP 和去风险路线图 | 公开营销的合作模式;未披露已启动项目 | 可以创造高价值定制合作,甚至孵化新公司 | 经常性经济性、已启动项目和客户名称都未公开 |
| 广泛自助或市场型用户 | 公开没有证据 | 没有公开自助工作流、定价页或社区采用表面 | 披露为 0 | 看不到 | 是否存在任何长尾仍未经验证 |
各行拆分了可能的内部生态需求、目标外部 ICP 和未经验证的长尾需求,避免本章夸大客户质量。
[CU001, CU002, CU003, CU010, CU015, CU026]Lila 潜在客户如何从科学瓶颈走向由合作伙伴推进的商业化路径。
Lila 没有发布带完整前后工作流的客户案例,因此该旅程图综合公开产品页面和报道生成。
[CU002, CU003, CU006, CU012, CU028, CU042]6.2 已命名证据缺口:商业化意图可见,但已命名客户证据仍缺席
外部需求最强的公开证据仍是间接的。Reuters 在 2025 年 10 月报道称,Lila 计划通过企业软件和自动化实验室向商业客户开放平台,并已吸引能源、半导体和药物开发公司的兴趣。Fierce 称,同一轮融资将帮助公司引入首批客户。Biopharma Dive 补充称,Lila 预计与外部生物技术公司和其他 Flagship 初创公司合作,而不是推进自有疗法。这些都是有意义的商业化信号,但按更严格的尽调标准,它们还不构成客户证据。在已审阅集合中,没有已命名付费客户、没有公开案例研究、没有买方侧证言、没有采购记录、没有使用指标,也没有来自参考账户的已披露结果。因此,最接近的公开对应物是潜在客户细分和生态关系,而不是生产级客户证据。[CU012, CU013, CU014, CU015, CU030, CU032]
| 客户 / 对手方 | 客群 | 部署 / 用例 | 生产 / 试点 | 结果 / 证明 | 主要限制 |
|---|---|---|---|---|---|
| Flagship 组合公司 / 内部项目 | 内部生态 / 可能最早用户 | 用 Lila 加速创业发现项目和新公司孵化 | 可能是内部或近似试点;未公开确认是一组付费客户 | BioPharma Dive 称 Lila 将与其他 Flagship 初创公司合作;March Capital 将 Geoff 与 Generate 和 Tessera 联系起来 | 没有已点名组合公司公开确认活跃使用、预算或结果 |
| 外部生物技术公司 | 外部生物技术潜在客户 | 通过平台访问或项目加速治疗项目早期发现 | 潜在 / 未验证 | BioPharma Dive 称外部生物技术公司是计划的一部分 | 没有已点名生物技术账户、部署、里程碑或案例研究 |
| 能源、半导体和药物开发公司 | 跨行业企业潜在客户 | 用企业软件加自动化实验室访问做科学发现 | 仅有潜在客户兴趣 | Reuters 称该平台吸引了这些行业公司的兴趣 | 没有名称、试点、采购记录或 ROI 指标 |
| 使用 Creation 的战略伙伴 / 投资者 | 合作伙伴主导的公司创建 | 提出问题论点,获得已验证资产、IP 和新项目蓝图 | Creation 项目 / 合作模式 | Creation 页面明确面向投资者或战略伙伴,并承诺交付已验证成果 | 没有公开案例显示已启动客户公司或披露经常性合同 |
公开客户证明非常薄,因此本表使用最接近、可验证的对手方类别,而不是假装存在已点名生产账户。
[CU010, CU012, CU015, CU024, CU026, CU042]对公开来源可见交易对手类别的证据强度作定性比较。
单元格是分析师只基于公开来源作出的定性判断;低分往往反映披露缺口,而不是已知失败。
[CU015, CU026, CU030, CU031, CU035, CU041]6.3 ICP 和商业化路径:制药 / 生物技术 / 材料买家先行,下游由合作伙伴主导开发
Lila 的 ICP 异常宽,但仍然连贯:它瞄准的是发现速度、实验吞吐量和与实体实验室集成比通用软件席位更重要的科学问题。疗法和生物技术页面强调遗传药物、抗体、小分子、生物工艺、试剂和测定工作流。化学品、先进材料和能源页面强调催化剂、吸附剂、涂层、磁体,以及在商业对齐条件下的工业测试。商业化路径在这些行业中也看起来一致。Catalyst 让现有团队直接访问 Lila Iris 和 AI Science Factories,使客户能加速自身路线图上的项目。Creation 更进一步,接收合作伙伴的投资逻辑或问题陈述,并返回已验证资产、数据包、IP 和去风险后的技术路线图。Reuters 明确称,预期由 Lila 的合作伙伴而非 Lila 自身把分子推进临床试验,或扩展新能源突破。这让 Lila 的收入模型更像企业发现产能和上游科学基础设施,而不是下游产品所有权。[CU003, CU004, CU005, CU006, CU012, CU016]
| 指标 / 里程碑 | 数值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| Flagship 发布并向合作伙伴开放 | 平台亮相;向生命科学和材料科学合作伙伴开放 | 2025-03-10 | Flagship + PR Newswire | 高 | Lila 打算对外商业化的最早公开表述 | 没有合作伙伴名称或承诺量 |
| 生物技术之外 / Flagship 初创公司的合作路径 | Biopharma Dive 称 Lila 将与其他 Flagship 初创公司和外部生物技术公司合作 | 2025-03-10 | BioPharma Dive | 中 | 暗示首批客户场景更可能是协作式发现,而非自助软件 | 没有已点名初创公司或外部生物技术伙伴 |
| Catalyst 和 Creation 产品化 | 两个明确商业动作:平台访问 / LaaS 和端到端项目交付 | 2026 | Lila 解决方案页面 | 高 | 相比单独的发布新闻稿,GTM 设计更清楚 | 没有公开转化率或赢单率数据 |
| 首批客户信息 | 融资轮被描述为帮助公司引入首批客户 | 2025-10-14 | Fierce Biotech | 中 | 意味着到 2025 年末,公开客户牵引仍处早期 | 没有已签客户数量 |
| 披露商业客户兴趣 | 能源、半导体和药物开发公司的兴趣 | 2025-10-14 | Reuters,经 U.S. News | 高 | 将 ICP 扩展到生物技术之外 | 没有公司名称、试点规模或支出披露 |
| 扩大客户交付能力 | 剑桥 235,500 平方英尺租约,加上工厂扩张计划 | 2025-10-14 | Reuters + TechStartups | 高 | 如果销售转化,说明 Lila 预期企业需求会有实质规模 | 没有利用率或已预订吞吐指标 |
| 面向企业买方的集成承诺 | 商业产品可以运行在客户现有数据和平台之上 | 2026 | Lila 能源文章 | 中 | 可能降低企业 R&D 团队的采用摩擦 | 没有参考账户证明实施速度 |
本表跟踪商业化里程碑,而不是客户数量增长,因为没有公开客户总数或活跃账户指标。
[CU003, CU008, CU010, CU012, CU013, CU014]从客户兴趣到由合作伙伴推进下游商业化的通用流程。
由于没有披露具名客户时间线,该流程从公开材料中概括而来。
[CU012, CU013, CU021, CU028, CU042]6.4 耐久性和集中度风险:没有留存证据,可能集中,产品化摩擦真实存在
耐久性是公开记录最弱的地方。已审阅来源没有披露客户数、活跃部署、已售吞吐量、定价、续约、NRR、GRR、流失、合同期限或满意度指标。仅这一点就让客户章节处在证据前状态。另一个主要问题是集中度。如果早期需求先来自 Flagship 关联项目、少数定制外部生物技术项目,或小量企业科学团队,那么第一批收入在战略上可能有价值,但经济上很窄。反向报道强化了这一点。Industry Examiner 认为,Lila 仍必须定义采购团队真正能购买的产品化工作单元;否则,该模型可能看起来像包裹在昂贵自动化实验室外的定制咨询。同一分析还指出,经济性对利用率、重跑和过度定制工作很敏感。因此,客户故事可作为商业化路径投资,但尚不能被承销为耐久、多元的客户基础。[CU026, CU029, CU030, CU031, CU033, CU034]
| 指标 | 数值 / 空值 | 客群 | 置信度 | 尽调问题 |
|---|---|---|---|---|
| 公开客户数量 | 所有外部客户 | 低 | 索取已签客户数、活跃客户数,以及按行业划分的客户结构 | |
| 公开部署 / 吞吐指标 | 所有外部客户 | 低 | 索取已预订实验、活跃项目,以及按账户划分的工厂利用率 | |
| NRR / GRR / 流失 / 续约 | 所有外部客户 | 低 | 索取续约队列、流失历史和合同期限披露 | |
| 客户满意度 / 证言证明 | 所有外部客户 | 低 | 索取客户访谈、NPS 数据和客户撰写的案例研究 | |
| 重复使用代理指标 | 除持续商业化建设和首批客户信息外,没有公开代理指标 | 潜在客户和早期合作伙伴 | 中 | 索取账户级扩张历史和重复项目节奏 |
| 实施摩擦 | 可能中到高,因为 Lila 向科学环境销售软件加自动化实验室工作流 | 企业 R&D 买方 | 中 | 索取从签约到首次实验、首次验证结果的平均时间 |
公开记录未披露留存或满意度数据时,空值是有意保留的占位。
[CU021, CU029, CU030, CU031, CU034, CU036]| 扩张驱动 | 集中 / 执行风险 | 影响 | 尽调路径 |
|---|---|---|---|
| Flagship 生态作为首个需求渠道 | 早期使用可能集中在关联项目内部,而不是形成独立客户证明 | 有利于吞吐,但外部市场验证较弱 | 索取 Flagship 关联项目与第三方活跃项目清单 |
| Catalyst 平台访问 | 如果每次合作都需要大量定制,可能仍像定制服务 | 可能限制毛利质量和可复制性 | 索取标准单位定义、定价逻辑和平均实施范围 |
| Creation 项目和创业发布 | Creation 可能产生战略价值,但会模糊客户收入与创业孵化 | 经常性软件耐久度更难判断 | 索取平台访问、服务、里程碑和创业经济之间的收入拆分 |
| 生物技术之外的跨行业扩张 | 能源和半导体需求只以兴趣形式被引用,不是转化 | 如果真实,可能快速多元化;否则仍只是幻灯片级论点 | 索取已点名非生物技术账户和首批交付结果 |
| 工厂产能建设 | 大型实验室面积提高固定成本风险,尤其在客户利用率爬坡慢时 | 参考账户成熟前,可能压迫利润率 | 索取按工厂划分的利用率、重跑和排队时间指标 |
| 合作伙伴主导的下游商业化 | Lila 依赖合作伙伴把产出推进成产品或试验 | 即便下游经济捕获延迟,上游价值也可能真实 | 索取里程碑结构、数据权利条款和下游参与经济性 |
核心客户风险不是缺少目标市场,而是早期需求能否足够快地变成可重复、产品化和多元化需求。
[CU033, CU034, CU035, CU036, CU037, CU038]针对可能早期客户原型的示意性连续性情景;仅因 Lila 未披露留存数据而使用。
这些百分比是分析师启发式假设,不是公司披露的留存。它们把当前披露模式转成尽调框架,不应解读为实际留存表现。
[CU030, CU031, CU033, CU036, CU041]6.5 图表
07风险
7.1 科学有效性和自主规模化风险
Lila 的核心承诺异常激进:一个系统横跨多个科学领域,实时生成假设、设计并运行实验,并从新数据中学习。这个雄心很重要,因为核心失败模式不是普通软件失误;而是平台在内部闭环里看起来很强,一旦暴露在合作伙伴工作流、混乱生物系统或长周期材料测试中,却无法产出可重复、能说服外部的结果。今天的公开证据更清楚地支持雄心,而不是证明。Lila 自身声称广泛的科学超额表现,但其公开界面没有提供基准表、盲测比较或复现实验包。Fierce Biotech 明确指出,若干标志性技术主张仍缺少公开支撑数据。因此,科学风险问题不是概念是否有趣,而是自主实验在规模化时能否避免优化到虚假代理指标、实验室特定伪影或隐藏的人类脚手架。广度放大了这一风险。Lila 并不聚焦一个狭窄测定或一个边界清晰的垂直。它同时谈疗法、化学和先进材料。每个领域都有不同的验证规范、错误成本和时间线。自动化能加速迭代,但无法抹去可重复性、校准或领域迁移风险。[CR001, CR002, CR003, CR004, CR009, CR010]
| 失效模式 | 可能性 | 严重度 | 缓释成熟度 | 剩余暴露 | 未解决缺口 |
|---|---|---|---|---|---|
| 内部模型胜利无法在合作方或外部实验室环境复现 | 高 | 关键 | 低 — 公开基准和复现材料缺失 | 关键 | 没有公开复现包、基准 notebook 或合作方验证研究 |
| 自主实验闭环优化虚假代理指标,或依赖隐性的人工脚手架 | 中高 | 高 | 低中 — 架构已有描述,但控制措施没有披露 | 高 | 没有公开的人类接管阈值、审计日志或失败案例处理细节 |
| 仪器漂移、实验室运营差异或数据管线损坏在 AI Science Factories 中层层放大 | 中 | 高 | 低中 — 工厂是建设重点,尚无公开证据显示已形成成熟网络 | 高 | 没有仪器校准、正常运行时间或跨站点可复现性的公开质量指标 |
| 敏感生物工作流带来安全或误用担忧,速度快过治理成熟 | 中 | 高 | 低 — 安全招聘可见,但生物安全控制没有披露 | 高 | 没有公开的红队、序列筛查或封闭控制披露 |
| 网站层面的声明跑在公开证据前面,在采购阶段削弱客户信任 | 高 | 高 | 低 — 法律免责声明存在,但证据包未公开 | 高 | 没有具名客户成果、基准表或第三方验证集 |
可能性和严重度采取审慎尽调视角,锚定公开基准、复现和运营质量证据的缺失,而非任何已知事件。
[CR001, CR002, CR003, CR004, CR014, CR016]仅基于公开证据,对 Lila 主要风险作发生可能性与剩余严重度对照。最深色单元集中在科学验证、商业化、治理和执行风险;这些风险在公开记录中仍缺乏充分缓释。
[CR004, CR008, CR017, CR030, CR036, CR038]7.2 监管、生物安全和数据治理风险
公开材料显示,Lila 在网站层面的法务卫生做过基本配置,但还看不到与生物领域自主科学敏感度相匹配的产品级治理。这个缺口很关键:平台一旦从泛 AI 叙事进入生物实验、受监管健康数据使用或合成生物学工作流,问题就不再是“有意思的 AI 公司”,而是“失效模式可能触发隐私、生物安全和双重用途暴露的公司”。NIST、NIH、HHS、EDPS、RAND 和 Johns Hopkins Center for Health Security 指向同一个方向:触及敏感数据或生物工作流的前沿 AI 系统,需要明确治理、可信控制,并在部分场景下设置隔离或监督程序。Lila 的隐私政策承认 GDPR、UK-DPA、跨境传输和监管披露义务,条款也搭起了 Massachusetts 法律和免责声明基础。它有帮助,但不够。这些文件管的是网站,不是部署到合作伙伴项目里的自主科学平台。公开材料没有说明产品级数据隔离、生物安全筛查、红队测试或审计流程。考虑到公司明确关注疗法和生物相邻工作,这一缺口应当被视为真实尽调问题,而不是文书积压。[CR014, CR015, CR016, CR017, CR018, CR019]
| 规则 / 案例 | 司法辖区 | 公开状态 | 可能性 | 严重性 | 缓解措施 | 剩余暴露 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| AI 系统的生物数据治理缺口 | 美国 / 全球 | Center for Health Security 和 RAND 都称,现有框架尚未覆盖 AI-生物技术双用途风险 | 中高 | 关键 | 可见的 AI 安全招聘和通用法律页面;没有公开的产品专属治理包 | 高 | 索取生物数据分类政策、模型使用限制和安全治理签核 |
| 如果重组或合成核酸工作属于 Lila 工作流的一部分,则适用 NIH 生物安全 / 隔离要求 | 美国 | NIH 发布隔离和安全要求,但 Lila 没有公开把其实验室映射到这些控制 | 中 | 高 | 通用安全招聘和公司主导实验室叙事;没有公开 IBC 或隔离细节 | 高 | 按项目索取生物安全等级图谱、IBC 监督安排和事件响应流程 |
| GDPR 和 UK-DPA 下的跨境隐私与数据传输义务 | 欧盟 / 英国 / 美国 | Lila 隐私政策承认 GDPR、UK-DPA,以及向美国和其他司法辖区传输数据 | 中 | 高 | 网站隐私政策已存在;产品专属 DPA 和安全架构未公开 | 中高 | 索取 DPA 模板、子处理方清单、传输机制和客户安全审查材料 |
| 如果合作方数据包含患者信息,HIPAA 或受监管健康数据处理要求 | 美国 | HIPAA 是现行法律框架,但 Lila 未公开描述 PHI 处理或 BAA | 中 | 高 | 除通用隐私表述外,没有医疗数据运营控制的公开证据 | 高 | 索取 BAA 模板、PHI 隔离政策和审计追踪设计 |
| 对公开声明和网站内容的法律信心 | Massachusetts / 网站使用 | 条款确立 Massachusetts 法律、Suffolk County 管辖地、知识产权保护和强保修免责声明 | 中 | 中高 | 基础法律框架已经搭好,但网站免责声明降低了营销表述的尽调价值 | 中 | 承销前要求合同级陈述、技术附表,以及对声明依据的法律审查 |
各行仅依据公开法律、监管和政策证据,按剩余严重度排序;公开来源没有披露 Lila 已完整落地的合规栈。
[CR014, CR016, CR017, CR018, CR019, CR020]7.3 商业化和竞争风险
即便 Lila 的核心科学栈真实存在,商业化风险仍然很重,因为公司想压缩的是结构上本就周期很长的类别。外部药物发现资料说得很直接:多数临床前项目进不了人体试验,临床批准率依然很低,开发常常超过十年,总成本可能高达数十亿美元。先进材料商业化的机制不同,但后果相似:验证、集成和客户采用都需要时间。这意味着 Lila 暴露在典型深科技陷阱里:市场还无法验证可重复产品化之前,公司就先靠平台承诺融资。公开商业化证据也很薄。管理层称正在迎接第一批客户,但客户名称、合同规模、收入和结果数据都没有公开。与此同时,竞争版图并不空。Recursion、Isomorphic Labs、Insilico、Absci 和 CuspAI 都在对外销售垂直领域证据点、管线或专门技术定位。因此 Lila 不仅要打赢科学领域的既有玩家,还要打赢那些能向投资人和买家讲出更简单故事的专业同业。多领域平台在 TAM 上看起来更大,但到销售节点仍可能输在聚焦、紧迫性和信任上。[CR006, CR007, CR008, CR025, CR026, CR027]
| 依赖 | 交易对手 | 角色 | 集中度 | 失效场景 | 严重度 | 缓释 | 剩余暴露 |
|---|---|---|---|---|---|---|---|
| 资本和战略支持 | Flagship 加广泛投资人银团 | 发起方、资金来源和生态伙伴 | 高 | 科学和商业证明追不上烧钱速度,风险下降前被迫再融资 | 关键 | 大额现金余额争取时间;尚无持久收入的公开证据 | 高 |
| 早期客户转化 | 未披露的首批客户队列 | 标杆账户和初步商业化证明 | 高 | 首批队列未转化为具名部署、续约或可发表成果 | 关键 | 管理层称已有首批队列,但没有具名公开客户证明 | 高 |
| 科学仪器和工厂铺开 | AI Science Factory 建设和站点运营 | 支撑软件声明的实体实验层 | 高 | 工厂扩张落后于招聘、校准或利用率,投入增加但学习速度下降 | 高 | 资本已指定用于建设工厂,但公开运营指标缺失 | 高 |
| 相对聚焦同行的跨领域可信度 | 竞争对手:Recursion、Isomorphic Labs、Insilico、Absci、CuspAI | 人才、合作方和客户注意力的竞争替代项 | 高 | 专门化竞争对手凭更窄的证明点取胜,而 Lila 仍停留在宽平台叙事 | 高 | 多领域选择权真实存在,但聚焦度尚未得到外部证明 | 中高 |
| 数据与安全能力 | AI 安全、AI 数据、前沿能力和领域科学岗位在招 | 负责任扩张所需的运营能力 | 高 | 关键岗位长期空缺,同时拖慢治理和执行 | 高 | 多地招聘活跃,但公开完成状态未知 | 高 |
风险最高的依赖不只是供应商,而是把 Lila 平台转化为可重复证明所需的外部关系和运营能力。
[CR007, CR008, CR010, CR011, CR031, CR032]Lila 叙事背后的关键外部和内部依赖:资本、AI 工厂、数据治理能力、安全人员配置和标杆客户。任何缺口都会拖慢证据生成和商业化。
[CR002, CR010, CR011, CR012, CR036, CR037]7.4 资本、招聘和执行风险
以一家这么年轻的公司而言,Lila 已经融到异常充裕的资本,但公开证据显示,这笔钱买来的是尝试搭建的权利,不是已经在运营上去风险的证明。公司称将用资金扩张 AI Science Factories、引入首批客户,并增加更多顶尖人才。招聘版图显示这件事铺得有多宽:Greenhouse 职位页仍在招聘 AI 安全、蛋白工程、前沿能力、细胞生物学自主科学、机器学习研究和技术项目管理岗位。这些不是边缘岗位;任何想安全规模化运行自主科学的公司,都需要这些核心职能。Cambridge、San Francisco 和 London 的多地点布局进一步增加管理复杂度,尤其是公司还同时跨越多个科学终端市场。由此形成一组熟悉的深科技风险:前期投入重、验证周期长,执行瓶颈可能从招聘、协调、安全审查延迟或物理基础设施利用不足中冒出来。如果 Lila 不能足够快地把资本转化为外部能读懂的科学和商业里程碑,下一轮融资可能会先于证据到来。[CR006, CR011, CR012, CR013, CR037, CR038]
| 角色 / 职能 | 依赖或缺口 | 可能性 | 严重度 | 缓释 | 尽调路径 |
|---|---|---|---|---|---|
| AI 安全和技术缓释 | 公开招聘显示该职能仍在补人 | 高 | 关键 | 专项招聘可见 | 索取组织架构图、红队责任归属,以及向 CEO 或董事会的汇报线 |
| 领域科学整合 | 平台范围横跨多个领域,但蛋白工程、细胞生物学和前沿能力岗位仍在开放 | 高 | 高 | 跨职能、跨站点招聘 | 按领域索取人员配置、负责人任期和各垂直项目归属 |
| Cambridge、London 和 San Francisco 的跨站点项目管理 | 多站点协调抬高沟通和实验室运营复杂度 | 中高 | 高 | 公司已跨多个地点运营 | 索取站点级里程碑节奏、升级路径和利用率指标 |
| 从平台到客户价值的商业转化 | 尽管已有首批队列表述,公开具名客户成果仍缺失 | 高 | 高 | 客户 onboarding 似乎已经开始 | 按阶段索取商业管线、设计伙伴清单和续约假设 |
| 资本配置纪律 | 治疗和材料两端铺得太宽,可能分散管理层注意力 | 中高 | 高 | 大额资金基础和合作伙伴网络 | 索取董事会批准的优先级矩阵和季度 go / no-go 标准 |
执行风险登记表强调公开招聘界面仍显示在推进中的角色和协调机制。
[CR011, CR012, CR013, CR037, CR038, CR039]7.5 监测、缓释措施和投资逻辑破裂触发点
公开材料确实显示出一些早期缓释信号:法律和隐私页面已经存在,公司在招聘 AI 安全岗位,管理层把融资与平台及工厂建设绑定,而不是假装商业化已经解决。但相对于公司的风险暴露面,可见缓释措施仍然偏通用。它们还没有展示产品级治理、合作伙伴验证,或同一系统能跨多个科学领域产出可重复价值的硬证据。因此,投资姿态应明确保持里程碑驱动。近期承销问题不是 Lila 是否可能变得重要,而是它能否在资本强度和竞争压力固化之前,把一个庞大、昂贵、多领域平台压成狭窄且经外部验证的证据。最清晰的投资逻辑破裂标准也应当是经验性的:如果到下一次重大融资检查点,公司仍缺少具名客户证据、外部可信基准数据或披露的治理控制,风险画像就应被视为恶化,而不是改善。公开雄心很多;可公开证伪的证据仍然稀缺。[CR008, CR013, CR014, CR016, CR017, CR030]
| 风险 | 可监测触发点 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 科学有效性风险 | 外部技术证明 | 下一轮重大融资前,没有合作方验证的基准包、复现研究或负面结果披露 | 不按模型优越性承销;要求基于里程碑的分批投资,或等待 |
| 商业化风险 | 客户证明 | 首批队列表述又经历一个刷新周期后,仍没有具名付费客户、ACV 或成果案例研究 | 将商业化视为未证实,并把估值支撑标为薄弱 |
| 生物安全和数据治理风险 | 治理披露 | 高敏感项目扩张前,没有产品专属生物安全、DPA、BAA 或敏感数据治理包 | 任何资本承诺前都要求法律和安全尽调 |
| 执行风险 | 招聘完成度和组织稳定性 | AI 安全、前沿能力和领域科学岗位在多站点持续空缺或快速流失 | 假设爬坡更慢、烧钱更高;下调里程碑时间 |
| 聚焦风险 | 组合纪律 | 管理层无法明确一两个滩头领域,并给出清晰 go / no-go 规则和资本配置 | 将业务宽度视为负面因素,避免按溢价倍数承销平台选择权 |
这些打破投资逻辑的标准刻意设置为可监测项,应对照下一轮融资、重大客户公告或治理审查来检查,而不只看叙事更新。
[CR008, CR013, CR017, CR030, CR038, CR039]Lila 最大风险如何从科学证据和治理,传导到客户采用、融资杠杆,并最终影响投资逻辑。
[CR008, CR017, CR030, CR036, CR038, CR039]08估值
8.1 建议、信心和价格纪律
Lila 已经完成 AI 原生科学领域最强的早期私募融资之一:先完成 $235 million 首次交割,再通过 $115 million 扩展轮把 2025 年 Series A 推至 $350 million,累计融资达到 $550 million,最新披露估值超过 $1.3 billion。这次资本形成很重要。它说明成熟投资人愿意为 Flagship 孵化、异常宽的平台野心,以及自主实验室可能在疗法、材料、化学和其他领域复利扩张的可能性付费。单看融资,Lila 已经像一项溢价资产,而不是常规 Series A 公司。 问题在于,价格更多由财团质量和平台可选性定出来,而不是由公开披露的商业证据支撑。Reuters 称 Lila 计划主要通过合作伙伴商业化,而不是自己推进分子;Fierce 也指出,公司尚未公开发布数据来支持其最强技术主张。已审阅来源中仍看不到具名付费客户、披露的经常性收入、披露的毛利率或公开股权结构条款。因此,即便当前估值可以理解,也很难称得上有吸引力。 因此,我的建议是观察,而不是买入。信心为中,风险为高。公开证据支持一个结论:Lila 是高质量融资故事,但还不是有吸引力的价格。当前估值看起来偏高而非非理性:高于普通 Series A 定价,低于最激进的私募 AI 科学融资;如果证据仍然薄弱,也容易遭遇类似公开市场的重估。[CV006, CV007, CV008, CV011, CV012, CV013]
| 维度 | 评估 | 证据基础 | 决策含义 |
|---|---|---|---|
| 投资建议 | 观察 / 尽调闸门 | 当前估值高于 $1.3B,但公开证明仍稀疏 | 仅在价格或证明改善时跟进 |
| 置信度 | 中 | 融资事实得到充分交叉印证,但技术和商业证明不足 | 承销时避免虚假精确 |
| 风险评级 | 高 | 资本密集、商业化前披露不足和板块重估风险都仍然重要 | 假设下行保护有限 |
| 估值立场 | 偏高 | 当前定价可以解释,但基准情景下安全边际很小 | 不要只靠动量追这轮 |
| 进入纪律 | 仅基于里程碑 | 具名付费合作方、公开验证数据和干净条款是上调路径 | 仅在尽调后或进入价格更好时重新评估 |
评估仅使用公开证据;它刻意对价格敏感,而不是单纯的质量打分。
[CV008, CV012, CV013, CV039, CV042, CV043]| 论点 | 证据 | 什么会改变判断 |
|---|---|---|
| 投资逻辑:Lila 正在搭建真正差异化的自主科学平台 | 多领域定位、AI Science Factories,以及顶级投资人提供的 $550M 支持 | 独立验证数据或具名付费合作方会显著强化这一投资逻辑 |
| 投资逻辑:Flagship 孵化支持早期溢价 | Flagship 多次打造资本密集型平台公司,并能带来战略资本 | 只有 Lila 证明商业转化,而不只是融资能力强,溢价才应扩大 |
| 投资逻辑:私募市场对 AI 科学的胃口仍可能很大 | Xaira 和 Isomorphic 说明,品类龙头可以以异常规模融资 | 如果公开市场重置持续压缩最终结果,仅靠私募胃口还不够 |
| 反向逻辑:公开证明太薄,撑不起当前估值 | 没有具名付费客户或收入披露;Fierce 指出关键技术声明没有公开数据 | 可背书的客户名单和可复现基准会削弱这一反向逻辑 |
| 反向逻辑:公开 AI 药物可比公司已经大幅重置 | Recursion 和 Exscientia 都损失了大部分公开市值;Exscientia 以约 $688M 出售 | 持久的公开市场重估或清晰的私募证明会缓和这一警示 |
| 反向逻辑:板块经济性仍未证实 | 尚无 AI 发现药物获批,后期疗效仍是投资人的瓶颈 | 后期胜利或获批产品会证明支付更高溢价有依据 |
各行框定多空两侧最高信念的论点,以及会推动判断变化的具体证据。
[CV003, CV004, CV012, CV013, CV015, CV016]从高价资本形成和平台宽度,到证据缺口和行业重估风险,最终推导观察建议。
[CV007, CV008, CV012, CV016, CV025, CV030]8.2 融资背景、可比公司和 Flagship 溢价
今天给 Lila 估值,最好的办法不是套收入倍数;收入没有公开披露,公司也没有把自己包装成完全一体化的疗法业务。更合适的框架,是对照三个可比集群做概率加权的里程碑定价:高溢价私募 AI 科学轮、Flagship 平台型同业,以及公开 AI 药物公司的重估。上行参照里,Xaira 的 $1 billion 启动融资,以及 Isomorphic Labs 继 2025 年 $600 million 融资后又在 2026 年完成 $2.1 billion 融资,说明只要市场相信 AI 平台能成为基础设施,私募市场就会以异常规模资助类别龙头。Generate:Biomedicines 是更接近 Flagship 风格的参照:同样资本密集,但临床管线更可见,自 2020 年以来已融资近 $700 million。 下行参照则严酷得多。Recursion 到 2026 年 6 月市值只有约 $2.01 billion,尽管它多年搭建平台、已上市且拥有多项合作;其 10-K 仍警告公司没有获批产品,并预计还需要大量额外融资。Exscientia 的重估更尖锐:2021 年融资充足地 IPO,随后在 2024 年以约 $688 million 完成合并;BioPharma Dive 等报道强调,Recursion 和 Exscientia 到那时都已损失大部分价值。这些公开市场结果并不自动说明 Lila 估值过高,但确实限制了投资人该为缺少证据的叙事支付多少。 相比普通早期公司,Flagship 确实应享有一定溢价,因为它能发起深技术团队、组织战略资本并讲出类别故事。但这个溢价不应无限。没有具名合作伙伴、已发表验证数据或单位经济模型,公开证据还不足以支持按“Lila 已经把平台承诺转化为耐久、可重复现金流”的方式定价。[CV015, CV016, CV018, CV019, CV020, CV021]
| 可比公司 | 指标 | 倍数 / 估值 / 状态 | 相关性 | 局限 |
|---|---|---|---|---|
| Xaira | 启动融资 | 启动时融资 >$1B;私募估值未披露 | 说明晚期证明出现前,私募资本愿意激进支持 AI 药物平台 | 纯治疗领域聚焦,比 Lila 的跨领域范围更窄 |
