BioMap
战略型 AI 生物技术平台,但仍缺公开估值锚
一个战略上很有吸引力的 AI-biotech 平台,已有真实伙伴和客户证明,但估值故事披露仍不足。
封面要素
公司概况
BioMap 是一家中国 AI 生物技术公司,2020 年成立;公开资料显示 Robin Li 任创始人兼董事长,Wei Liu 任联合创始人兼 CEO。公司把 BioMap OS 和 xTrimo 模型家族定位成干湿闭环发现操作系统,覆盖蛋白生物学、生物药、精准医学、合成生物学和前沿生命科学研究。公开证据显示,BioMap 获得 HKIC 战略支持,拿到 Sanofi 旗舰合作,机构触达面广,并已保密递表赴港 IPO,但核心财务披露仍然有限。
- 成立时间
- 2020-09-25
- 创始人
- Robin Li, Wei Liu
- 创立地点
- Beijing, China
- 总部
- Beijing, China, with Hong Kong expansion
- 产品
- xTrimo 基础模型加 BioMap OS,构成一套干湿闭环发现操作系统,贯穿数据洞察、参数优化、de novo 设计、高通量验证和模型迭代改进。
- 客户
- 跨国药企、中国生物医药公司、CDMO、科研机构、合成生物学团队、绿色科技企业,以及与仪器联动的生命科学工作流。
- 商业模式
- 混合 B2B 变现:平台合同、发现项目、合作经济、湿实验工作流支持,以及里程碑、特许权使用费等潜在资产增值结构。
- 阶段
- Late-stage private, pre-IPO
- 融资情况
- 公开披露历史包括 $100M Series A 轮、截至 2026 年 3 月媒体报道的累计融资超过 $200M、HKIC 战略支持,以及据报数亿美元规模的港交所 IPO 募资目标。
执行摘要
主要优势
- 少见地同时具备 foundation-model 深度、干湿实验流程整合和企业生物学定位。
- HKIC、Sanofi、Harbour/MegaStream 给出战略验证,而不只是 VC 叙事。
- 公开披露 200+ 个签约 AIGP 用户和 800+ 个机构用户,机构触达面较广。
- 围绕 xTrimoPGLM、PFMBench 及 BioMap 相关研究资产的公开技术足迹较强。
主要风险
- 没有公开经审计收入、毛利率、留存、burn 或 runway,无法干净锚定估值。
- 具名证明集中在少数头部伙伴,可能带来集中度风险。
- 运营模式资本密集且复杂,横跨模型、实验室、数据和伙伴流程。
- 保密 IPO 时点和估值叙事可能走在当前公开证明集前面。
未决问题
- 没有可对账的收入分母,也没有按平台、服务和里程碑拆分的收入结构。
- 没有公开 top-customer 集中度、续约或流失指标。
- 没有公开优先权堆栈、稀释路径或完整 cap table 可见度。
- 没有覆盖可靠性、实施成功率或支持质量的公开运营 dashboard。
目录
01公司概览
1.1 身份、平台与运营足迹
BioMap 对外呈现的是生命科学 AI 基础设施公司,而不是单一资产药物开发商。公司自有材料称 BioMap 是生命科学 AI 基础模型的全球先行者,并把 BioMap OS 定位成干湿闭环发现系统,把数据洞察、预测设计、实验控制和模型训练串起来。同一批官方材料列出北京、苏州、香港和硅谷四地布局;外部画像则称公司总部在北京、属于中国公司。一个重要不一致仍未解决:HKSTP 2024 年 6 月称 BioMap “总部在美国”,但 Tracxn 和中文报道都把它放在北京。最稳妥的结论是,BioMap 确实在运营上跨境,根基在中国,并向香港 / 美国扩张;但公开来源没有把法律顶层公司与运营总部的问题交代清楚。[CO001, CO004, CO005, CO006, CO007, CO008]
| 指标 | 数值 / 状态 | 截至 | 置信度 | 缺口或说明 |
|---|---|---|---|---|
| 成立 | 2020-09-25 | 2020 | 高 | 官方成立日期已由 BaiduWiki 和 Tracxn 交叉印证 |
| 运营版图 | 北京、苏州、香港、硅谷 | 2026 | 中 | 公开列有办公室;法律顶层控股公司 / 总部口径不一致 |
| 首次外部融资 | $100M Series A 轮 | 2021-07-30 | 高 | 轮次规模被广泛报道;后续轮次未公开拆分 |
| 战略药企合作 | Sanofi 交易:$10M 首付款,>$1B 里程碑 | 2023-10 | 高 | 交易经济条款已公开;实际里程碑兑现时间未披露 |
| 香港战略支持 | HKIC 投资 + 合作 | 2024-06-24 | 高 | 主权资本支持已宣布,具体出资金额未披露 |
| AIGP 签约用户 | 200+ | 2024-06-24 | 中 | HKIC 数字早于后续全球机构口径 |
| 机构用户 | 全球 800+ | 2026-06 | 中 | 公司 / 合作伙伴披露,未经审计 |
| 企业客户背书 | 30+ 家领先企业 | 2026-03 | 中 | Yicai 援引公司说法报道 |
| IPO 状态 | 据报道已向 HKEX 保密递表 | 2026-03 | 中 | 因保密递表,未有公开招股书 |
| 公开员工数 | 未可靠披露 | 2026 | 低 | 可取得公开资料未给出可对账的最新数字 |
混合了官方、合作伙伴与媒体披露。使用量和 IPO 指标并非经审计财务申报,只能视为公开信号型 KPI。
[CO001, CO006, CO012, CO015, CO020, CO025]创始人、模型、产品、合作伙伴和香港布局如何共同强化 BioMap 的平台战略。