| Isomorphic Labs | 私募融资规模 | 2025 年 $600M 轮;2026 年 $2.1B Series B;私募估值未披露 | 当前最好的顶端 AI 科学资本胃口参照 | 背后有 DeepMind/Alphabet 规模支持,Lila 无法匹配 |
| Generate:Biomedicines | Flagship 平台融资 | 2023 年 $273M Series C;2020 年以来股权融资近 $700M | 有用的 Flagship 风格可比,平台推进路径可见 | Generate 披露的管线成熟度和临床资产更多 |
| Recursion | 公开市值 | 截至 2026 年 6 月市值约 $2.01B | 有合作伙伴关系的规模化 AI 药物平台公开基准 | 公开市场折价比私募估值更严苛 |
| Exscientia | 公开重置 / M&A 价值 | 2021 年以每 ADS $22 IPO,另有 $160M 同步配售;2024 年以约 $688M 合并 | 说明证明滞后时,AI 药物估值可以多快重置 | 单一公司的治理和执行问题也影响了结果 |
| Lila Sciences | 当前参考点 | $350M Series A 和累计融资 $550M 后,估值 >$1.3B | 本章当前承销锚点 | 没有公开收入或具名客户数据,难以三角校准精度 |
可比集合只提供方向而非穷尽列表,因为 Lila 横跨多个科学终端市场,且没有披露收入。
[CV008, CV016, CV018, CV019, CV020, CV023]Lila 隐含估值对证据和定价里程碑的敏感度,而非收入倍数。
柱状值为示意性投后估值,锚定可比轮次和证据里程碑,不是 DCF 输出。
[CV033, CV034, CV035, CV040, CV041, CV045]8.3 乐观、基准和悲观估值框架
乐观情景取决于 Lila 能否从异常亮眼的融资表象,走向可验证的运营证据。这意味着披露具名付费合作伙伴,展示可复现实验基准或客户成果,并证明自主实验室相较传统团队能实质性提升发现吞吐量。如果这些信号出现,Lila 有可能拿到下一档高溢价私募估值,大约落在 $2.3 billion 到 $3.0 billion 区间。即便如此,它仍会低于 Isomorphic 已经达到的资本规模,并接近私募 AI 科学投资人在没有公开市场检验时愿意出资的上沿。 基准情景更温和,而且重要的是,它离当前估值近得不舒服。在该情景下,Lila 继续吸引强势支持者和有限的合作伙伴试点,但披露的经济性不足以带来决定性重估。只要资本市场保持建设性,$1.1 billion 到 $1.6 billion 左右的估值区间可以支撑。这个区间意味着相较当前标记几乎没有安全边际,因为今天的价格已经折现了相当多未来证据。 悲观情景是公开市场重估传导进私募市场:技术主张仍不透明,合作伙伴采用继续模糊,更广泛的 AI 药物发现行业不断提醒投资人,尚无 AI 发现药物获得批准,后期疗效也未被证明。在这种环境里,Lila 可能被迫重估到 $0.5 billion 到 $0.9 billion,甚至下轮降价。这些情景的概率加权中点接近当前估值,这正是为什么类股票答案不是 Lila 不好,而是价格还不够慷慨。[CV030, CV031, CV036, CV037, CV038, CV039]
| 情景 | 假设 | 估值 / 回报逻辑 | 关键风险 | 概率信号 |
|---|---|---|---|---|
| 乐观 | 至少两个领域出现具名付费合作方、公开技术验证、溢价后续轮或战略交易 | ~$2.3B-$3.0B 投后;相对 $1.3B 进入估值,毛回报大约 1.8x-2.3x | 执行证明必须快速出现,并且保持可信 | ~20% |
| 基准 | 有一些合作方转化,资本获取继续,但单位经济披露仍有限 | ~$1.1B-$1.6B;相对当前估值大约 0.8x-1.2x | 基准情景离今天估值太近,无法提供舒适度 | ~50% |
| 悲观 | 证明不透明、客户转化弱,并出现类似 Recursion/Exscientia 重置的板块再压缩 | ~$0.5B-$0.9B;相对当前估值大约 0.4x-0.7x | 下轮降价或战略重置变得可能 | ~30% |
情景区间是按概率加权的投后估计,锚定当前融资证据和可比结果。
[CV036, CV037, CV038, CV039]基于当前估值标记的低、基准和高估值及总结果区间。
区间是情景估算,以当前 >$1.3B 估值标记作为参考进入点。
[CV036, CV037, CV038, CV039, CV045]面向 IC 的评分,覆盖市场雄心、证据质量、商业证据、下行保护和估值纪律。
分数带有判断性,只反映本章收集的公开证据集。
[CV003, CV008, CV012, CV015, CV033, CV043]8.4 退出准备度、尽调问题和投资逻辑破裂触发点
从公开证据看,Lila 还没有准备好退出。如果它成为横跨多个垂直领域的科学发现基础设施层,最终可能支撑一个非常大的结果,但这仍是战略雄心,不是已披露的运营画像。若要明确改变建议,需要看到合作伙伴兴趣转化为付费项目、实验室带来可衡量的生产率提升,并确认当前估值没有藏着会损害未来回报的优先权或稀释悬挂。 因此,最重要的尽调工作很实际,而非哲学。投资人需要看到是否有客户在付费、这些项目可重复性如何、产出经济性长什么样,以及公司能否把技术主张打包成一组可供怀疑者承销的验证集。投资人还需要资本结构文件:股权结构表、优先股堆叠、期权池和治理权。没有这些,即便顶层估值判断正确,最终回报仍可能错。 投资逻辑破裂触发点也很清楚。如果 12 到 18 个月过去仍没有客户披露,如果 Lila 仍拿不出可复现技术证据,或公开 AI 药物可比公司再下一台阶,当前溢价就应压缩。反过来,披露付费合作伙伴、基准数据和更干净的条款,则足以快速重审建议。[CV010, CV013, CV040, CV041, CV042, CV046]
| 触发点 | 阈值 | 对投资逻辑的传导 | 行动含义 |
|---|---|---|---|
| 没有具名付费客户 | 下一次重大融资前或 12-18 个月内没有客户披露 | 平台选择权仍是叙事,而非商业证明 | 不要加码;假设溢价应压缩 |
| 没有公开技术验证 | 下一轮尽调周期仍没有可复现基准数据或合作方案例研究 | 科学护城河仍未验证,也更难变现 | 将估值情景推向悲观 |
| 板块重估 | 公开 AI 药物可比公司再次大幅回撤,或类似私募公司出现降价轮 | 私募胃口可能不再支撑今天的溢价 | 用公开市场重置折价重新承销 |
| 不利条款或遗留压力 | 优先股堆叠、治理条款或稀释比预期更苛刻 | 即使账面估值不变,回报结构也可能被打穿 | 暂停部署资金,直到厘清资本结构风险 |
触发项都可观察,并直接对应估值压缩或停止部署资金的决定。
[CV038, CV040, CV041, CV046, CV047]| 主题 | 缺失证据 | 重要性 | 负责人或尽调路径 |
|---|---|---|---|
| 商业验证 | 具名客户、合同金额、续约情况和合作伙伴背书 | 需要判断平台兴趣能否转成可重复收入 | 管理层和客户;要求安排客户访谈并提供合同摘要 |
| 实验室经济性 | 吞吐量、单次实验成本、成功率和失败模式 | 决定 AI 科学工厂是能复利扩张,还是只会吃掉资本 | 运营复核;要求提供 KPI 时间序列和队列分析 |
| 技术验证 | 可复现实验基准和第三方案例研究 | 当前溢价靠可测量护城河支撑,不能只靠叙事 | 科学尽调;要求提供资料室文件包和复现实证 |
| 资本结构 | 股权结构表、清算优先权、期权池和治理权利 | 账面估值可能大幅误导投资人的实际结果 | 法律尽调;要求提供投资条款清单和分配瀑布模型 |
| 可比对象精度 | 经经纪商或管理层确认的 Xaira 与 Isomorphic 私募估值标记 | 有助于收紧可比组里的溢价和折价假设 | 二级市场数据供应商、经纪商和管理层沟通 |
每一项都是真正的承销阻断点,而不是可有可无的后续问题。
[CV013, CV033, CV040, CV041, CV042, CV043]免责声明
本报告是 AI 辅助生成的尽调材料,仅基于截至 2026-06-02 的公开信息;不构成投资建议。私营公司融资条款、经营指标、科学结果、客户合同和商业化时间线,可能与公开披露存在重大差异;作出任何投资或合作决策前,应以一手文件核验所有重大事实。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Lila describes itself as the world’s first scientific superintelligence platform for life, chemical, and materials science. | 高 | SO001, SO013 |
| CO002 | Lila says its advanced AI model is the brain and its AI Science Factory instruments are the body of the platform. | 高 | SO001, SO002 |
| CO003 | Official materials say Lila’s system generates hypotheses, designs experiments, runs them, and learns from new data in real time. | 高 | SO001, SO005, SO015 |
| CO004 | Public descriptions position Lila against use cases in therapeutics, chemistry, materials, energy, semiconductors, and defense rather than consumer AI. | 中 | SO001, SO016, SO021 |
| CO005 | Lila was founded in 2023 inside Flagship Pioneering’s labs and publicly unveiled in March 2025. | 高 | SO013, SO015, SO021 |
| CO006 | Lila says it spent about three years building inside Flagship labs before the March 2025 reveal, indicating substantial incubation before public launch. | 中 | SO005, SO015 |
| CO007 | The March 2025 launch was paired with $200 million of committed seed financing. | 高 | SO005, SO015 |
| CO008 | Seed backers included Flagship Pioneering, General Catalyst, March Capital, Altitude Life Science Ventures, ARK Venture Fund, Blue Horizon Advisors, Modi Ventures, the State of Michigan Retirement System, and an ADIA subsidiary. | 高 | SO005, SO015 |
| CO009 | Geoffrey von Maltzahn is Lila’s co-founder and CEO. | 高 | SO003, SO004, SO014 |
| CO010 | Geoffrey previously founded or co-founded Generate:Biomedicines, Tessera Therapeutics, Quotient Therapeutics, Indigo Ag, Sana Biotechnology, and Seres Therapeutics, and his bio credits him with more than 200 patents or applications. | 高 | SO004, SO014 |
| CO011 | Andrew Beam is CTO, leads AI for scientific discovery, and previously co-founded Generate:Biomedicines while serving as a Senior Fellow at Flagship. | 高 | SO003, SO009 |
| CO012 | Jawad Ahsan serves as COO and CFO and previously held CFO roles at Axon and Market Track/Numerator. | 高 | SO003, SO008 |
| CO013 | Chris Fussell serves as Lila’s operations leader after a career that included U.S. Navy SEAL service and leadership at McChrystal Group. | 中 | SO003, SO010 |
| CO014 | Julie Shah is Chief Robotics Officer and also leads MIT’s Department of Aeronautics and Astronautics. | 中 | SO003, SO012 |
| CO015 | Rafael Gómez-Bombarelli is co-founder and CSO of Physical Sciences and is an MIT materials scientist focused on AI plus physics-based simulations. | 中 | SO003, SO011 |
| CO016 | Noubar Afeyan is Lila’s co-founder and chairman while also serving as Flagship’s founder and CEO, embedding sponsor influence in governance. | 高 | SO014, SO015, SO021 |
| CO017 | John Kim appears on the current leadership roster as President, Corporate Development. | 中 | SO003 |
| CO018 | The first public Series A close totaled $235 million and was co-led by Braidwell and Collective Global. | 中 | SO018, SO026, SO027 |
| CO019 | An October 2025 extension added $115 million and brought Lila’s total Series A financing to $350 million. | 高 | SO006, SO016, SO018, SO019 |
| CO020 | Overall capital raised reached $550 million across the $200 million seed and $350 million Series A. | 高 | SO006, SO016, SO019, SO021 |
| CO021 | Reuters, Goodwin, CNBC, and multiple syndications placed Lila’s post-extension valuation at more than $1.3 billion. | 高 | SO016, SO019, SO021, SO023, SO024 |
| CO022 | The extension round brought in NVentures, Analog Devices, IQT, Dauntless Ventures, Catalio Capital Management, Pennant Investors, and other new backers. | 中 | SO006, SO018 |
| CO023 | The broader Series A syndicate also included Flagship, Altitude, Alumni Ventures, ARK Venture Fund, Common Metal, General Catalyst, March Capital, Mathers Foundation, Modi Ventures, NGS Super, the State of Michigan Retirement System, and an ADIA subsidiary. | 中 | SO006, SO026, SO027 |
| CO024 | Management says the fresh capital will improve scientific performance, scale AI Science Factories, open the platform to commercial partners, and hire aggressively. | 中 | SO006, SO020 |
| CO025 | Reuters says Lila plans to open its platform to commercial customers via enterprise software and has seen interest from firms in energy, semiconductors, and drug development. | 中 | SO016, SO024 |
| CO026 | Von Maltzahn told Reuters that Lila does not plan to take molecules into clinical trials or scale energy breakthroughs itself; partners and startups on the platform are intended to do that. | 中 | SO016 |
| CO027 | Reuters, AGBI, and Economic Times said Lila recently signed a 235,500-square-foot lease in Cambridge, Massachusetts. | 高 | SO016, SO023, SO024 |
| CO028 | Bisnow reported that the Cambridge footprint is at 1 and 5 Alewife Park, leased from IQHQ. | 中 | SO022 |
| CO029 | Reuters and CNBC described the Cambridge facility as one of Greater Boston’s largest lab leases of 2025. | 高 | SO016, SO021, SO022 |
| CO030 | Flagship’s company page says Lila is growing its team in Cambridge, San Francisco, and London. | 中 | SO013 |
| CO031 | Independent coverage says Lila also plans additional hubs in Boston, San Francisco, and London to house AI Science Factories. | 中 | SO025, SO027 |
| CO032 | Lila’s differentiating thesis is that scientific AI leadership will come from proprietary experimental data generated in automated labs, not only from internet-scale model training. | 高 | SO001, SO016 |
| CO033 | Official materials claim the platform has already delivered thousands of discoveries or benchmark-beating results across life sciences, chemistry, and materials. | 中 | SO005, SO006, SO016 |
| CO034 | Fierce Biotech noted that Lila had not publicly released data to support several of its bold scientific-performance claims. | 中 | SO018 |
| CO035 | CNBC wrote that hype around Lila may be running ahead of reality because many AI platforms have struggled to outperform traditional research models consistently. | 中 | SO021 |
| CO036 | CafePharma summarized the September 2025 first-close round as a unicorn financing at roughly $1.2 billion valuation, showing momentum even before the October extension. | 低 | SO026 |
| CO037 | Across reviewed public sources, Lila does not disclose revenue or run-rate, so there is no supportable public revenue KPI for this chapter. | 中 | SO001, SO006, SO016, SO021 |
| CO038 | Across reviewed public sources, Lila does not disclose named customers or a customer count, although management says a first cohort is being welcomed and partner interest exists. | 中 | SO006, SO016 |
| CO039 | Across reviewed public sources, Lila does not disclose current headcount, so hiring intensity is visible only qualitatively through recruiting language and expansion plans. | 低 | SO003, SO007, SO020 |
| CO040 | The combination of Geoffrey’s company-creation track record, Andrew Beam’s AI-science background, Jawad Ahsan’s public-company finance experience, and Julie Shah’s robotics leadership gives Lila unusually senior functional coverage for a young platform company. | 中 | SO004, SO008, SO009, SO012 |
| CO041 | CNBC listed ten founders or founding executives, including Geoffrey von Maltzahn, John Kim, Chris Fussell, Andy Beam, Rafael Gómez-Bombarelli, John Gregoire, Ben Kompa, Alex Sneider, Josh Waitzkin, and Noubar Afeyan. | 中 | SO021 |
| CO042 | Lila’s messaging positions enterprise platform access and AI Science Factories—not an internal drug pipeline—as the primary route to commercialization. | 中 | SO001, SO006, SO016 |
| CM001 | The market relevant to Lila is the overlap of lab automation, laboratory informatics, AI drug discovery, and emergent self-driving laboratory orchestration rather than one canonical public category. | 中 | SM001, SM006, SM010, SM016 |
| CM002 | Public lab automation sources include robotic systems, workstations, liquid handling, screening workflows, and workflow software used in drug discovery and adjacent lab processes. | 中 | SM001, SM004 |
| CM003 | Public laboratory informatics sources define a separate software and data layer built around LIMS, ELN, LES, cloud delivery, and compliance tooling. | 中 | SM006, SM007, SM008 |
| CM004 | Public AI drug discovery coverage centers on software and services for target identification, molecular screening, repurposing, de novo design, and preclinical decision support. | 中 | SM009, SM010 |
| CM005 | Self-driving laboratory literature consistently describes the category as a closed-loop combination of automated instruments, AI decision-making, and orchestration software. | 高 | SM014, SM015, SM016, SM018 |
| CM006 | Routine diagnostics operations, generic enterprise AI, broad clinical-development services, and general industrial automation outside an experimental loop should be treated as excluded adjacencies for Lila’s market boundary. | 中 | SM001, SM006, SM010, SM015 |
| CM007 | MarketsandMarkets projects the global lab automation market at USD 6.60 billion in 2026 and USD 8.62 billion in 2031, a 6.6% CAGR. | 中 | SM001 |
| CM008 | Precedence Research estimates the global lab automation market at USD 8.91 billion in 2026. | 中 | SM003 |
| CM009 | Future Market Insights estimates the lab automation market at USD 2.7 billion in 2026 and USD 6.9 billion by 2036, implying a 9.7% CAGR. | 中 | SM004 |
| CM010 | Business Research Insights estimates the global lab automation market at USD 12.12 billion in 2026. | 低 | SM002 |
| CM011 | Published lab automation estimates vary by more than four times from low to high, which makes boundary and methodology sensitivity a material diligence issue. | 中 | SM001, SM002, SM003, SM004 |
| CM012 | Mordor projects the laboratory informatics market at USD 4.05 billion in 2026 and USD 6.08 billion by 2031, a 8.46% CAGR. | 中 | SM006 |
| CM013 | Business Research Insights estimates the laboratory informatics market at USD 5.4 billion in 2026. | 低 | SM007 |
| CM014 | Grand View frames laboratory informatics at USD 4.1 billion in 2025 and USD 6.0 billion by 2033, a 4.9% CAGR from 2026 to 2033. | 中 | SM008 |
| CM015 | Mordor estimates the AI drug discovery market at USD 3.25 billion in 2026 and USD 10.29 billion by 2031, a 25.94% CAGR. | 中 | SM010 |
| CM016 | Global Market Insights says AI drug discovery exceeded USD 3.1 billion in 2025 and will grow 30.5% annually from 2026 to 2035. | 中 | SM009 |
| CM017 | AI drug discovery appears smaller than adjacent automation and informatics categories by current revenue but materially faster-growing. | 中 | SM001, SM006, SM009, SM010 |