[CO002, CO003, CO005, CO015, CO020, CO022]1.2 创始人与管理梯队
公司由 Baidu 创始人 Robin Li 和 Baidu Ventures 老将 Wei Liu 发起,Liu 负责日常经营。BioMap 目前公开的管理梯队不止两位创始人,这直接影响执行风险:网站资料列出 Xiaoming Zhang 负责 AI R&D,Xiaoyue Sun 负责运营并担任总法律顾问,Ziyao Xu 负责解决方案、战略和创新,Patrick Zhang 负责全球企业发展,Stan Z. Li 担任 AI 大模型首席科学家。这是一支技术可信的队伍,覆盖大模型工程、计算生物学、法务 / 运营和医药商务拓展;但叙事和战略控制也集中在 Robin Li 与 Wei Liu 身上。公开可得来源仍未披露董事会构成和正式治理权利,因此创始人影响力清晰可见,机构治理仍不透明。[CO002, CO003, CO009, CO010, CO011]
| 姓名 | 职务 | 背景 | 执行作用 | 关键人风险 |
|---|---|---|---|---|
| Robin Li | 创始人兼董事长 | Baidu 创始人、长期 AI 领军者;BioMap 主要发起人 | 战略背书、品牌与资本入口 | 高 - 象征性创始人集中风险 |
| Wei Liu | 联合创始人兼 CEO | 前 Baidu Ventures CEO、Baidu Group 副总裁;20 年 VC / 孵化背景 | 运营负责人,牵头合作、融资和香港扩张 | 高 - 核心高管发言人 |
| Xiaoming Zhang | 副总裁,AI 研发负责人 | 前 Ant Group AI 工程负责人;大模型架构专家 | 负责核心模型工程落地 | 中 |
| Xiaoyue Sun | 副总裁,运营负责人兼总法律顾问 | 10+ 年法律、治理与运营经验 | 跨境运营和合规离不开这个角色 | 中 |
| Ziyao Xu, PhD | 副总裁,解决方案、战略与创新负责人 | 计算生物学和药物发现专家 | 把模型接入客户工作流和科学场景 | 中 |
| Patrick Zhang | 副总裁,全球企业发展总裁 | 前 GenScript、Fosun Pharma 高管,具备资本市场背景 | 业务拓展与战略合作 | 中 |
| Stan Z. Li, PhD | AI 大模型首席科学家 | Westlake 讲席教授、高被引 AI 科学家 | 科学可信度和前沿模型研究 | 中 |
高管名单来自 BioMap 当前网站包和公开公司资料;公开来源未披露董事会组成或高管持股。
[CO002, CO003, CO009, CO010, CO011]1.3 资本形成、战略伙伴与香港扩张
BioMap 的公开融资时间线从 2021 年 7 月宣布的 $100 million Series A 开始;该轮由 GGV Capital 领投,Baidu、Legend Capital、BlueRun Ventures、Zhenzhi Capital、Xiang He Capital 参投,Robin Li 继续参与。此后,公开记录从经典风险融资转向战略平台合作。2023 年 10 月,Sanofi 联盟给 BioMap 带来一个标杆药企背书点:$10 million 首付款,以及超过 $1 billion 的潜在里程碑。2024 年 6 月,公司通过 HKIC 获得香港主权资本支持,并借此推出 BioMap InnoHub 和 BioX 加速器。HKSTP 与 HKIC 都把香港建设描述为不止开办公室:它被设计成面向跨国客户的信任、人才、生态和商业化楔子。到 2026 年 3 月,多家媒体报道公司已保密递表赴港 IPO,承销支持来自 CICC、Morgan Stanley 和 UBS。[CO012, CO013, CO014, CO015, CO016, CO017]
| 利益相关方 | 角色 / 关系 | 重要性 | 公开证据 |
|---|---|---|---|
| GGV Capital | Series A 轮领投方 | 锚定首轮外部机构融资 | Series A 轮报道 |
| Baidu / Robin Li | 创始发起支持方和早期投资人 | 提供创立可信度、资本和 AI 生态连接 | BaiduWiki、VCBeat、Yicai |
| Legend Capital | Series A 轮参投方 | 医疗投资人验证和企业网络支持 | ACN Newswire 新闻稿 |
| HKIC | 战略投资人 / 主权合作伙伴 | 带来政策支持、香港市场入口和生态聚合能力 | HKIC 和 HKSTP 新闻稿 |
| HKSTP / OASES | 香港生态赋能方 | 提供本地设施、合作伙伴入口和战略企业身份 | HKSTP 新闻稿 |
| Sanofi | 战略药企合作方 | 验证平台能服务生物药发现,并带来潜在里程碑收入 | Pharmaphorum 和 PMLive |
| Harbour BioMed | 2026 年创业孵化合作伙伴 | 把 BioMap 从 SaaS / 平台延伸到 AI 原生管线孵化 | PRNewswire 新闻稿 |
仅覆盖最影响决策的投资人与战略伙伴;可取得公开资料大多未披露具体股权、董事席位和现金承诺。
[CO012, CO013, CO015, CO018, CO022, CO023]| 日期 | 事件 | 类型 | 金额 / 状态 | 含义 |
|---|---|---|---|---|
| 2020-09-25 | BioMap 正式成立 | 成立 | 公司成立 | Baidu 关联 AI 生物学平台起点 |
| 2021-07-30 | Series A 轮发布 | 融资 | $100M | 平台建设获得首笔重要外部资本 |