| CM018 | A broad adjacent-market envelope relevant to AI science factories sits in the low-teens billions of dollars using conservative 2025-2026 published category estimates, but those categories overlap and do not constitute a clean additive TAM. | 低 | SM001, SM006, SM010, SM015 |
| CM019 | Public evidence does not provide a standardized standalone TAM for autonomous or self-driving laboratories; the literature describes an emergent architecture rather than a mature revenue category. | 高 | SM014, SM015, SM016, SM018 |
| CM020 | Pharmaceutical and biotechnology companies held 53.14% of laboratory informatics spending in 2025 in Mordor’s market segmentation. | 中 | SM006 |
| CM021 | CROs are a meaningful secondary buyer group because they are explicit lab automation end users and a fast-growing laboratory informatics cohort. | 中 | SM004, SM006 |
| CM022 | Thermo Fisher’s 2024 revenue profile shows scientific-tool demand concentrated in pharma and biotech at 57% of revenue, with academic and government, industrial and applied, and diagnostics and healthcare each materially smaller. | 中 | SM020 |
| CM023 | Agilent positions itself across life sciences, diagnostics, and applied markets and says most of the world’s labs use Agilent solutions, reinforcing that buyer demand spans both life-science and applied-lab environments. | 中 | SM017 |
| CM024 | NIH says it invests nearly USD 48 billion in medical research and that about 82% of that budget funds extramural research distributed across almost 50,000 grants and more than 2,500 institutions. | 中 | SM023 |
| CM025 | CRS estimates the federal FY2026 R&D request at approximately USD 181.4 billion, showing a large public research backdrop but one that is mission-driven and agency-specific rather than a single commercial buyer pool. | 高 | SM021, SM012 |
| CM026 | In an integrated AI science factory deployment, the economic buyer is most plausibly the platform-R&D or lab-operations owner, while users include bench scientists, automation engineers, and computational scientists. | 中 | SM006, SM010, SM015, SM018 |
| CM027 | High-throughput screening demand is a primary adoption driver in lab automation. | 中 | SM001, SM004 |
| CM028 | Labor scarcity and the need to reduce manual intervention are direct drivers of laboratory automation adoption. | 中 | SM001, SM002 |
| CM029 | Regulatory demands, data integrity requirements, and the shift to cloud-native platforms are core drivers of laboratory informatics adoption. | 中 | SM006, SM007 |
| CM030 | AI drug discovery budgets are supported by pressure to compress multiyear discovery cycles and by the high cost of commercializing a molecule, which Mordor summarizes at roughly USD 2.6 billion on average. | 中 | SM010 |
| CM031 | IQVIA says biopharmaceutical R&D remained resilient in 2025 but that growing scientific complexity and longer timelines are putting renewed pressure on productivity. | 中 | SM013 |
| CM032 | Current scientific reviews say early self-driving labs were constrained by limited scope, poor interoperability, and reliance on human-curated heuristics. | 高 | SM014, SM016 |
| CM033 | Materials-science self-driving-lab literature argues that traditional discovery-to-market timelines of 10 to 20 years are too slow for important technology domains. | 高 | SM015, SM016 |
| CM034 | Bruker’s 2026 Chemspeed/SciY launch says many labs still face siloed tools and integration gaps in heterogeneous environments that limit efficiency and scalability. | 中 | SM018 |
| CM035 | Legacy-system integration is a leading challenge in lab automation adoption. | 中 | SM001 |
| CM036 | High upfront investment and unclear ROI remain material barriers to automation adoption, especially outside the largest labs. | 中 | SM001, SM002, SM004 |
| CM037 | Implementation cost and data-security concerns remain material constraints in laboratory informatics adoption. | 中 | SM006, SM007 |
| CM038 | Hype risk is a real adverse factor in AI drug discovery: STAT quotes Insitro CEO Daphne Koller warning that people expect breakthroughs to happen “tomorrow.” | 中 | SM019 |
| CM039 | Public evidence does not show that autonomous labs are yet purchased as a stable standalone budget line; buyers more often assemble instruments, informatics, and services separately. | 中 | SM001, SM006, SM015, SM018 |
| CM040 | MarketsandMarkets identifies Thermo Fisher, Danaher, Agilent, Tecan, and Roche among the key lab automation incumbents. | 中 | SM001 |
| CM041 | Specialists such as Automata appear inside broader lab automation coverage, implying that newer vendors still compete inside stacks defined by larger incumbents. | 中 | SM001 |
| CM042 | Lila’s competitive context is fragmented across automation incumbents, informatics platforms, AI drug discovery software, and self-driving-lab orchestration specialists rather than one neatly bounded peer set. | 中 | SM001, SM006, SM010, SM015, SM016, SM018 |
| CM043 | Pharma and biotech represent the clearest initial SAM because analyst segmentation, company market mix, and productivity pressure all converge there. | 中 | SM006, SM010, SM013, SM020 |
| CM044 | Materials and chemistry discovery are strategically relevant but harder to size through standard market reports, making them a second wedge rather than the whole serviceable market. | 中 | SM015, SM016, SM017 |
| CM045 | The most important remaining diligence asks are Lila’s ACV by customer type, software-versus-automation-versus-services mix, implementation duration, renewal behavior, and evidence of expansion from pilot into broader factory deployments. | 中 | SM001, SM006, SM010, SM018 |
| CM051 | Commercial adoption maturity is uneven: large pharma and specialized CRO programs are more likely than academic/government or diagnostics buyers to support scaled full-factory deployments because their budgets are more concentrated and ROI can be measured at program level. | 中 | SM006, SM020, SM023 |
| CM046 | Across adjacent market reports, North America is typically the current revenue leader while Asia-Pacific is the faster-growing region. | 中 | SM001, SM003, SM004, SM006 |
| CM047 | MarketsandMarkets says drug discovery accounted for 39.0% of the lab automation market in 2025. | 中 | SM001 |
| CM048 | Mordor says cloud-based platforms held 58.35% of laboratory informatics spending in 2025. | 中 | SM006 |
| CM049 | Mordor says LIMS accounted for 51.42% of laboratory informatics spending in 2025. | 中 | SM006 |
| CM050 | Mordor says target identification and validation held 28.43% of AI drug discovery spending in 2025, while de novo design is one of the fastest-growing use cases. | 中 | SM010 |
| CP001 | Lila says its operating system for science autonomously generates hypotheses, designs experiments, runs them, and learns from results in real time. | 高 | SP001, SP003 |
| CP002 | Lila publicly frames itself as a single general platform for autonomous science rather than a set of narrow domain tools. | 高 | SP002, SP003 |
| CP003 | Lila's public framing pairs an advanced AI model as the brain with proprietary AI Science Factory instruments as the body. | 中 | SP001 |
| CP004 | Flagship said Lila launched with $200 million in committed seed capital in March 2025. | 中 | SP003 |
| CP005 | Recursion says its operating system combines proprietary biological and chemical datasets with automated wet labs that capture millions of cell experiments per week. | 高 | SP004, SP005 |
| CP006 | Recursion says it has generated more than 50 petabytes of proprietary biological and chemical data. | 中 | SP004 |
| CP007 | Recursion's public platform description spans CRISPR perturbation, high-throughput screening, transcriptomics, generative AI design, and feedback loops into molecule optimization. | 中 | SP005 |
| CP008 | Recursion's acquisition materials say Exscientia adds precision chemistry tools and automated small-molecule synthesis to Recursion's scaled biology and translational capabilities. | 高 | SP006, SP007, SP008 |
| CP009 | The Recursion-Exscientia deal materials framed the combined company as a full-stack or end-to-end small-molecule drug-discovery platform with about $850 million of combined cash at Q2 2024. | 高 | SP007, SP008 |
| CP010 | Public Recursion-Exscientia materials stay centered on small-molecule therapeutics rather than Lila's broader biology-chemistry-materials science-factory ambition. | 中 | SP008, SP009 |
| CP011 | Insilico markets Pharma.ai as generative AI and automation for drug discovery, scientific research, and sustainability. | 中 | SP010 |
| CP012 | Insilico says it is using AI to create an AI-driven drug-discovery pipeline from A to Z. | 中 | SP010 |
| CP013 | Insilico's public platform materials map work from target identification through hit-to-lead, lead optimization, IND-enabling, Phase I, and Phase II programs. | 中 | SP010 |
| CP014 | Insilico says it has collaborations with 10 of the top 20 global pharmaceutical companies by 2021 reported sales. | 中 | SP011 |
| CP015 | Isomorphic Labs says it is building predictive and generative AI models to accelerate drug discovery at digital speed. | 中 | SP012 |
| CP016 | Isomorphic's public narrative aims to solve disease through digital biology and AI drug design rather than through a cross-domain autonomous lab platform. | 中 | SP012, SP013 |
| CP017 | Isomorphic's public partner materials show distribution through Novartis, Lilly, and Johnson & Johnson rather than an open platform or self-serve commercial model. | 高 | SP013, SP014 |
| CP018 | PR Newswire said Lilly agreed to pay Isomorphic Labs $45 million upfront with up to $1.7 billion in milestone payments for a multi-target collaboration. | 中 | SP014 |
| CP019 | Isomorphic's news page listed a $600 million external investment round in June 2025. | 中 | SP015 |
| CP020 | Benchling markets a cloud-based notebook and data platform that digitizes labs, automates workflows, and exposes AI tools rather than autonomously running the full scientific method. | 高 | SP016, SP017 |
| CP021 | Benchling emphasizes open integrations, custom apps, and adaptable science workflows, making it a modular infrastructure substitute to a closed end-to-end factory. | 中 | SP016 |
| CP022 | Benchling Solutions says it has completed thousands of successful implementations and covers end-to-end R&D processes like experiment tracking, sample management, inventory, and process management. | 中 | SP017 |
| CP023 | Benchling's customer materials say the platform is trusted by 1,200 or more leading biotech organizations. | 中 | SP018 |
| CP024 | Benchling's AstraZeneca customer quote says the platform turned manual processes and in-house tools into fully automated steps, showing that pharma teams can build internal digital-science stacks on neutral software infrastructure. | 中 | SP018 |
| CP025 | Arcadia says it was founded in 2021 to rethink the entire research cycle and make biological discovery more systematic. | 中 | SP019 |
| CP026 | Arcadia says it releases apps, software pipelines, protocols, and other resources to the scientific community as it develops its platform. | 中 | SP020 |
| CP027 | TechCrunch described OpenBioML as an open research laboratory applying machine learning to DNA sequencing, protein folding, and computational biochemistry. | 中 | SP022 |
| CP028 | OpenBioML leaders said they want large-scale collaborations backed by compute resources normally available only to the largest industrial labs. | 中 | SP022 |
| CP029 | OpenBioML's GitHub organization shows an open-source portfolio of public repositories spanning datasets, biochemical language models, evaluation harnesses, and RL-OED workflows, but no evidence in this source set of integrated wet-lab execution. | 中 | SP021, SP022 |
| CP030 | Opentrons says labs can use the assays, instruments, and AI tools they want without being forced into a closed system. | 中 | SP023 |
| CP031 | Opentrons markets reconfigurable hardware, workflows, and throughput so labs can change automation setups without starting over. | 中 | SP023, SP024 |
| CP032 | Drug Discovery Trends said at least 15 companies were vying to become the operating-system layer for AI-enabled labs at SLAS 2026. | 中 | SP025 |
| CP033 | The same SLAS 2026 article said OpenAI and Ginkgo Bioworks ran more than 36,000 experiments in an autonomous lab campaign, showing that cloud-lab plus AI combinations can approximate parts of the science-factory promise without one vertically integrated vendor. | 中 | SP025 |
| CP034 | Royal Society Open Science said current self-driving labs can automate nearly the entire scientific method, but fully autonomous Level-5 AI researcher systems have not yet been realized. | 中 | SP026 |
| CP035 | UChicago researchers argued for an AI-advisor model in which humans and machines share the driver’s seat in autonomous labs rather than ceding leadership entirely to the machine. | 中 | SP027 |
| CP036 | Northwestern researchers argued that megalibraries can generate data and candidate materials faster than iterative self-driving labs in some materials-discovery workflows. | 中 | SP028 |
| CP037 | Genentech says it has made AI a core part of discovery through a lab-in-a-loop process where lab and clinic data feed models that generate hypotheses and molecules, then experiments feed back into the models. | 中 | SP030 |
| CP038 | Genentech says it is building a next-generation drug-discovery platform using decades of lab and clinical data together with NVIDIA-enabled generative AI. | 中 | SP030 |
| CP039 | Recursion-Exscientia, Insilico, and Isomorphic Labs are the closest direct overlaps to Lila because all market AI-enabled therapeutic discovery, but each public narrative is narrower than Lila's cross-domain science-factory pitch. | 中 | SP005, SP010, SP012, SP013 |
| CP040 | Benchling, Opentrons, and OpenBioML represent a modular substitute path that can cover informatics, automation, and model/community layers without adopting one closed general platform. | 中 | SP016, SP021, SP023 |