| 2022-09-09 | 北京中心实验室和 ImmuBot 披露 | 产品 | 实验室 + 内部药物概念 | 显示其野心不止软件,还要做湿实验室 |
| 2023-10-10 | Sanofi 战略合作发布 | 合作 | $10M 首付款;>$1B 潜在金额 | 头部药企验证 |
| 2024-06-24 | HKIC 战略合作签署 | 融资 | 战略投资 | 增加主权资本支持和香港切入点 |
| 2024-06-24 | BioMap InnoHub 与 BioX 在香港启动 | 扩张 | 枢纽 + 加速器 | 国际化与生态战略 |
| 2025-03-24 | Kexing Biopharm 合作对外披露 | 合作 | 战略合作 | 商业 AI 药物发现用例 |
| 2025-04-29 | 生成式发现系统发布对外披露 | 产品 | 已宣布 | 显示产品化节奏 |
| 2025-06-30 | PFMBench 开源 | 产品 | 基准发布 | 外部开发者 / 科学界参与 |
| 2025-07-11 | RNAGenesis 发布 | 产品 | RNA 基础模型 | 从蛋白扩展出去 |
| 2025-08-13 | ProteinReasoner 发布 | 产品 | 多模态蛋白推理模型 | 向更高阶推理推进 |
| 2025-10-23 | BioLab 发布 | 产品 | 自主研究系统 | 迈向智能体式湿实验室编排 |
| 2026-03-17 | 香港保密递表被报道 | 融资 | 目标融资数亿美元 | 后期流动性事件准备 |
| 2026-06-15 | 与 Harbour BioMed 推出 MegaStream TechBio | 合作 | 联合 AI 原生管线公司 | 平台到管线的变现路径 |
这是一条合并后的事件时间线,来自官方、合作伙伴和媒体来源。若干 2025 年发布事项主要由公司新闻标题和关联论文支撑, 而不是详细新闻稿正文。
[CO001, CO012, CO015, CO017, CO022, CO025]2020 至 2026 年的创立、融资、香港扩张、产品发布和 IPO 准备里程碑。
若干 2025 年事项按公司新闻档案中的公告日标注,而不是按产品正式发布(GA)日期。
[CO001, CO012, CO015, CO022, CO025, CO036]1.4 规模信号与产品节奏
对一家私人 AI 生物技术公司而言,BioMap 的公开规模图景异常强,但数字来自公司和合作伙伴渠道,不是审计报告。HKIC 称,截至 2024 年 6 月,BioMap 已与 200 多名 AIGP 用户签约;HKSTP 称公司已与 10 多个商业伙伴、200 多家学术机构合作;Yicai 与 Harbour BioMed 后来引用了 800 多家机构用户的支持,Yicai 还称有 30 多家领先企业。2025 与 2026 年,产品节奏也在加快。官方新闻标题和论文记录了生成式发现系统、RNAGenesis、ProteinReasoner、PFMBench 和 BioLab 等产品或研究的发布 / 披露;2026 年 6 月 Harbour 发布的信息显示,平台正在被商业化为一家新的 AI 原生 复杂生物药合资企业。含义很清楚:BioMap 不再只是模型构建者;它试图成为生物学的操作系统和孵化新公司的底层。[CO019, CO020, CO021, CO032, CO033, CO034]
围绕科研规模、商业化、主权资本支持和披露质量的公开核心指标。
若干指标来自公司声称或媒体报道,意在作为可投性信号,而非经审计 KPI。
[CO012, CO020, CO023, CO025, CO030, CO034]1.5 判断、矛盾与剩余尽调空白
公司概览层面的证据足以证明 BioMap 是一家严肃、资金和资源支持较强的后期 AI 生物平台,但还不足以精确支撑估值或治理判断。公开来源支持几件事的存在:有意义的企业和机构使用、可信的管理梯队、标杆 Sanofi 合作,以及强香港政策支持。它们没有给出经审计财务报表、可调和的当前估值、可靠员工数、董事会构成,或覆盖北京、香港与美国的清晰法律实体图。此外,总部表述不一致,许多牵引力 数字来自公司口径。因此,后续章节可以稳妥复用 BioMap 的身份、产品栈、战略合作和 IPO 轨迹,但应把资本效率、客户集中度和治理结构视为仍待尽调的问题,而不是既定事实。[CO007, CO008, CO020, CO025, CO026, CO028]
1.6 要点展示
02市场分析
2.1 市场边界与计入口径
BioMap 不应按整体药企 R&D 或整体生物技术软件支出来估值。最相关的核心类别是 AI 驱动药物发现平台:卖给药企、生物技术公司、CRO、CDMO 和科研机构的软件、模型、整理后的数据、湿实验整合和配套服务,用于靶点识别、命中发现、先导优化、ADMET、转化建模以及临床前配套工作。这个边界最贴近 BioMap 对 BioMap OS、AIGP 和基础模型栈的描述。更宽但仍相关的邻近市场是完整药物发现技术市场,涵盖仪器、试剂、软件和筛选系统。BioMap 还延伸到传统药企之外的合成生物学和绿色技术,因此没有单一第三方市场报告能完美覆盖公司。最干净的方法是使用多重视角,并明确区分较窄的 AI 发现市场与更宽的赋能技术栈市场。[CM001, CM002, CM003, CM007, CM034, CM036]
| 细分 / 类别 | 纳入支出 | 排除支出 | 主要买方 / 付款方 | 对 BioMap 的意义 |
|---|---|---|---|---|