| CP041 | Internal pharma AI programs and alliance-heavy competitors shift distribution power away from a standalone science-factory vendor because buyers can build or co-build inside existing R&D organizations. | 中 | SP013, SP018, SP030 |
| CP042 | Lila's clearest public differentiation is the claim to one general autonomous platform spanning idea generation through experiment execution across multiple scientific domains. | 中 | SP001, SP002, SP003 |
| CP043 | Lila's biggest competitive risk is that buyers may prefer narrower validated stacks, modular orchestration layers, or internal builds over one closed general platform. | 中 | SP025, SP026, SP030 |
| CP044 | Lila's public sources reviewed here do not disclose named external customers, public pricing tiers, or source-backed throughput metrics for its autonomous labs. | 中 | SP001, SP002, SP003 |
| CP045 | Recursion-Exscientia, Insilico, and Isomorphic all appear to monetize primarily through partnered drug programs, pipelines, or milestone economics rather than transparent self-serve software pricing. | 中 | SP006, SP010, SP013, SP014 |
| CP046 | Arcadia, OpenBioML, and other open efforts pressure closed systems mainly on openness, talent attraction, and tool/community diffusion rather than on industrialized end-to-end wet-lab execution. | 中 | SP020, SP021, SP022 |
| CI001 | Flagship unveiled Lila Sciences in March 2025 with $200M of committed seed capital. | 高 | SI003, SI006 |
| CI002 | Lila announced a $235M Series A first close in September 2025 co-led by Braidwell and Collective Global. | 高 | SI002, SI016, SI028 |
| CI003 | Lila added $115M in October 2025 in a round extension that included NVentures, Nvidia’s venture arm. | 高 | SI003, SI014, SI015 |
| CI004 | The two 2025 closes brought Lila’s Series A total to $350M. | 高 | SI003, SI014, SI015, SI017 |
| CI005 | Lila’s disclosed capital raised reached $550M across its $200M seed and $350M Series A. | 高 | SI003, SI014, SI017, SI019 |
| CI006 | Bloomberg reported that Lila’s September 2025 round valued the company at roughly $1.23B. | 中 | SI016 |
| CI007 | Reuters reported that the October 2025 extension lifted Lila’s valuation to more than $1.3B. | 高 | SI014, SI017 |
| CI008 | Forge displayed a $1.42B Series A valuation snapshot for Lila in 2026. | 中 | SI021 |
| CI009 | Lila says it is welcoming its first cohort of customers now. | 中 | SI003 |
| CI010 | Reuters said Lila plans to offer enterprise software access to its AI models and automated labs. | 中 | SI014 |
| CI011 | Sacra described Lila’s current monetization as project-based discovery programs for research-intensive customers. | 中 | SI019 |
| CI012 | Sacra said Lila also plans to introduce subscription or usage-based lab-as-a-service access. | 中 | SI019 |
| CI013 | Flagship said Lila’s platform will be open to partners across the life and material sciences industries. | 中 | SI006 |
| CI014 | No reviewed official or market-data source disclosed public list pricing, ACV, or standard contract terms for Lila’s offerings. | 中 | SI001, SI003, SI019, SI020, SI021 |
| CI015 | No reviewed public source disclosed revenue, ARR, or active paying-customer count for Lila. | 中 | SI001, SI003, SI014, SI019, SI020, SI021 |
| CI016 | Lila is expanding AI Science Factories and teams across Boston or Cambridge, San Francisco, and London. | 高 | SI002, SI003, SI023, SI026 |
| CI017 | Reuters reported that Lila signed a 235,500-square-foot Cambridge lease, one of Greater Boston’s largest lab leases of 2025. | 高 | SI014, SI017 |
| CI018 | Lila’s Director of Facilities role covers multi-site budgets, capital planning, vendor governance, KPI reporting, and renovations or expansions. | 中 | SI012 |
| CI019 | Lila’s Facilities Support role references process gases, lab water and air systems, wastewater, loading docks, and heavy-equipment handling. | 中 | SI013 |
| CI020 | Job boards show Lila hiring across AI research, lab operations, product, partnerships, enterprise sales, and government affairs. | 中 | SI023, SI026, SI027 |
| CI021 | Flagship and AWS said Lila is among the companies using AWS cloud and AI support, implying meaningful compute infrastructure needs. | 中 | SI007 |
| CI022 | Sacra said customers use Lila’s platform to avoid building their own AI and automation capabilities. | 中 | SI019 |
| CI023 | Lila’s likely cost stack combines facilities, robotics and lab equipment, compute or cloud, scientific labor, and compliance or vendor management. | 中 | SI012, SI013, SI014, SI021 |
| CI024 | Fierce Biotech wrote that Lila has not yet publicly released data supporting several breakthrough claims. | 中 | SI015 |
| CI025 | Industry Examiner argued that the model is capital-hungry and that margins will depend on utilization, low rerun rates, and standardization rather than custom consulting. | 中 | SI017 |
| CI026 | Industry Examiner said proof of economics would require named reference accounts, capacity metrics, conversion rates, and time-to-project-start evidence. | 中 | SI017 |
| CI027 | Public sources reviewed do not disclose gross margin, CAC, payback, retention, or customer concentration. | 中 | SI014, SI015, SI017, SI019 |
| CI028 | Public sources reviewed do not disclose current cash, monthly burn, or runway. | 中 | SI003, SI014, SI019, SI020, SI021 |
| CI029 | Public sources reviewed do not name paying customers or publish measurable commercial ROI outcomes. | 中 | SI003, SI014, SI017, SI019 |
| CI030 | Nasdaq Private Market and Forge still present Lila as a private or pre-IPO company rather than a public issuer. | 中 | SI020, SI021 |
| CI031 | SEC and NASAA filings show AVSF - Lila Sciences 2025, LLC as a Delaware pooled investment fund filed in late September 2025. | 高 | SI024, SI025 |
| CI032 | The Form D disclosed a $817,500 offering amount and named Alumni Ventures as the issuer’s sole manager. | 高 | SI024, SI025 |
| CI033 | The Form D structure indicates that at least one feeder or syndication vehicle participated around the 2025 financing process. | 中 | SI024, SI025 |
| CI034 | Official fundraising materials say the new capital is earmarked for AI Science Factory buildout, commercial partner opening, and hiring. | 高 | SI002, SI003, SI006 |
| CI035 | The 2025 syndicate blended healthcare and science investors, deep-tech VCs, strategic technology capital, and institutional asset owners. | 中 | SI002, SI003, SI008, SI009, SI010, SI011 |
| CI036 | Near-term financing risk appears lower than execution risk because Lila raised $550M before disclosing public operating metrics. | 中 | SI003, SI014, SI015, SI019 |
| CI037 | Revenue quality today is better described as prospective and partner-led than as proven recurring software. | 中 | SI003, SI014, SI019 |
| CI038 | If enterprise software access remains tied to custom scientific programs and physical factory throughput, gross margins may trail pure-software benchmarks. | 中 | SI014, SI017, SI019 |
| CI039 | High utilization of factory capacity is likely necessary to absorb fixed lease, equipment, and staffing costs. | 中 | SI012, SI013, SI014, SI017 |
| CI040 | No reviewed public source disclosed debt facilities or project-finance obligations. | 低 | SI003, SI014, SI019, SI020, SI021 |
| CE001 | Lila describes itself as the world's first scientific superintelligence platform and autonomous lab for life, chemistry, and materials science. | 高 | SE019, SE020 |
| CE002 | Flagship says Lila combines an AI platform with fully autonomous labs that assist scientists in designing and conducting new experiments. | 高 | SE019, SE021 |
| CE003 | Lila says it is training a scientific reasoning model on experiment-generated evergreen tokens rather than exhausted internet data. | 中 | SE001 |
| CE004 | Lila's public architecture pairs scale verifiers and scientific tools with autonomous design workflows and continuous policy optimization. | 中 | SE001 |
| CE005 | Lila says its model learns the scientific method across DNA, RNA, proteins, molecules, cells, surfaces, nano, pores, coatings, and catalysts. | 中 | SE001 |
| CE006 | Lila says AI Science Factories are an extensible network of instruments built for AI-driven scientific discovery. | 中 | SE001 |
| CE007 | The tech page names molecular dynamics simulators, protein structure predictors, quantum chemistry solvers, gene editors, and robotic lab workflows as scientific tools in the loop. | 中 | SE001 |
| CE008 | Lila's solutions page says AI-driven discovery and physical experimentation operate as one on-demand resource. | 中 | SE002 |
| CE009 | Catalyst gives partner teams direct access to Lila Iris, AI Science Factories, and scientific experts. | 中 | SE003 |
| CE010 | Catalyst is positioned as Lab-as-a-Service that converts fixed lab capacity and capex into on-demand experimental throughput. | 中 | SE003 |
| CE011 | Creation uses Lila Iris and AI Science Factories to generate hypotheses, design experiments, run them, and iteratively optimize candidates. | 中 | SE004 |
| CE012 | Creation promises validated assets, including structures, protocols, and data packages, rather than insight reports alone. | 中 | SE004 |
| CE013 | Lila says Creation campaigns can produce new molecules, materials, or platforms with validated science, IP, and de-risked technical roadmaps. | 中 | SE004 |
| CE014 | Lila's about page says the company is building one general platform for autonomous science rather than many narrow domain tools. | 中 | SE005 |
| CE015 | Lila's about page says the platform is intended to accelerate discovery across medicine, materials, energy, and defense. | 中 | SE005 |
| CE016 | Lila says its culture is guided by safety, human impact, and scientific rigor rather than reckless experimentation. | 中 | SE005 |
| CE017 | The therapeutics page says the platform hypothesizes, experiments, and refines while generating verified real-world data each iteration. | 中 | SE006 |
| CE018 | Lila says its therapeutics workflows cover genetic medicines across programmable payloads, delivery vehicles, potency, durability, safety, and manufacturability. | 中 | SE006 |
| CE019 | Lila says its therapeutics workflows also cover antibody and ligand engineering across binding, specificity, stability, solubility, aggregation risk, and expression. | 中 | SE006 |
| CE020 | The biotech page says Lila couples AI models with autonomous experimentation to design, test, and refine biology products and workflows. | 中 | SE007 |
| CE021 | The biotech page says Lila compresses innovation and development cycles from months into weeks. | 中 | SE007 |
| CE022 | Lila says its biotech workflows optimize constructs, parts, libraries, host systems, expression platforms, and formulation conditions. | 中 | SE007 |
| CE023 | The biotech page says integrated platforms translate novel methods into reliable high-throughput systems under real manufacturing constraints. | 中 | SE007 |
| CE024 | The chemicals page says Lila combines molecular design, computational modeling, and high-throughput experimentation to engineer chemicals and fuels. | 中 | SE008 |
| CE025 | The chemicals page says Lila explores large materials spaces to build predictive models for catalyst activity, selectivity, and stability. | 中 | SE008 |
| CE026 | The chemicals page says Lila can select reactor formats and test candidates in devices under commercially aligned conditions. | 中 | SE008 |
| CE027 | The advanced materials page highlights discovery of durable coatings and critical infrastructure components, including extreme-environment thin films. | 中 | SE009 |
| CE028 | The energy and environment page adds electrocatalysts, rare-earth-free magnets, sorbents, and catalyst optimization to the public program map. | 中 | SE010 |
| CE029 | Julie Shah serves as Chief Robotics Officer at Lila Sciences and brings a background in human-robot collaboration across manufacturing, healthcare, transportation, and defense. | 高 | SE012, SE029 |
| CE030 | Milad Abolhasani's Lila profile says he leads chemistry efforts spanning self-driving labs, autonomous experimentation, flow chemistry, microfluidics, multimodal analytics, robotics, and autonomous science. | 中 | SE013 |
| CE031 | Rafael Gómez-Bombarelli's Lila profile says he leads AI for chemistry and materials across experimental data and physics-based simulations. | 中 | SE014 |
| CE032 | Kenneth Stanley leads open-ended discovery and creativity methods for AI systems at Lila. | 中 | SE015 |
| CE033 | Greenhouse listings show current hiring across foundation models for life sciences, frontier capabilities, AI safety, protein engineering, ML research, AI data, and autonomous science for cell biology. | 中 | SE022 |
| CE034 | CareersInRobotics listings show Lila hiring for robotics program management, simulation engineering, robotics engineering, dexterous manipulation, and robotics scientist roles. | 中 | SE023 |
| CE035 | CareersInRobotics role tags mention simulation-to-real, MoveIt, LiDAR, SLAM, Gazebo, PyBullet, NVIDIA Isaac Sim, and NVIDIA Omniverse. | 中 | SE023 |
| CE036 | Lila's Series A announcement says the company has raised $350 million in Series A financing and $550 million total capital. | 高 | SE016, SE024 |
| CE037 | Lila's Series A announcement says NVentures, NVIDIA's venture arm, is among the new investors. | 高 | SE016, SE024 |
| CE038 | Lila says the new investors bring technical collaborations to accelerate global growth plans. | 中 | SE016 |
| CE039 | Lila says the new capital will scale AI Science Factories through more instruments under AI control than any company on earth. | 中 | SE016 |
| CE040 | Lila says it is opening the platform to commercial partners and welcoming its first cohort of customers in strategic scientific domains. | 中 | SE016 |
| CE041 | Flagship says Lila was founded in 2023 inside Flagship labs and launched publicly in March 2025 with $200 million in seed capital to build the first AI Science Factories. | 高 | SE019, SE021 |
| CE042 | Geoffrey von Maltzahn said the hard problem is enabling AI to run each step from idea generation to reduction to practice with robotics and automation. | 中 | SE019 |
| CE043 | Industry Examiner says Lila added $115 million to the Series A, reached a valuation above $1.3 billion, and planned a 235,500-square-foot Cambridge site. | 中 | SE024 |
| CE044 | Industry Examiner says Lila is positioning AI Science Factories as discovery capacity for customers beyond biotech, including pharma, chipmakers, and energy groups. | 中 | SE024 |
| CE045 | Excedr says Lila is trying to teach AI to make discoveries through autonomous AI labs rather than build another text or image model. | 中 | SE025 |
| CE046 | MIT DMSE says Lila is at the forefront of AI-directed automated labs that plan, run, and analyze materials experiments to shorten discovery timelines from decades to years or less. | 中 | SE026 |
| CE047 | BioPharmaTrend says Lila's platform combines AI models, robotics, and custom software to automate the scientific method from hypothesis generation through learning from results. | 中 | SE027 |
| CE048 | BioPharmaTrend says the first AI Science Factory had already run hundreds of thousands of AI-driven experiments across life sciences, chemistry, and materials science. | 中 | SE027 |
| CE049 | The Nature self-driving labs review cites Abolhasani's work on universal self-driving laboratories as part of the core literature for autonomous experimentation. | 中 | SE028 |
| CE050 | Catalyst and Creation pages both advertise a 900-fold increase in experimental validation for Lila's DNA Design agent and cite 100% agent performance. | 高 | SE003, SE004 |
| CE051 | Lila's website privacy policy says it uses physical, technical, and organizational measures and need-based access controls to protect website personal data. | 中 | SE017 |
| CE052 | Lila's candidate privacy notice says recruiting-data controls include access controls, role-based permissions, encryption in transit and at rest, anomaly monitoring, and regular security reviews of third-party recruiting tools. | 中 | SE018 |
| CE053 | The public materials reviewed here do not name product-level certifications, regulated quality systems, public uptime targets, or a public status page for AI Science Factories. | 低 | SE005, SE011, SE016, SE017, SE018 |
| CU001 | Lila says its scientific superintelligence is meant to serve customer programs and discovery challenges across multiple industries. | 高 | SU001, SU003 |
| CU002 | Public-facing materials present Lila as on-demand scientific infrastructure rather than a single finished application. | 高 | SU001, SU003, SU011 |
| CU003 | Lila publicly offers two commercial modes: Catalyst for platform access and Creation for end-to-end campaign delivery. | 高 | SU011, SU012 |
| CU004 | Catalyst is positioned as access to Lila Iris, AI Science Factories, and scientific experts for existing programs. | 中 | SU011 |
| CU005 | Creation is positioned for investors or strategic partners that want validated assets, IP, and a de-risked technical roadmap. | 中 | SU012 |
| CU006 | Lila says customers can access AI-driven discovery without funding and building their own full lab stack. | 高 | SU003, SU011 |
| CU008 | Flagship said at launch that the Lila platform would be open to partners across life and material sciences. | 高 | SU015, SU016 |
| CU009 | BioPharma Dive reported that Lila does not plan to develop its own therapeutic candidates. | 中 | SU018 |
| CU010 | BioPharma Dive reported that Lila plans to partner with other Flagship startups and outside biotech companies. | 中 | SU018 |
| CU011 | Lila’s team page lists dedicated commercialization roles including Chief Revenue & Product Officer, Business Development, and Corporate Development leadership. | 中 | SU005 |
| CU012 | Reuters reported that Lila planned to open its platform to commercial customers through enterprise software and automated labs. | 高 | SU020, SU021 |
| CU013 | Reuters reported that Lila had interest from firms in energy, semiconductors, and drug development but did not name any specific companies. | 高 | SU021, SU024 |
| CU014 | Fierce Biotech said the October 2025 financing would help bring in Lila’s first customers. | 中 | SU020 |
| CU015 | No reviewed public source names a paying external customer, pilot partner, procurement win, or case-study reference account as of the run date. | 中 | SU011, SU018, SU020, SU021, SU023 |
| CU016 | Lila’s therapeutics page targets genetic medicines, antibodies, ligands, and small molecules. | 中 | SU006 |
| CU017 | Lila’s biotech page targets bioprocessing, reagents, assays, and scalable production workflows under manufacturing constraints. | 中 | SU007 |
| CU018 | Lila’s chemicals page targets sorbents and catalyst discovery under commercially aligned conditions. | 中 | SU008 |
| CU019 | Lila’s advanced materials page targets extreme-environment coatings and infrastructure-oriented materials. | 中 | SU009 |
| CU020 | Lila’s energy and environment page targets electrocatalysts, rare-earth-free magnets, sorbents, and catalysts tested under commercially aligned conditions. | 中 | SU010 |
| CU021 | Lila says its commercial product can run on top of a customer’s existing data and platforms without a broad IT transformation. | 中 | SU013 |
| CU022 | Lila says it aims to make each customer’s R&D dollars and team much more efficient. | 中 | SU013 |
| CU023 | Lila’s tech page says frontier science should become possible without building a full in-house R&D organization. | 高 | SU003, SU004 |
| CU024 | March Capital said it had worked with Geoffrey von Maltzahn through Generate Biomedicines and Tessera Therapeutics before backing Lila. | 中 | SU022 |
| CU025 | March Capital said Lila is opening its platform to partners across healthcare, materials, energy, and national resilience. | 高 | SU020, SU022 |
| CU026 | The combination of Flagship origin, outside-biotech partnering language, and March Capital’s Generate/Tessera ties makes Flagship ecosystem companies the likeliest early users, but public proof of actual usage is absent. | 低 | SU015, SU018, SU022 |
| CU027 | Lila’s public ICP spans enterprise R&D teams in pharma, biotech, chemicals, materials, energy, and related industrial sectors. | 高 | SU001, SU006, SU007, SU008, SU009, SU010 |
| CU028 | The public go-to-market looks enterprise-led rather than self-serve because Lila emphasizes partnerships, Lab-as-a-Service, custom campaigns, and direct contact CTAs. | 高 | SU001, SU003, SU011, SU012 |
| CU029 | No public pricing, marketplace listing, or broad user-review footprint appears in the reviewed materials. | 中 | SU001, SU003, SU011, SU012 |
| CU030 | No public customer counts, deployment counts, active-user counts, or booked-throughput metrics were found in the reviewed materials. | 中 | SU011, SU012, SU020, SU021, SU023 |
| CU031 | No public NRR, GRR, churn, renewal-rate, contract-length, or satisfaction metrics were found in the reviewed materials. | 中 | SU011, SU012, SU021, SU023 |
| CU032 | The first visible commercialization milestones are productizing offerings and expanding factory capacity, not publishing reference accounts. | 中 | SU011, SU012, SU020, SU021 |
| CU033 | If early revenue comes first from Flagship-linked programs or a handful of bespoke projects, concentration risk could be high until independent reference accounts appear. | 低 | SU018, SU022, SU023 |
| CU034 | Industry Examiner argues Lila still has to define productized units of work that procurement teams can actually buy. | 中 | SU023 |
| CU035 | Industry Examiner says first non-biopharma reference accounts and published capacity metrics would be real proof points for the model. | 中 | SU023 |
| CU036 | Industry Examiner says factory economics are sensitive to utilization, reruns, and excessive custom work. | 中 | SU023 |
| CU037 | Reuters said partners rather than Lila will bring molecules into clinical trials or scale new energy breakthroughs. | 高 | SU021, SU024 |
| CU038 | Lila’s customer value proposition therefore sits primarily in upstream discovery acceleration rather than downstream product commercialization. | 中 | SU018, SU021, SU023 |
| CU039 | The commercialization team buildout implies Lila is assembling sales and product infrastructure ahead of public customer disclosure. | 中 | SU005, SU020 |
| CU040 | Fierce and TechStartups both frame the 2025 financing around factory buildout and first-customer acquisition rather than existing customer traction. | 中 | SU020, SU024 |
| CU041 | The current customer-quality verdict is promising target-market breadth with extremely limited public adoption proof. | 高 | SU001, SU011, SU021, SU023 |
| CU042 | The most credible external-customer path is to sell platform access or discovery campaigns into enterprise R&D and let partners advance outputs downstream. | 高 | SU011, SU012, SU018, SU021 |
| CU043 | Lila’s 2026 blog continues to market Creation as a route to launch products and create new companies. | 高 | SU012, SU014 |
| CU044 | Public materials blur the line between customer acquisition and venture creation, making repeat-revenue quality hard to underwrite from outside. | 中 | SU012, SU014, SU023 |
| CR001 | Lila says its platform uses advanced AI and autonomous labs to generate hypotheses, design and run experiments, and learn from new data in real time. | 中 | SR001 |
| CR002 | Lila describes its system as an advanced AI model paired with proprietary AI Science Factory instruments, implying a tightly coupled software-and-lab stack rather than a software-only tool. | 中 | SR001 |
| CR003 | Lila publicly claims that its system consistently outperforms other models across scientific domains. | 中 | SR001 |
| CR004 | Fierce Biotech reported that Lila had not publicly released data supporting its claims about scientific reasoning, genetic medicine constructs, or newly generated binders. | 中 | SR010 |
| CR005 | Flagship's launch announcement says Lila was founded in Flagship's labs in 2023. | 中 | SR008 |
| CR006 | Lila's Series A announcement says total capital raised reached $550 million after a $350 million Series A. | 中 | SR002, SR008 |
| CR007 | Lila says the new capital will accelerate AI Science Factory buildout and open its platform to commercial partners. | 中 | SR002 |
| CR008 | Lila said in its Series A post that it was welcoming its first cohort of customers, but the post did not name customers or disclose revenue. | 中 | SR002 |
| CR009 | Lila's advanced-materials page says it is targeting use cases from durable coatings to critical infrastructure components. | 中 | SR003 |
| CR010 | Across its homepage, materials page, and Flagship profile, Lila presents itself as spanning life science, chemistry, materials, energy and environment, aerospace and defense, and biotech rather than a single beachhead market. | 中 | SR001, SR003, SR007 |
| CR011 | Lila's Greenhouse board shows open roles in AI safety, AI safety technical mitigations, AI data, protein engineering, autonomous science for cell biology, and frontier capabilities. | 中 | SR011 |
| CR012 | Lila's Greenhouse board lists roles across Cambridge, London, and San Francisco. | 中 | SR011, SR007 |
| CR013 | The breadth of open scientific, engineering, safety, and program-management roles implies that core operating capacity is still being assembled publicly. | 低 | SR011 |
| CR014 | NIST says AI risk management should address risks to individuals, organizations, and society across the design, development, use, and evaluation of AI systems. | 中 | SR017 |
| CR015 | NIST highlights a generative-AI profile because frontier models create risk-management issues beyond the base AI RMF. | 中 | SR017 |
| CR016 | NIH biosafety policy says research involving recombinant or synthetic nucleic acid molecules requires specific safety practices and containment procedures under the NIH Guidelines. | 中 | SR023 |
| CR017 | The Center for Health Security says AI models trained on sensitive biological datasets create a dual-use risk and that a regulatory gap exists for governing this information-based risk. | 中 | SR027 |
| CR018 | RAND says rapid AI and biotechnology development creates biosecurity risks that current global treaties and data systems cannot sufficiently address. | 中 | SR026 |
| CR019 | Lila's privacy policy says the company may collect personal information, IP addresses, usage details, and cookies and references GDPR and the UK Data Protection Act 2018. | 中 | SR005 |
| CR020 | Lila's privacy policy says personal data may be transferred to the United States and other jurisdictions and disclosed to comply with court orders, laws, or regulatory requests. | 中 | SR005 |
| CR021 | Lila's terms say website use is governed by Massachusetts law and disputes are subject to Suffolk County, Massachusetts courts. | 中 | SR006 |
| CR022 | Lila's terms say the site content is provided as-is, disclaim warranties, and cap aggregate liability at fifty dollars. | 中 | SR006 |
| CR023 | The EDPS says AI systems depend on ever-larger datasets and monitoring of human behaviour, creating privacy and data-protection challenges. | 中 | SR025 |
| CR024 | HHS presents HIPAA as part of the laws and regulations that govern health information and privacy in the United States. | 中 | SR024 |
| CR025 | FDA says most drugs that undergo preclinical testing never reach human testing, and the few that do face rigorous review of trial design, side effects, and manufacturing. | 中 | SR019 |
| CR026 | The Wyss Institute says traditional drug discovery typically takes 13 to 15 years, fewer than 10% of Phase I candidates are approved, and average R&D investment exceeds $2.5 billion. | 中 | SR021 |
| CR027 | UCSF QBI says industrial estimates put the cost of bringing a drug to market at about $4 billion and require a vertically integrated research enterprise. | 中 | SR022 |
| CR028 | The PMC review describes biotechnology product development as a business with very high failure rates, high and rising costs, and extended timelines. | 中 | SR020 |
| CR029 | The National Academies' reproducibility report shows that reproducibility and replicability remain live scientific-system challenges rather than solved problems. | 中 | SR018 |
| CR030 | Fierce Biotech reported that Lila had not publicly released data to substantiate several marquee technical claims as of its October 2025 fundraising coverage. | 中 | SR010 |