| AI 驱动的药物发现平台 | 模型软件、数据、工作流工具、湿实验室集成及相关服务 | 下游商业药品销售、完整临床商业化 | 药企、生物科技公司、CRO、研究机构 | 最接近 BioMap OS / AIGP / xTrimo 商业模式 |
| 药物发现技术(广义技术栈) | 软件、仪器、试剂、筛选和验证技术 | 后期生产与商业分销 | 大药企 R&D、生物科技实验室、平台运营方 | AI 系统可能影响预算的上限 |
| 发现信息学 / 分析 | 靶点数据、测序分析、对接、分析软件 / 服务 | 软件工作流之外的实体实验室基础设施 | 发现信息学团队、转化医学团队 | BioMap 软件价值捕获的中低位相邻市场 |
| 生物药发现平台 | 抗体 / 蛋白设计、多参数优化、可开发性工具 | 孤立的小分子专用工作流 | 生物药和蛋白疗法团队 | BioMap 切在高增长生物药工作流里,因而重要 |
| 合成生物学 / 绿色科技设计平台 | 面向工业和可持续场景的蛋白 / 通路设计 | 与生物学无关的通用工业软件 | 合成生物学公司、工业创新团队 | 解释 BioMap 的 TAM 为何不止药企 |
这张边界表把狭义 AI 平台市场与更广的发现技术、跨行业生物学预算分开,避免夸大 BioMap 核心可变现市场。
[CM001, CM002, CM003, CM007, CM008, CM034]三层嵌套预算池:狭义 AI 发现 TAM、更广的发现技术邻近市场,以及 BioMap 隐含的跨行业机会集。
[CM004, CM007, CM034, CM036, CM040]2.2 测算视角:窄口径 AI 发现与广口径发现技术栈
Global Market Insights 给出的最窄可用视角,把药物发现 AI 市场规模估为 2025 年 $3.1 billion、2026 年 $4.0 billion,并以约 30.5% 的复合增速到 2035 年达到 $43.9 billion。Precedence Research 描述了类似市场,并按工作流、模态、AI 栈、治疗领域和区域拆分,进一步说明这已经是一个多分部平台市场,而不是单产品小众赛道。MarketsandMarkets 给出的更宽技术栈视角大得多:药物发现技术市场在 2025 年为 $30.58 billion,到 2030 年达到 $51.51 billion。这个宽口径高估了 BioMap 当前可变现市场,因为它包含仪器和试剂;但它确实捕捉到 AI 原生 平台试图替代或编排的更大预算池。承销时,BioMap 的现实市场落在这两个端点之间。[CM004, CM005, CM006, CM007, CM008, CM009]
| 发布方 / 口径 | 年份 / 周期 | 地域 | 数值 | 方法 / 范围 | 置信度 | 局限 |
|---|---|---|---|---|---|---|
| Global Market Insights - 药物发现 AI | 2026 / 2035 | 全球 | 2026 年 $4.0B -> 2035 年 $43.9B | 药物发现 AI 市场 | 中 | 供应商市场报告;类别较宽,但仍聚焦 AI |
| Global Market Insights - 药物发现 AI | 2025 | 全球 | $3.1B in 2025 | 同一类别的上一年基准 | 中 | 不是 BioMap 专属 SAM |
| Precedence Research - AI 驱动的药物发现平台 | 2025-2035 | 全球 | 可访问摘要只有叙述式分拆 | 按工作流、模态和地区划分的平台市场 | 低 | 可访问摘录没有给出清晰的整体市场规模 |
| MarketsandMarkets - 药物发现技术 | 2025 / 2030 | 全球 | 2025 年 $30.58B -> 2030 年 $51.51B | 更广的发现技术栈,包含软件、试剂和仪器 | 中 | 包含非软件支出,会高估 BioMap 直接 TAM |
| MarketsandMarkets - 药物发现信息学 | 2025 | 全球 | $3.5B by 2025 | 软件 / 信息学子集 | 低 | 历史预测摘录,不是 2026 年点估计 |
| MarketsandMarkets - 生命科学分析 | 2025 / 2030 | 全球 | 2025 年 $40.03B -> 2030 年 $68.81B | 贯穿生命科学工作流的分析 | 低 | 极宽的相邻市场,不是 BioMap 核心市场 |
核心 TAM 用 AI 专属口径;更广的发现技术口径只作为外沿上限。两类口径衡量的支出池不同,因此刻意不做平均。
[CM004, CM007, CM008, CM009, CM010, CM040]用公开第三方报告,对 BioMap 相关支出池给出低 / 基准 / 高三种市场视角。
这不是对同一对象的三组估算。图中有意把较窄到较宽的类别作为有边界的视角并列呈现。
[CM004, CM007, CM010, CM040]2.3 买方图谱与采用路径
公开买方证据显示,BioMap 进入的是科研预算,而不是通用 IT 预算。HKIC 称,当前 AIGP 用户包括国际药企、CDMO、创新药开发商、合成生物学公司、绿色科技企业和科研机构。Harbour BioMed 把 BioMap 描述为服务抗体 / 蛋白、精准医学、合成生物学和前沿研究用户的平台。可比公司指向类似需求中心:Recursion 通过自有和合作管线变现,Absci 通过内部和合作的生成式生物学项目变现,Schrödinger 通过发现软件变现,Certara 通过软件加监管与开发服务变现,Insilico 则同时通过平台软件和管线授权变现。实际采用路径通常从围绕靶点识别 或先导优化的试点开始,扩展到工作流集成和数据积累,最后才进入更深的共同开发、收入分享或合资孵化。[CM018, CM019, CM029, CM030, CM031, CM032]