| CR031 | Recursion says it has over a decade of AI-drug-discovery work, strategic partnerships, and an advanced pipeline. | 中 | SR012 |
| CR032 | Isomorphic Labs says it is using predictive and generative AI models built on and beyond AlphaFold to transform drug discovery. | 中 | SR013 |
| CR033 | Insilico Medicine publicly markets programs ranging from target identification through Phase II and emphasizes generative AI plus automation. | 中 | SR015 |
| CR034 | Absci says it has internal and partnered programs and claims an AI-designed antibody advanced from concept toward the clinic in 24 months. | 中 | SR016 |
| CR035 | CuspAI publicly positions itself as an AI materials company with a high-profile scientific leadership and advisor bench. | 中 | SR014 |
| CR036 | The presence of specialized peers in AI drug discovery and AI materials means Lila is competing against companies with narrower scopes and more specific proof points. | 中 | SR012, SR013, SR014, SR015, SR016 |
| CR037 | Lila's public materials and partner pages say the company is growing teams in Cambridge, San Francisco, and London while building AI Science Factories. | 中 | SR002, SR003, SR007 |
| CR038 | Building AI Science Factories plus global multidisciplinary teams implies heavy capital needs before durable commercial proof appears, even after $550 million raised. | 中 | SR002, SR007, SR011, SR021 |
| CR039 | Lila has public legal and privacy pages and visible AI-safety hiring, but it does not publicly show named customer outcomes, benchmark datasets, or detailed biosecurity controls. | 低 | SR001, SR005, SR006, SR010, SR011 |
| CR040 | Because Lila is simultaneously pursuing therapeutics and advanced materials, it must clear very different validation and commercialization pathways before investors can underwrite repeatability at scale. | 中 | SR003, SR019, SR021, SR022 |
| CV001 | Lila was founded in Flagship Pioneering's labs in 2023. | 中 | SV002 |
| CV002 | Lila launched publicly in March 2025 with $200 million of committed seed capital. | 高 | SV002, SV011 |
| CV003 | Lila positions itself as a scientific superintelligence platform for life, chemical, and materials science. | 高 | SV001, SV002, SV003 |
| CV004 | Lila says its AI Science Factories combine AI, software, and robotics to run closed-loop experimentation. | 高 | SV004, SV005 |
| CV005 | Lila announced a $235 million Series A co-led by Braidwell and Collective Global. | 高 | SV004, SV008 |
| CV006 | Lila's October 2025 extension added $115 million and brought total Series A financing to $350 million. | 高 | SV005, SV006, SV007, SV008, SV009 |
| CV007 | Lila's total capital raised reached $550 million after the Series A extension. | 高 | SV005, SV006, SV007, SV011 |
| CV008 | Reuters and Goodwin said the Series A extension lifted Lila's valuation to more than $1.3 billion. | 高 | SV006, SV007, SV010, SV011 |
| CV009 | The Series A syndicate added NVentures, Analog Devices, IQT, and other strategic backers in addition to Flagship and earlier investors. | 高 | SV005, SV006, SV008 |
| CV010 | Lila says the new capital will scale AI Science Factories and open the platform to customers and partners. | 高 | SV005, SV006 |
| CV011 | Reuters reported that Lila does not plan to bring molecules into clinical trials itself and expects partners or startups to commercialize outputs. | 中 | SV007 |
| CV012 | Fierce Biotech reported that Lila had not yet publicly released data to support its technical claims. | 中 | SV008 |
| CV013 | Public sources reviewed do not disclose named paying customers, revenue, pricing, or gross margin for Lila. | 中 | SV005, SV007, SV008 |
| CV014 | Sacra independently tracked Lila at about a $1.30 billion valuation and $550 million of funding in 2025. | 中 | SV011 |
| CV015 | Flagship said its ecosystem has produced more than $60 billion of aggregate value across platform companies such as Moderna and Generate. | 中 | SV002 |
| CV016 | Xaira launched in 2024 with $1 billion of financing, showing that frontier AI-biotech companies can raise more capital than Lila before late-stage proof. | 高 | SV016, SV017 |
| CV017 | Xaira investors said biology is data poor and that building AI drug companies requires billions of dollars, underscoring sector capital intensity. | 中 | SV016 |
| CV018 | Isomorphic Labs raised $600 million in its first external round in 2025 led by Thrive with GV and Alphabet support. | 高 | SV012, SV013, SV014 |
| CV019 | Isomorphic Labs raised another $2.1 billion in 2026, showing the top end of private AI-science capital appetite. | 中 | SV015 |
| CV020 | Generate:Biomedicines raised $273 million of Series C funding in 2023 and said it had raised nearly $700 million in equity since 2020. | 高 | SV018, SV019, SV020 |
| CV021 | Generate disclosed 17 programs and at least one first-in-human trial, giving it more visible pipeline maturity than Lila. | 中 | SV018 |
| CV022 | Recursion's 2025 10-K says the company had no approved products for commercial sale and expects to need substantial additional funding. | 中 | SV021 |
| CV023 | CompaniesMarketCap puts Recursion's market capitalization at about $2.01 billion as of June 2026. | 中 | SV029 |
| CV024 | Exscientia's 2021 IPO priced 13.85 million ADS at $22 for $304.7 million and added $160 million of concurrent private placements. | 中 | SV028 |
| CV025 | Recursion and Exscientia agreed a 2024 all-stock merger valuing Exscientia at about $688 million. | 高 | SV024, SV025, SV026, SV027 |
| CV026 | The merger exchange ratio was 0.7729 Recursion shares per Exscientia share, leaving Exscientia holders with roughly 26% of the combined company. | 高 | SV023, SV027 |
| CV027 | BioPharma Dive said Recursion and Exscientia had each lost most of their value since going public by the time of the merger. | 中 | SV026 |
| CV028 | Drug Discovery Trends reported Exscientia's stock fell from $21.97 in October 2021 to $4.68 in August 2024. | 中 | SV027 |
| CV029 | CompaniesMarketCap recorded Exscientia at about a $0.63 billion market cap on January 22, 2025. | 中 | SV030 |
| CV030 | DrugPatentWatch concluded AI has improved preclinical success but not late-stage efficacy, which is the gap that matters most to investors. | 中 | SV031 |
| CV031 | All About AI said no AI-discovered drug had yet received FDA approval as of 2024 despite more than $60 billion of AI investment. | 中 | SV032 |
| CV032 | Lila's breadth across therapeutics, materials, and chemistry means pure-play AI drug discovery comparables are directionally useful but imperfect. | 中 | SV001, SV002, SV016, SV018 |
| CV033 | The strongest support for Lila's current mark is syndicate quality and platform optionality rather than public commercial proof. | 中 | SV006, SV007, SV008, SV015, SV016 |
| CV034 | A stage-appropriate method for Lila is probability-weighted milestone and comparable-round valuation rather than a revenue multiple because revenue is undisclosed. | 中 | SV007, SV011, SV016, SV018, SV021 |
| CV035 | Flagship incubation likely deserves a premium versus an ordinary Series A company, but that premium should shrink if proof stays non-public. | 中 | SV002, SV015, SV020, SV026, SV031 |
| CV036 | A bull case for Lila assumes named paid partners, reproducible technical data, and a next financing or strategic transaction at roughly $2.3 billion to $3.0 billion. | 低 | SV005, SV007, SV015, SV016, SV020 |
| CV037 | A base case for Lila assumes limited partner conversion and continued premium capital access, supporting roughly $1.1 billion to $1.6 billion. | 低 | SV007, SV008, SV011, SV020, SV023 |
| CV038 | A bear case for Lila assumes opaque proof, slower partner uptake, and sector de-rating, implying roughly $0.5 billion to $0.9 billion. | 低 | SV008, SV026, SV027, SV031, SV032 |
| CV039 | From a current mark above $1.3 billion, the bull case can work, but the base case offers little margin of safety and the bear case implies material capital loss. | 中 | SV007, SV015, SV026, SV031, SV032 |
| CV040 | The most material diligence gap is whether any partner has converted from interest into paid, repeatable programs with measurable output. | 中 | SV007, SV008, SV010 |
| CV041 | The next-most material diligence gap is lab productivity economics, including throughput, cost per experiment, and hit-to-validation rate. | 中 | SV004, SV005, SV017 |
| CV042 | Recommendation: track the company, but do not underwrite the current mark as attractive until proof or price changes. | 中 | SV007, SV008, SV026, SV031, SV032 |
| CV043 | Confidence is medium because financing and investor quality are clear, but commercial and technical evidence remains sparse. | 中 | SV005, SV007, SV008, SV011 |
| CV044 | Risk rating is high because Lila is capital intensive, pre-commercial in public evidence, and exposed to sector re-rating. | 中 | SV008, SV016, SV021, SV026, SV031 |
| CV045 | Valuation stance is stretched rather than irrational because Lila sits above ordinary Series A pricing but below the most aggressive AI-science private capital pools. | 中 | SV007, SV015, SV016, SV020, SV023 |
| CV046 | The view would improve with named paid partners, public validation datasets, and cleaner cap-table visibility. | 中 | SV005, SV007, SV008 |
| CV047 | The view would worsen if 12 to 18 months pass with no customer disclosures or if sector de-rating deepens further. | 中 | SV008, SV026, SV030, SV031 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Lila Sciences | LILA | Scientific Superintelligence | LILA's advanced AI model is the brain. Our proprietary AI Science Factory™ instruments are the body. |
| SO002 | Lila Sciences | About | LILA | The World's First Operating System for Science | Scale is the key to accelerating the scientific method. |
| SO003 | Lila Sciences | Team | LILA | Scientific Superintelligence | |
| SO004 | Lila Sciences | Geoffrey von Maltzahn, PhD | Lila | Geoffrey von Maltzahn is Co-founder and CEO of Lila Sciences, where he is leading the company’s mission to build scientific superintelligence. |
| SO005 | Lila Sciences | Join Our Mission | Lila | We’ve been building behind the scenes for about three years within the labs of Flagship Pioneering... We are honored to announce $200 million in seed capital. |
| SO006 | Lila Sciences | Announcing Lila’s $350M Series A and Incredible Partners on Our Mission | Today we’re announcing the close of our $350M Series A, bringing Lila’s total capital raised to $550M. |
| SO007 | Lila Sciences | Careers | LILA | Scientists and engineers, technologists and experimentalists work side by side to turn questions into ideas, and ideas into breakthroughs. |
| SO008 | Lila Sciences | Jawad Ahsan | Lila | Jawad Ahsan is Chief Operating Officer and Chief Financial Officer at Lila Sciences. |
| SO009 | Lila Sciences | Andrew Beam, PhD | Lila | Andrew Beam is Chief Technology Officer at Lila Sciences, where he leads development of AI for scientific discovery. |
| SO010 | Lila Sciences | Chris Fussell | Lila | Chris Fussell is President of Business Operations at Lila Sciences. |
| SO011 | Lila Sciences | Rafael Gómez-Bombarelli, PhD | Lila | Rafael Gómez-Bombarelli, PhD, is a Co-founder and Chief Scientific Officer of Physical Sciences at Lila Sciences. |
| SO012 | Lila Sciences | Julie Shah, PhD | Lila | Julie Shah is Chief Robotics Officer at Lila Sciences. |
| SO013 | Flagship Pioneering | Lila Sciences | Flagship Pioneering | Lila is growing its team in Cambridge, San Francisco, and London. |
| SO014 | Flagship Pioneering | Geoffrey von Maltzahn | Flagship Pioneering | Through his role in Flagship Labs... Geoffrey has created companies that include Lila Sciences, Quotient Therapeutics, Mirai Bio, Tessera Therapeutics, Generate:Biomedicines, Indigo Agriculture, Sana Biotechnology, and Seres Therapeutics. |
| SO015 | PR Newswire | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence in Science | Company has raised $200M in seed financing to further develop platform and build first AI Science Factories. |
| SO016 | Reuters | AI startup Lila Sciences raises extension round and tops $1.3B valuation | The latest funding brings Lila's total Series A to $350 million and overall capital raised to $550 million. |
| SO017 | Yahoo Finance / Reuters | Exclusive: AI lab Lila Sciences tops $1.3 billion valuation with new Nvidia backing | AI startup Lila Sciences has raised $115 million in an extension funding round from investors including Nvidia's venture arm, lifting its valuation to more than $1.3 billion. |
| SO018 | Fierce Biotech | Flagship’s Lila adds $115M to series A, bringing total haul to $350M and securing Nvidia backing | The company has not yet publicly released any data to support the claims. |
| SO019 | Goodwin | Goodwin Advises Lila on $350 Million Series A | The Technology and Life Science teams advised Lila Sciences on its $350 million Series A financing... lifting its valuation to more than $1.3 billion. |
| SO020 | Built In Boston | Lila Sciences Raises $350M Series A to Expand Its Reach | Massachusetts-based Lila Sciences closed a Series A funding round worth $350 million. |
| SO021 | CNBC | 25. Lila Sciences | As with all things AI, there are questions around whether the hype surrounding Lila is running ahead of reality. |
| SO022 | Bisnow | AI Biotech Startup Signs 235K SF Alewife Lease: The Boston Deal Sheet | AI startup Lila Sciences leased 235K SF at 1 and 5 Alewife Park in Cambridge from IQHQ. |
| SO023 | The Economic Times | AI lab Lila Sciences tops $1.3 billion valuation with new Nvidia backing | The latest funding brings Lila’s total Series A to $350 million and overall capital raised to $550 million. |
| SO024 | AGBI | AI lab Lila Sciences tops $1bn valuation with Nvidia backing | Lila said the funds will accelerate development of its 'AI Science Factories'. |
| SO025 | StartupWired | Lila Sciences Hits $1.3B with Nvidia’s AI Lab Backing | The company recently signed a 235,500-square-foot lease in Cambridge, Massachusetts—one of the largest lab leases in the Greater Boston area this year. |