| 细分市场 | 买方 | 用户 | 付款方 / 预算负责人 | 工作流 | 采用触发点 |
|---|---|---|---|---|---|
| 大药企生物药发现 | 研究平台负责人 / 治疗领域 VP | 计算生物学家、抗体团队 | R&D 平台预算 | 靶点识别、先导优化、可开发性 | 需要加速生物药项目或提高命中率 |
| 新兴生物科技 / TechBio | 创始人 CEO、平台负责人 | ML 科学家、转化生物学团队 | 公司 R&D / 风险资本支持预算 | 从靶点到管线的一体化发现技术栈 | 希望用有限团队和更快迭代参与竞争 |
| CDMO / CRO / 服务提供商 | 服务线总经理或创新负责人 | 工艺科学家、检测团队 | 资本开支 + 平台赋能预算 | 给服务交付加 AI 层 | 需要差异化通量和客户成效 |
| 学术 / 研究机构 | PI、研究所负责人、创新办公室 | 研究人员、博士、博士后 | 课题 / 机构经费 | 模型辅助研究和数据解读 | 需要使用基础模型,但不自建 |
| 合成生物学 / 绿色科技 | CTO / 平台负责人 | 蛋白工程师、菌株团队 | 研发预算 | 蛋白 / 通路设计与优化 | 非药物生物场景需要更快设计 |
买方地图把第三方市场定义与 BioMap 已披露的用户垂直领域放在一起;不同客户类别的采购机制并不相同,且未公开披露。
[CM018, CM019, CM026, CM029, CM032, CM033]跨细分比较科学工作流强度、预算集中度、湿实验耦合度和战略合作上行空间。
[CM029, CM030, CM031, CM032, CM033, CM035]商业采用通常从限定范围试点开始,走向平台化集成,再进入收益共担更深的模式。
[CM029, CM030, CM033, CM037, CM038, CM039]2.4 增长驱动与采用约束
这个市场的多头逻辑很直接:生物药复杂度上升、R&D 成本更高、开发周期需要压缩、算力和模型架构更好,以及药企对 AI 合作的强需求。IQVIA 还指出,2025 年融资仍处高位,虽然低于 2024 年;AI 赋能 项目开始呈现更强成功信号。但约束同样重要。市场报告反复点名数据质量、带标签数据有限、专门人才短缺和湿实验验证需求是瓶颈。McKinsey 强调,药企不能照搬消费软件打法来规模化生成式 AI,因为受监管数据、验证和工作流改造会拖慢采用。对 BioMap 来说,市场规模不是最关键的闸门;从试点走到规模化平台标准,能否证明 ROI 才是。[CM011, CM012, CM013, CM014, CM015, CM016]
| 驱动 / 约束 | 方向 | 时间 | 含义 | 尽调问题 |
|---|---|---|---|---|
| 生物制剂需求上升,多目标设计更复杂 | 驱动 | 现在 | 利好基础模型平台,而非单点工具 | 量化 BioMap 需求中有多少来自生物制剂 |
| 需要缩短发现周期、提升生产率 | 驱动 | 现在 | 支撑药企研发部门的 AI 试点预算 | 索取客户 ROI 案例和命中周期指标 |
| 亚太政策支持与涉华交易 | 驱动 | 2025-2026 | 支撑香港 / 中国平台采用和资本获取 | 检验需求中政策驱动和自下而上各占多少 |
| 数据质量和标注数据稀缺 | 约束 | 现在 | 限制模型表现,拖慢部署 | 追问专有数据优势和反馈闭环 |
| AI 与生物学交叉人才稀缺 | 约束 | 现在 | 抬高客户和供应商的落地成本 | 索取 BioMap 按模型、科学和解决方案岗位拆分的人力配置 |
| 湿实验验证瓶颈 | 约束 | 现在 | 无法形成纯软件经济,也拖慢价值证明 | 审计干湿闭环吞吐量和实验室利用率 |
| 受监管工作流变更和信任要求 | 约束 | 中期 | 拉长企业销售周期,加重验证负担 | 按客户类别索取续约和扩张数据 |
| 试点转平台的转化风险 | 约束 | 中期 | 大 TAM 不等于规模化合同 | 获取 200+ / 800+ 用户中的生产与试点拆分 |
本表把已披露的市场驱动因素和实际商业化阻力配对。最后三行是不能把表面 TAM 直接转成激进 BioMap 收入假设的主要原因。
[CM011, CM012, CM013, CM014, CM020, CM021]2.5 判断与剩余市场缺口
市场显然跨过了“足够大”的门槛。即使用最窄的 AI 发现视角,也能看到一个快速增长的数十亿美元级品类;更宽的发现技术栈视角则指向 AI 原生 系统长期可以影响的更大支出池。更难的问题不是 TAM,而是捕获能力。BioMap 必须把科学可信度和跨境政策支持转化为药企及邻近生物学客户内部可重复的预算。公开来源仍未按买方、地区或模态拆出 BioMap 的真实 SAM,也没有披露合同规模、续约行为,或当前用户中试点与规模化生产的占比。因此,本章支持积极的市场观点,但在缺少客户级证据前,不能支撑精确渗透率假设。[CM034, CM035, CM036, CM037, CM038, CM040]
2.6 要点展示
03竞争格局
3.1 格局:直接对手、邻近玩家、替代方案与现状