| SO026 | CafePharma | Lila Sciences raises $235M Series A, reaches unicorn status with ambitious AI-science platform | Lila is entering a crowded field... Ensuring safety, reproducibility, and oversight when experiments are largely automated will be important. |
| SO027 | Robotics & Automation News | Lila Sciences raises $235 million in Series A funding to advance AI-driven scientific research | The round also included participation from Altitude Life Science Ventures, Alumni Ventures, ARK Venture Fund, Common Metal, Flagship Pioneering, General Catalyst, March Capital, the Mathers Foundation, Modi Ventures, NGS Super, the State of Michigan Retirement System, and a wholly owned subsidiary of the Abu Dhabi Investment Authority (ADIA). |
| SM001 | MarketsandMarkets | Lab Automation Market Report 2026-2031, By Product, Application, and Geo | The global lab automation market is projected to grow from USD 6.60 billion in 2026 to USD 8.62 billion by 2031, at a CAGR of 6.6% during the forecast period. |
| SM002 | Business Research Insights | Lab Automation Market Size, Share | Global Research [2035] | Global Lab Automation Market size is valued at USD 12.12 Billion in 2026, expected to reach USD 25.2 Billion by 2035. |
| SM003 | Precedence Research | Lab Automation Market Size to Surpass USD 14.78 Bn By 2034 | The global lab automation market size is predicted to increase from USD 8.91 billion in 2026 to approximately USD 14.78 billion by 2034. |
| SM004 | Future Market Insights | Lab Automation Market | Global Market Analysis Report - 2036 | The lab automation market is expected to expand from USD 2.7 billion in 2026 to USD 6.9 billion by 2036. |
| SM005 | Research and Markets | Lab Automation Market Report 2026 - Research and Markets | |
| SM006 | Mordor Intelligence | Laboratory Informatics Market Size, Share & Growth | Forecast Report - 2031 | The Laboratory Informatics Market size is projected to be USD 4.05 billion in 2026 and reach USD 6.08 billion by 2031. |
| SM007 | Business Research Insights | Laboratory Informatics Market Segmentation & Forecast 2026–2035 | The global Laboratory Informatics Market is anticipated to be worth USD 5.4 Billion in 2026. |
| SM008 | Grand View Research | Laboratory Informatics Market Size | Industry Report, 2033 | Market Size, 2025 (US$B) $4.1B; Forecast, 2033 (US$B) $6.0B; CAGR, 2026 - 2033 4.9%. |
| SM009 | Global Market Insights | Artificial Intelligence in Drug Discovery Market Size, Share – 2035 | AI in drug discovery market size exceeded USD 3.1 billion in 2025 and is expected to grow at a CAGR of 30.5% from 2026 to 2035. |
| SM010 | Mordor Intelligence | AI in Drug Discovery Market Size, Growth & Drivers Research Report 2031 | The Artificial Intelligence In Drug Discovery Market size is estimated to grow from USD 3.25 billion in 2026 to reach USD 10.29 billion by 2031. |
| SM011 | Research and Markets | Artificial Intelligence in Drug Discovery Market - Global Forecast 2026-2032 | |
| SM012 | National Center for Science and Engineering Statistics | Federal R&D Funding, by Budget Function 2024-2026 | The data for FY 2026 are the funding levels proposed by the president’s Budget of the United States Government, Fiscal Year 2026. |
| SM013 | IQVIA Institute | Global R&D Trends 2026 | Biopharmaceutical R&D remained resilient in 2025, with investment and dealmaking increasingly concentrated in high value science. |
| SM014 | Royal Society Open Science | Autonomous self-driving laboratories: a review of technology and ... | |
| SM015 | ACS Omega | Self-Driving Laboratories: Translating Materials Science from Laboratory to Factory | We argue that self-driving laboratories represent not merely another step in automation, but a fundamental reimagining of the materials development pipeline. |
| SM016 | Materials Horizons | Toward self-driving laboratory 2.0 for chemistry and materials discovery | While early SDLs demonstrated the feasibility of closed-loop discovery, their impact has been constrained by limited scope, poor interoperability, and reliance on human-curated heuristics. |
| SM017 | Agilent Technologies | Agilent Technologies, Inc. - Investor Overview | Agilent Technologies Inc. is a global leader in the life sciences, diagnostics, and applied markets. |
| SM018 | Bruker | Chemspeed and SciY Announce Self‑Driving Laboratory Platform Integrating Automation, Analytics and AI Orchestration | Today, many labs face significant challenges from siloed tools and integration gaps in heterogeneous lab environments that limit efficiency and scalability. |
| SM019 | STAT | AI & drug discovery: A biotech CEO, a scientist, and a venture capitalist separate hype from reality | “I am very worried about the hype,” said Daphne Koller. |
| SM020 | Thermo Fisher Scientific / SEC | Thermo Fisher Scientific 2024 Annual Report | Pharma & Biotech 57%; Academic & Government 15%; Industrial & Applied 14%; Diagnostics & Healthcare 14%. |
| SM021 | Congressional Research Service | Federal Research and Development (R&D) Funding: FY2026 | CRS calculated that President Trump’s budget proposal for FY2026 included approximately $181.4 billion for R&D. |
| SM022 | AAAS | FY 2026 R&D Appropriations Dashboard | |
| SM023 | National Institutes of Health | Budget | The NIH invests most of its nearly $48 billion budget in medical research for the American people. |
| SM024 | Deloitte | 2026 Life Sciences Outlook | |
| SM025 | Research and Markets | Laboratory Informatics Market Report 2026 - Research and Markets | |
| SP001 | LILA | LILA | Scientific Superintelligence | LILA's operating system for science executes the entire scientific method autonomously — generating hypotheses, designing experiments, running them, and learning from results in real time. |
| SP002 | LILA | About | LILA | The World's First Operating System for Science | We are focused on creating a single, general platform for autonomous science, rather than many narrow, domain-specific tools. |
| SP003 | Flagship Pioneering | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence in Science | Lila Sciences, a company building the world's first scientific superintelligence platform and fully autonomous labs for life, chemical, and materials sciences. |
| SP004 | Recursion | Pioneering AI Drug Discovery | Recursion | Over the last decade, we have generated and aggregated one of the largest fit-for-purpose proprietary biological and chemical datasets in the world — >50 petabytes... Our automated wet lab utilizes robotics and computer vision to capture millions of cell experiments per week. |
| SP005 | Recursion | Technology | Central to our mission is the Recursion Operating System (OS), a platform powered by one of the world’s largest proprietary biological and chemical datasets. |
| SP006 | Recursion | Recursion to Acquire Exscientia, Combining AI Drug Pioneers | |
| SP007 | Securities and Exchange Commission | Exscientia plc Form 6-K: Transaction Agreement with Recursion | |
| SP008 | BioSpace | Recursion and Exscientia Enter Definitive Agreement to Create a Global Technology-Enabled Drug Discovery Leader with End-to-End Capabilities | |
| SP009 | pharmaphorum | AI biotechs Exscientia and Recursion agree $688m merger | Recursion will absorb its smaller UK counterpart... [to] create a 'full-stack technology-enabled small molecule discovery platform' powered by AI and with 10 programmes in clinical testing. |
| SP010 | Insilico Medicine | Pharma.ai | Insilico Medicine is using AI to create an entirely new AI-driven drug discovery pipeline from A to Z. |
| SP011 | Insilico Medicine | About Insilico | The company has received strong external validation... with collaborations with leading industry partners around the globe, including 10 of the top 20 global pharmaceutical companies in terms of 2021 reported sales. |
| SP012 | Isomorphic Labs | Reimagining Drug Discovery Process with AI - Isomorphic Labs | Our interdisciplinary team ... has built powerful new predictive and generative AI models that accelerate scientific discovery at digital speed. |
| SP013 | Isomorphic Labs | Partnerships - Isomorphic Labs | The initial scope of our research collaboration was focused on the discovery of small molecule therapeutics against three particularly challenging targets. That has now been expanded - adding up to three additional research programs. |
| SP014 | PR Newswire | ISOMORPHIC LABS ANNOUNCES STRATEGIC MULTI-TARGET RESEARCH COLLABORATION WITH LILLY | Isomorphic Labs will partner with Lilly to discover small molecule therapeutics against multiple targets and will receive an upfront cash payment of $45 million. |
| SP015 | Isomorphic Labs | News - Isomorphic Labs | Isomorphic Labs announces $600m external investment round. |
| SP016 | Benchling | Cloud-based platform for biotech R&D | Benchling | Digitize your lab, automate workflows, and increase productivity with AI. |
| SP017 | Benchling | Benchling Solutions | Benchling Solutions contemplate the full end-to-end R&D process, including core capabilities such as experimental tracking, sample management, inventory, and process management. |
| SP018 | Benchling | Benchling | Customers in Life Sciences R&D | Trusted by 1,200+ leading biotech organizations. |
| SP019 | Arcadia Science | About | Arcadia Science | Arcadia was founded in 2021 with a long time horizon to rethink the entire research cycle. |
| SP020 | Arcadia Science | Arcadia Science | As we develop our platform, we release apps, software pipelines, protocols, and other resources to the scientific community. |
| SP021 | GitHub | OpenBioML | OpenBioML/datasets’s past year of commit activity. |
| SP022 | TechCrunch | Stability AI backs effort to bring machine learning to biomed | TechCrunch | The company’s founders describe OpenBioML as an 'open research laboratory'. |
| SP023 | Opentrons | Opentrons Labworks Inc | Use the assays, instruments, and AI tools you want, now and later, without being forced into a closed system. |
| SP024 | Opentrons | Opentrons Labworks Inc | Reconfigure hardware, workflows, and throughput as your science evolves and the needs of your lab change, without starting over. |
| SP025 | Drug Discovery Trends | SLAS 2026: Orchestration patforms, API-first instruments and the rise of semiautonomous labs | The lab OS wars: 15 companies vying to enable AI-enabled labs at SLAS 2026. |
| SP026 | Royal Society Open Science | Autonomous ‘self-driving’ laboratories: a review of technology and policy implications | Level-5 SDL ... full automation of the scientific method ... has not yet been realized. |
| SP027 | University of Chicago | ‘AI advisor’ helps scientists steer autonomous labs | We promote human-machine collaboration to boost discovery together. |
| SP028 | Northwestern University | Megalibraries in pole position for autonomous discovery over self-driving labs | Compared to the megalibrary ... self-driving labs are basically crawling. |
| SP029 | Nasdaq | Recursion and Exscientia Shareholders Approve the Proposed Combination | |
| SP030 | Genentech | Redefining Drug Discovery with AI | The foundation of our strategy centers on creating a 'lab in a loop,' where data from the lab and clinic feed AI models ... and generate new molecules. |
| SI001 | Lila Sciences | LILA | Scientific Superintelligence | |
| SI002 | Lila Sciences | Welcoming New Partners in Our Mission to Build Scientific Superintelligence | Today I’m thrilled to share a milestone for Lila Sciences: a $235M Series A, co-led by Braidwell and Collective Global. |
| SI003 | Lila Sciences | Announcing Lila’s $350M Series A and Incredible Partners on Our Mission | Today we’re announcing the close of our $350M Series A, bringing Lila’s total capital raised to $550M. |
| SI005 | Lila Sciences | Careers | LILA | |
| SI006 | Flagship Pioneering | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence in Science | Company has raised $200M in seed financing to further develop platform and build first AI Science Factories. |
| SI007 | Flagship Pioneering | Flagship Pioneering and AWS Announce Collaboration to Accelerate Drug Discovery and Life Sciences Innovation | |
| SI008 | Altitude Life Science Ventures | Announcing Lila’s $350M Series A and Incredible Partners on Our Mission | |
| SI009 | Braidwell | Braidwell | |
| SI010 | Collective Global | collectiveglobal.com | |
| SI011 | NVIDIA Newsroom | News Archive | |
| SI012 | General Catalyst Job Board | Director, Facilities @ Lila Sciences | |
| SI013 | General Catalyst Job Board | Facilities Support Specialist (Contractor) @ Lila Sciences | |
| SI014 | Reuters via Yahoo Finance | Exclusive: AI lab Lila Sciences tops $1.3 billion valuation with new Nvidia backing | The latest funding brings Lila's total Series A to $350 million and overall capital raised to $550 million. |
| SI015 | Fierce Biotech | Lila Sciences adds Nvidia-backed $115M to series A, bringing total haul to $350M | The company has not yet publicly released any data to support the claims. |
| SI016 | Bloomberg | AI Unicorn: Lila Sciences Raises $235 Million in Latest Round | The startup announced it had raised $235 million at a roughly $1.23 billion valuation. |
| SI017 | Biotech Industry Examiner | The AI science factory arrives: why Lila’s $1.3bn valuation matters beyond biotech | Factories are capital-hungry and unforgiving. |
| SI019 | Sacra | Lila Sciences valuation, funding & news | |
| SI020 | Nasdaq Private Market | Sell or Invest in Lila Sciences Stock Pre-IPO | |
| SI021 | Forge | Lila Sciences IPO: Investment Opportunities & Pre-IPO Valuations | |
| SI023 | Built In | Lila Sciences Jobs + Careers | |
| SI024 | Securities and Exchange Commission | SEC FORM D for AVSF - Lila Sciences 2025, LLC | Name of Issuer: AVSF - Lila Sciences 2025, LLC. |
| SI025 | North American Securities Administrators Association EFD | View Form D - Electronic Filing Depository | Offering Amount: $817,500. |
| SI026 | Greenhouse | Lila Sciences | |
| SI027 | Built In | Lila Sciences Careers, Perks + Culture | |
| SI028 | WebProNews | Lila Sciences Secures $235M Funding, Hits Unicorn Status in AI Science | |
| SE001 | Lila Sciences | Tech | LILA | |
| SE002 | Lila Sciences | Solutions | |
| SE003 | Lila Sciences | LILA Catalyst | LILA Iris | AI Science Factories | |
| SE004 | Lila Sciences | Lila Creation | Lila Iris | AI Science Factories | |
| SE005 | Lila Sciences | About | LILA | The World's First Operating System for Science | |
| SE006 | Lila Sciences | Therapeutics | LILA | |
| SE007 | Lila Sciences | Biotech | LILA | |
| SE008 | Lila Sciences | Chemicals | LILA | |
| SE009 | Lila Sciences | Advanced Materials | LILA | |
| SE010 | Lila Sciences | Energy and Environment | LILA | |
| SE011 | Lila Sciences | Careers | LILA | |
| SE012 | Lila Sciences | Julie Shah, PhD | Lila | |
| SE013 | Lila Sciences | Milad Abolhasani, PhD | Lila | |
| SE014 | Lila Sciences | Rafael Gómez-Bombarelli, PhD | Lila | |
| SE015 | Lila Sciences | Kenneth Stanley, PhD | Lila | |
| SE016 | Lila Sciences | Announcing Lila’s $350M Series A and Incredible Partners on Our Mission | |
| SE017 | Lila Sciences | Privacy Policy | LILA | |
| SE018 | Lila Sciences | Candidate Privacy Policy Notice | |
| SE019 | Flagship Pioneering | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence… | |
| SE020 | Flagship Pioneering | Lila Sciences | |
| SE021 | PR Newswire | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence in Science | |
| SE022 | Greenhouse / Lila Sciences | Lila Sciences | |
| SE023 | CareersInRobotics | Lila Sciences Careers | 7 jobs | CareersInRobotics | |
| SE024 | Biotech Industry Examiner | The AI science factory arrives: why Lila’s $1.3bn valuation matters beyond biotech - Biotech Industry Examiner | |
| SE025 | Excedr | Lila Sciences Builds Scientific Superintelligence Through Autonomous AI Labs | |
| SE026 | MIT Department of Materials Science and Engineering | MIT Technology Review: AI-driven labs aim to accelerate materials discovery - MIT Department of Materials Science and Engineering | |
| SE027 | BioPharmaTrend | Lila Sciences Raises $235M to Build Autonomous AI Labs, Joins Unicorn Ranks | |
| SE028 | Nature Synthesis | The rise of self-driving labs in chemical and materials sciences | |
| SE029 | MIT Department of Mechanical Engineering | MECHE PEOPLE: jshah@mit.edu | MIT Department of Mechanical Engineering | |
| SU001 | Lila Sciences | LILA | Scientific Superintelligence | |
| SU002 | Lila Sciences | About | LILA | The World's First Operating System for Science | |
| SU003 | Lila Sciences | Solutions | Access to LILA's AI Science Factories works the way modern infrastructure should — on demand, at the scale your program requires, without the capital commitment of building it yourself. |
| SU004 | Lila Sciences | Tech | LILA | |
| SU005 | Lila Sciences | Team | LILA | Scientific Superintelligence | |
| SU006 | Lila Sciences | Therapeutics | LILA | |
| SU007 | Lila Sciences | Biotech | LILA | |
| SU008 | Lila Sciences | Chemicals | LILA | |
| SU009 | Lila Sciences | Advanced Materials | LILA | |
| SU010 | Lila Sciences | Energy and Environment | LILA | |
| SU011 | Lila Sciences | LILA Catalyst | LILA Iris | AI Science Factories | Partners gain access to Lila Iris™, our proprietary AI platform powered by Scientific Superintelligence™. By tapping into LILA's Lab-as-a-Service (LaaS™), teams convert fixed lab capacity and capex into an on-demand resource. |
| SU012 | Lila Sciences | Lila Creation | Lila Iris | AI Science Factories | Investors or strategic partners present a problem space or thesis; Lila runs focused Creation campaigns to discover novel molecules, materials, or platforms with clear technical and commercial differentiation. |
| SU013 | Lila Sciences | AI is not going to solve all the problems in the energy sector. But it might fix this one. | As a commercial product, Lila’s system operates on top of a company's existing data and platforms, so using it requires no IT transformation or grand digitization project. |
| SU014 | Lila Sciences | Scientific Superintelligence: The Deep Blue Moment | |
| SU015 | Flagship Pioneering | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence in Science | The Lila platform will be open to partners across the life and material sciences industries to jointly bring forth solutions in human health and sustainability. |
| SU016 | PR Newswire | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence in Science | |
| SU017 | Fierce Biotech | With $200M in seed funding, Flagship-backed Lila Sciences touts ambitious AI vision | |
| SU018 | BioPharma Dive | Flagship startup raises $200M in pursuit of scientific superintelligence | Lila will not make its own therapeutic candidates. Instead, the company will partner with other Flagship startups and outside biotech companies to help them speed their research. |
| SU019 | pharmaphorum | Scientific superintelligence firm Lila launches with $200m | |
| SU020 | Fierce Biotech | Lila Sciences adds Nvidia-backed $115M to series A, bringing total haul to $350M | These are “superhuman scientific performance;” building more automated labs, which Lila calls AI science factories; bringing in the company's first customers; and hiring “the world's most brilliant minds,” the CEO said. |
| SU021 | U.S. News & World Report / Reuters | Exclusive-AI Lab Lila Sciences Tops $1.3 Billion Valuation With New Nvidia Backing | It also plans to open its platform to commercial customers, offering access to its AI models and automated labs via enterprise software. Lila said the platform has drawn interest from firms in energy, semiconductors and drug development, although it did not name specific companies. |
| SU022 | March Capital | Lila: Building Scientific Superintelligence | We have partnered with Geoffrey von Maltzahn since 2021 through ventures including Generate Biomedicines and Tessera Therapeutics. |
| SU023 | Biotech Industry Examiner | The AI science factory arrives: why Lila’s $1.3bn valuation matters beyond biotech | The near-term commercial test is practical: can Lila define units of work that feel productised to a procurement team? |
| SU024 | Tech Startups | Lila Sciences hits $1.3B valuation after $115M raise from Nvidia to build AI Science Factories | |
| SU025 | P05.org | Company of the Week: Lila Sciences – A Red and Blue Team Analysis | |
| SR001 | Lila Sciences | LILA | Scientific Superintelligence | |
| SR002 | Lila Sciences | Announcing Lila’s $350M Series A and Incredible Partners on Our Mission | |
| SR003 | Lila Sciences | Advanced Materials | LILA | |
| SR004 | Lila Sciences | Careers | LILA | |
| SR005 | Lila Sciences | Privacy Policy | LILA | |
| SR006 | Lila Sciences | Terms of Use | LILA | |
| SR007 | Flagship Pioneering | Lila Sciences | |
| SR008 | PR Newswire / Flagship Pioneering | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence in Science | |
| SR009 | CNBC | 25. Lila Sciences | |
| SR010 | Fierce Biotech | Lila Sciences adds Nvidia-backed $115M to series A, bringing total haul to $350M | |
| SR011 | Greenhouse | Lila Sciences | |
| SR012 | Recursion | Pioneering AI Drug Discovery | Recursion | |
| SR013 | Isomorphic Labs | Reimagining Drug Discovery Process with AI - Isomorphic Labs | |
| SR014 | cusp.ai | cusp.ai | |
| SR015 | Insilico Medicine | Main | Insilico Medicine | |
| SR016 | Absci | Home | Absci | |
| SR017 | NIST | AI Risk Management Framework | |
| SR018 | National Academies of Sciences, Engineering, and Medicine | Reproducibility and Replicability in Science | |
| SR019 | FDA | The FDA's Drug Review Process: Ensuring Drugs Are Safe and Effective | |
| SR020 | National Center for Biotechnology Information | Pharma Success in Product Development—Does Biotechnology Change the Paradigm in Product Development and Attrition | |
| SR021 | Wyss Institute at Harvard University | From Data to Drugs: The Role of Artificial Intelligence in Drug Discovery | |
| SR022 | UCSF Quantitative Biosciences Institute | QBI - Drug Discovery | |
| SR023 | NIH Office of Science Policy | Biosafety and Biosecurity Policy | |
| SR024 | U.S. Department of Health & Human Services | HIPAA Home | |
| SR025 | European Data Protection Supervisor | Artificial Intelligence | |
| SR026 | RAND | Biosecurity Governance Across Uncertain Artificial Intelligence Futures | |
| SR027 | Johns Hopkins Center for Health Security | Risk-Based Categorization and Governance of Biological Data in AI Systems | |
| SR028 | FDA | Artificial Intelligence in Software | |
| SR029 | OECD | The OECD Artificial Intelligence Policy Observatory | |
| SR030 | Lila Sciences | Lila Wants to Create "Scientific Superintelligence" | |
| SV001 | Lila Sciences | About | LILA | The World's First Operating System for Science | |
| SV002 | Flagship Pioneering | Flagship Pioneering Unveils Lila Sciences to Build Superintelligence… | Company has raised $200M in seed financing to further develop platform and build first AI Science Factories. |
| SV003 | Flagship Pioneering | Lila Sciences | |
| SV004 | Lila Sciences | Welcoming New Partners in Our Mission to Build Scientific Superintelligence | Today I’m thrilled to share a milestone for Lila Sciences: a $235M Series A, co-led by Braidwell and Collective Global. |
| SV005 | Lila Sciences | Announcing Lila’s $350M Series A and Incredible Partners on Our Mission | Today we’re announcing the close of our $350M Series A, bringing Lila’s total capital raised to $550M. |
| SV006 | Goodwin | Goodwin Advises Lila on $350 Million Series A | News & Events | Goodwin | The Technology and Life Science teams advised Lila Sciences on its $350 million Series A financing, bringing the company’s total capital raised to $550 million and lifting its valuation to more than $1.3 billion. |
| SV007 | Reuters via Yahoo Finance | Exclusive-AI lab Lila Sciences tops $1.3 billion valuation with new Nvidia backing | AI startup Lila Sciences has raised $115 million in an extension funding round ... lifting its valuation to more than $1.3 billion. |
| SV008 | Fierce Biotech | Lila Sciences adds Nvidia-backed $115M to series A, bringing total haul to $350M | The company has not yet publicly released any data to support the claims. |
| SV009 | Built In Boston | Lila Sciences Raises $350M Series A to Expand Its Reach | Built In Boston | |
| SV010 | The Economic Times | AI lab Lila Sciences tops $1.3 billion valuation with new Nvidia backing - The Economic Times | |
| SV011 | Sacra | Lila Sciences valuation, funding & news | Valuation $1.30B ... Funding $550.00M. |
| SV012 | Isomorphic Labs | Isomorphic Labs announces $600m external investment round - Isomorphic Labs | Isomorphic Labs announces it has raised $600 Million in its first external funding round. |
| SV013 | PR Newswire | Isomorphic Labs announces $600 million funding to further develop its next-generation AI drug design engine and advance therapeutic programs into the clinic | |
| SV014 | TechCrunch | Alphabet's AI drug discovery platform Isomorphic Labs raises $600M from Thrive | TechCrunch | |
| SV015 | Isomorphic Labs | Isomorphic Labs announces Series B investment round - Isomorphic Labs | Isomorphic Labs announces it has raised $2.1 Billion in Series B funding. |
| SV016 | TechCrunch | Xaira, an AI drug discovery startup, launches with a massive $1B, says it's 'ready' to start developing drugs | TechCrunch | ARCH Venture Partners and Foresite Labs ... funded the AI biotech with $1 billion. |
| SV017 | pharmaphorum | Enter Xaira, with $1bn for its AI in drug discovery platform | |
| SV018 | Generate:Biomedicines via Business Wire | Generate:Biomedicines Announces Close of $273M Series C Financing to Advance Its Generative AI Pipeline of Preclinical and Clinical Protein Therapeutics | Generate:Biomedicines ... has raised $273 million in Series C financing. ... Company has raised nearly $700 million in equity financing since 2020. |
| SV019 | BioPharma Dive | Flagship-backed Generate raises $273M as its first drugs move to the clinic | |
| SV020 | Goodwin | Generate:Biomedicines Completes $273 Million Series C | News & Events | Goodwin | |
| SV021 | Securities and Exchange Commission | rxrx-20251231 | We are a clinical-stage biotechnology company with a limited operating history and no products approved by regulators for commercial sale. |
| SV022 | Securities and Exchange Commission | Document | |
| SV023 | Securities and Exchange Commission | Document | Exscientia shareholders received 0.7729 shares ... of Recursion Class A common stock for each Exscientia ordinary share. |
| SV024 | Fierce Biotech | After a tough year, Exscientia folds into Recursion to create an AI superpower | |
| SV025 | pharmaphorum | AI biotechs Exscientia and Recursion agree $688m merger | Recursion Pharma has agreed to join with Exscientia in an all-stock transaction valued at $688 million. |
| SV026 | BioPharma Dive | Recursion to absorb Exscientia in ‘techbio’ deal | The two AI drug discovery firms, which have each lost most of their value since going public ... |
| SV027 | Drug Discovery & Development | Recursion-Exscientia merger consolidates AI in drug discovery field | Exscientia’s stock price has fallen from a high of $21.97 in October 2021 to $4.68 in August 2024. |
| SV028 | Exscientia via Business Wire | Exscientia Announces Pricing of $304.7 Million Upsized Initial Public Offering and $160.0 Million Concurrent Private Placements | |
| SV029 | CompaniesMarketCap | Recursion Pharmaceuticals (RXRX) - Market capitalization | As of June 2026 Recursion Pharmaceuticals has a market cap of $2.01 Billion USD. |
| SV030 | CompaniesMarketCap | Exscientia (EXAI) - Market capitalization | On January 22, 2025 Exscientia had a market cap of $0.63 Billion USD. |
| SV031 | DrugPatentWatch | AI Drug Discovery’s $110B Productivity Bet: What the Clinical Data Actually Shows | AI has demonstrably improved preclinical success rates. It has not yet cracked late-stage efficacy. The gap between those two statements contains most of what matters for investors. |
| SV032 | All About AI | AI in Drug Development Statistics 2026: The $60 Billion Reality vs. Hype Analysis | Despite $60+ billion in global AI investments ... no AI-discovered drug has yet received FDA approval as of 2024. |