BioMap 的竞争集合比“其他 AI 药物发现创业公司”更宽。直接对手是那些试图压缩发现周期、构建专有生物学模型,并常常捕获管线经济的 AI 原生平台公司:Insilico、Recursion、Absci、Owkin、Valo、insitro 和 BenevolentAI 都属于这一桶。邻近竞争者包括 Schrödinger、Certara 等发现软件和仿真供应商;它们未必把自己包装成蛋白基础模型公司,但已经嵌在买方工作流和预算里。现状替代方案仍是药企内部发现团队,把点状工具、CRO / CDMO 和既有服务商拼在一起。这一点很关键,因为 BioMap 不是在单一赛道对抗一个敌人;它面对的是内部自建、专用工具和一体化 TechBio 平台叠出来的一整套栈。也正因这种多层格局,评估 BioMap 时,不应只看某个对手首页是否写了类似主张,而要看在具体购买时点、具体发现项目、企业预算和合作结构下,哪一类对手最容易替换客户工作流。[CP001, CP002, CP003, CP004, CP005, CP006]
按商业模式深度和生物平台广度拆分竞争场。
序位评分综合了公开能力和商业模式证据;不是财务倍数图。
[CP013, CP015, CP016, CP017, CP018, CP026]3.2 画像对比与规模
公开可比公司显示,市场分成偏管线的 TechBio 和偏经常性软件的供应商。Recursion、Absci 和 Insilico 强调内部及合作管线;Owkin、insitro、Valo 和 BenevolentAI 强调 AI 原生 R&D 平台、自主或因果生物学工作流;Schrödinger 与 Certara 强调软件、仿真和开发服务。2026 年 7 月,Recursion、Absci、Schrödinger 和 Certara 的公开市值都在约 $1.0–$1.8 billion,为公开投资者如何给邻近发现平台定价提供了现实校准。BioMap 的优势是仍能围绕 xTrimo 和亚洲相关扩张讲私人市场增长故事;劣势是公开披露比上市可比公司更薄。[CP011, CP012, CP013, CP014, CP019, CP020]
| 竞争对手 | 类别 | 规模 / 融资信号 | 目标细分市场 | 差异化 | 局限 |
|---|---|---|---|---|---|
| Insilico Medicine | 直接对标 / TechBio | 未上市、多项目平台,叙事围绕许可展开 | 药企、生物技术公司、内部管线 | 端到端 AI 技术栈,覆盖靶点识别至 II 期 | 未上市公司披露仍然选择性较强 |
| Recursion | 直接对标 / 上市 TechBio | $1.75B 市值(2026 年 7 月) | 药企合作 + 内部管线 | 临床阶段 TechBio,有公开市场披露 | 管线经济性叠加生物技术风险 |
| Absci | 直接对标 / 上市生物制剂 AI | $1.72B 市值(2026 年 7 月) | 生物制剂发现买方 | 聚焦生成式 AI 生物制剂 | 模态侧重点比广义生物学平台更窄 |
| Owkin | 直接对标 / AI 生物学平台 | 未上市平台叙事 | 医疗健康 / 研发 / 数据丰富买方 | 主打自主 AI 科学家和多模态生物学 | 对蛋白设计的聚焦不够明显 |
| insitro | 直接对标 / ML 驱动药物公司 | 未上市平台叙事 | 药物发现机构 | 大规模数据驱动的 ML 药物公司模式 | 公开商业化包装细节有限 |
| Valo Health | 直接对标 / 因果生物学平台 | 未上市平台叙事 | 药物发现团队 | 闭环化学 + 人类因果生物学 | 定价 / 合同公开细节较少 |
| BenevolentAI | 直接对标 / AI 制药研发平台 | 公开定位为 AI 药物发现 | 研发决策支持和发现团队 | 生命科学情报与研发决策平台 | 跨行业生物学角度不够明显 |
| Schrödinger | 相邻软件老牌厂商 | $1.22B 市值(2026 年 7 月) | 计算化学 / 分子设计 | 基于物理的软件,已嵌入工作流 | 基础模型原生叙事较弱 |
| Certara | 相邻软件 / 服务老牌厂商 | $1.05B 市值(2026 年 7 月) | 药物开发软件 + 服务买方 | 覆盖生物制药、学术界、监管方的 2,400+ 客户 | 更偏开发赋能,而非前沿生成式生物学 |
公开画像混合了公司官网、投资者关系、SEC 浏览页面和市值追踪器。未上市公司规模往往只能定性判断,因为确切收入和估值数据未披露。
[CP002, CP003, CP004, CP005, CP006, CP007]| 采购维度 | BioMap | Insilico | Recursion | Absci | Schrödinger | Certara |
|---|---|---|---|---|---|---|
| 蛋白基础模型 | 强 | 中 | 中 | 中 | 低 | 低 |
| 干湿闭环编排 | 强 | 中 | 中 | 中 | 低 | 低 |
| 生物制剂 / 抗体侧重 | 强 | 中 | 中 | 强 | 低 | 低 |
| 公开市场披露 | 低 | 低 | 高 | 高 | 高 | 高 |
| 纯软件工作流适配度 | 中 | 中 | 低 | 中 | 强 | 强 |
| 跨行业生物学(合成 / 绿色科技) | 强 | 低 | 低 | 低 | 低 | 低 |
能力分档按证据判断,不做数字打分。「低」往往表示缺少可公开获取的证明,而不是没有能力。
[CP013, CP014, CP022, CP023, CP026, CP028]| 公司 | 价格 / 合同模式 | 包含能力 | 未知项 | 含义 |
|---|---|---|---|---|
| BioMap | 定制企业合同 / 合作伙伴 / 可能叠加里程碑 | BioMap OS、xTrimo、AIGP、湿实验室集成、合作伙伴 | 未公开标价或 ACV | 对标定价需要客户访谈 |
| Insilico | 软件 + 管线 / 许可叙事 | 覆盖靶点识别至 II 期的技术栈与软件 | 未公开自助式定价 | 更接近成果 / 许可经济,而非 SaaS |
| Recursion | 合作伙伴关系和管线经济 | TechBio 平台和内部项目 | 官网未完全公开合同条款 | 部分竞争点是共享上行收益,不是价目表 |
| Absci | 内部 + 合作生成式生物制剂项目 | 生物制剂设计和项目创建 | 未公开标价 | 大概率由企业合同和合作牵引 |
| Schrödinger | 软件平台 + 许可 / 企业模式 | 基于物理的分子发现软件 | 官网信息不足以形成可用定价对标 | 包装更像工具,而非 BioMap 式平台 |
| Certara | 软件 + 服务 | 生物仿真、开发与监管支持 | 未公开定价 | 最接近平台 + 服务经常性经济的类比 |
这一组主要同行都没有公开足够定价细节,无法做真正同口径 ACV 对标;市场仍由合作伙伴关系和企业合同牵引。
[CP019, CP020, CP021, CP022, CP023, CP033]对比 BioMap 相比直接和相邻替代方案在哪些能力更强。
[CP013, CP019, CP022, CP023, CP028, CP031]3.3 切换成本、多供应商并用与分发权
这个市场的切换成本可能低于企业基础设施软件。买方常常能并行试用多个供应商,许多平台也把自己呈现为靶点识别、优化、仿真或模型辅助工作流步骤中的模块化层,而不是完整排他的操作系统。只有当 BioMap 嵌入专有湿实验数据循环、微调模型或共享经济合作时,才会形成真正锁定。因此,分发权掌握在已经占住预算线和信任关系的公司手里:大型药企内部平台、软件既有玩家和规模化服务商。这并不意味着 BioMap 没有竞争力,但确实意味着单靠科学性能通常不够。采用必须沉淀成专有客户数据、工作流集成和 ROI 证明,护城河说法才会变硬。实际情况是,一个客户既用 Schrödinger 做分子建模,又用 Certara 做开发分析,并拥有内部生物药团队,也仍然可以试用 BioMap,而无需完全标准化到 BioMap 上。因此,公司需要证明,至少部分客户正在从评估走向更深工作流,在这些工作流里,数据尾迹、实验设计和多目标优化越来越依赖 BioMap 特定基础设施。[CP019, CP026, CP027, CP028, CP029, CP030]
| 护城河主张 | 威胁 | 严重性 | 缓解措施 / 尽调问题 |
|---|---|---|---|
| xTrimo 规模与生物学专精 | 基础模型商品化 | 高 | 验证客户是否从专有微调和数据闭环中受益 |
| 干湿闭环 | 买方把湿实验数据留在供应商栈外 | 高 | 审查真实客户的数据反馈集成 |
| 香港信任与生态 | 全球竞争对手披露更多,买方可能觉得更安全 | 中 | 索取跨国客户背书与合规材料 |
| 跨行业垂直覆盖 | 资源在制药、绿色科技、合成生物学之间被摊薄 | 中 | 按细分市场核查资源分配 |
| 平台到管线的项目公司孵化 | 执行复杂、资本消耗高 | 中 | 审计 MegaStream 类项目公司的经济性和治理 |
| 科学品牌 / 领导梯队 | 关键人物与叙事依赖 | 中 | 审查接班梯队深度和商务拓展能力 |
竞争风险不只是某个功能缺口,更在于 BioMap 能否赶在品类商品化前,把科学主张转成粘性客户可控数据和耐久经济性。
[CP026, CP027, CP028, CP029, CP030, CP031]3.4 护城河耐久性与被替代风险
BioMap 最强的差异化论证,不是“药物发现里只有它用 AI”,而是它是少数几家同时结合超大生物学基础模型、干湿闭环产品架构、香港生态支持,并愿意跨药企和邻近生物学垂直领域商业化的私人玩家。因此,护城河是有条件的,不是绝对的。如果基础模型商品化,而客户数据仍不在 BioMap 控制中,公司可能退化成“服务加模型”供应商。反过来,如果它持续把试点转化为专有反馈循环和 MegaStream 这类共享经济合资,其位置会更难复制。真正的竞争问题是 BioMap 能否跑得比市场商品化更快。如果 BioMap 成功,它会更像平台拥有者,而不是窄口径软件供应商:每一个伙伴特定循环都会改善模型输出。如果失败,披露更清晰的上市竞争者和采购路径更成熟的既有玩家,仍可能拿走 BioMap 瞄准的预算池。[CP013, CP014, CP026, CP027, CP028, CP029]
压缩读数:BioMap 哪里有优势,哪里仍会被竞争或商品化挤压。
定性 KPI 标签是有证据支撑的判断,来自竞争对手披露,以及 BioMap 已披露的产品 / 合作姿态。
[CP014, CP026, CP027, CP028, CP029, CP030]3.5 要点展示
04财务
4.1 收入来源与变现逻辑
BioMap 的公开收入模型只能从交易结构和客户数量披露中看出轮廓,而不是来自清晰价目表或收入科目。最明确披露的收入流是合作伙伴经济:Sanofi 合作包含 $10 million 首付款和超过 $1 billion 的潜在里程碑;MegaStream 则带来首付款、成功里程碑和特许权使用费分享的可能。HKIC 和 HKSTP 还描述了 200 多名签约 AIGP 用户,支持 BioMap 存在平台合同收入,而不是纯粹投机性使用。Yicai 提到的 60 多个项目和 800 多家机构说明,BioMap 也通过项目交付和企业赋能变现。同时,与 Kexing、Optoseeker 的合作公告显示,这是一种定制化、重集成的模式,BioMap 贡献模型、编排能力和实验工作流支持。这种混合结构在财务上很重要:它比纯研究叙事更能证明客户愿意付费,但也意味着收入质量取决于合同组合、里程碑时点和服务强度,而这些目前都没有公开拆分。[CI001, CI002, CI003, CI007, CI008, CI009]
| 收入流 | 机制 | 单位 | 当前状态 | 质量 | 尽调问题 |
|---|---|---|---|---|---|
| Sanofi 合作 | 预付款 + 里程碑 | 项目经济 | 预付款 $10M;潜在里程碑 >$1B | 高价值,但取决于里程碑 | 索取已确认收入和阶段关卡 |
| AIGP 平台合同 | 签约用户 | 账户 / 合同 | HKIC/HKSTP 披露 200+ 用户 | 需求证据不错,但价格未知 | 索取 ACV 和续约数据 |
| BioMap OS 发现项目 | 项目交付 | 项目 | Yicai 披露 60+ 个已验证项目 | 有牵引力,但收入未知 | 索取单项目收入和试点转化率 |
| 企业客户基础 | 机构 / 企业关系 | 客户 | 引用 800+ 家机构和 30+ 家头部企业 | 是规模信号,不是收入 | 拆分付费与非付费用户 |
| 实验工作流服务 | 湿实验室和筛选支持 | 服务项目 | 披露苏州中心部署和试用意向 | 大概率人力 / 资本开支密集 | 索取分服务线毛利率 |
| MegaStream 上行空间 | 预付款、里程碑、版税 | 资产经济 | 架构已宣布;未公开入账金额 | 潜在上行高,但仍偏推测 | 索取所有权、收入分成和会计处理 |
公开来源显示多条变现路径,但只有 Sanofi 披露的经济性足够精确,可以直接引用。
[CI001, CI002, CI003, CI004, CI005, CI008]| 项目 | 价格 / 价值 | 单位 | 公开状态 | 含义 |
|---|---|---|---|---|
| Sanofi 预付款 | $10M | 一次性预付款 | 已披露 | 证明大型药企愿意为平台付费 |
| Sanofi 里程碑 | >$1B | 开发 / 监管 / 商业化里程碑 | 已披露 | 经济收益偏后端 |
| 2026 年 IPO 目标 | 数亿美元 | 融资 | 已报道,但未公开披露 | 说明仍有持续资本需求 |
| 潜在订单总额 | $2B | 管线机会 | Sohu 报道 | 不等同于已确认收入 |
| AIGP 合同定价 | 未披露 | 按合同 / 账户 | 未公开 | 无法推断 ACV 或实际成交价 |
| 项目 / 服务定价 | 未披露 | 按项目 / 实验 | 未公开 | 无法分析利润率和收入结构 |
本章可以引用交易金额,但 BioMap 核心平台业务仍没有公开价目表,也没有披露合同金额。
[CI002, CI007, CI010, CI018]BioMap 如何把平台能力、合作和项目转成变现路径。
[CI001, CI002, CI003, CI004, CI008, CI009]4.2 公开牵引力与 GTM 效率代理指标
BioMap 的公开牵引力指标在商业上令人鼓舞,但财务上不好解读,因为它们混合了几种不同分析单位。香港官方来源引用超过 200 名签约 AIGP 用户;Yicai 和 PR Newswire 引用超过 800 家机构、30 多家领先企业和 60 多个已验证项目。这些数字说明,公司不是只靠一个合作伙伴讲故事,而是在药企、CDMO、科研团队、合成生物学玩家和绿色科技公司之间形成了真实企业级 GTM 动作。它们也说明公司可以以多种形式落地:平台合同、定制项目、数据密集型共同开发和实验工作流部署。问题是,没有任何来源调和这些机构中有多少在实质付费、多少是试点、头部客户支出集中度如何,或总量中有多少属于经常性收入。因此,BioMap 在漏斗顶部证明上比销售效率上更强。已披露指标支持 BioMap 具备有意义商业触达的观点,但还无法计算 CAC、回本周期、NRR,甚至无法可靠估算 ACV。[CI004, CI005, CI006, CI011, CI012, CI013]
| 指标 | 数值 / 状态 | 置信度 | 重要性 | 尽调要求 |
|---|---|---|---|---|
| 已确认收入 | 未公开 | 低 | 核心测算输入缺失 | 要求提供经审计收入桥接表 |
| 毛利率 | 未公开 | 低 | 决定软件化与服务化收入质量 | 要求按收入流拆分毛利率 |
| 客户规模代理指标 | 200+ 份合同;800+ 家机构;60+ 个项目 | 中 | 能显示需求广度,但单位口径混杂 | 拆分试点、活跃账户和付费客户 |
| 销售效率 | 只有间接代理指标 | 低 | 需要 CAC / 回本周期 / 转化率 | 要求提供漏斗和销售产能数据 |
| 交付成本驱动因素 | 算力 + 数据 + 湿实验室 + 服务结构 | 中 | 大概率压制毛利率 | 要求按工作流拆分成本 |
| 经常性收入占比 | Unknown | 低 | 决定现金生成的韧性 | 要求按合同类型拆分收入结构 |
| 在手订单质量 | $2B 潜在订单已有报道,转化率未知 | 低 | 在手订单可能高估近期变现能力 | 要求提供已签在手订单和转化节奏 |
几乎所有关键单位经济性字段仍未公开;公开牵引指标有帮助,但解决不了投资测算问题。
[CI004, CI005, CI006, CI015, CI017, CI024]仅靠公开证据,很难估算 BioMap 的利润率和回本周期。
公开来源能识别成本中心和销售渠道,但没有给出二者之间的数值转化关系。
[CI015, CI017, CI023, CI024, CI033]可用于投资测算的公开交易规模和商业化代理指标,不是已确认收入。
来源给出精确点值时,低 / 基准 / 高相同;IPO 行反映媒体所称「数亿美元」表述,并非法定申报金额。
[CI002, CI007, CI018, CI019]4.3 成本结构、资本强度与资金充足性
成本基座可能比“AI 平台”这个标签暗示的更重。BioMap 2021 年 Series A 据报约 $100 million,且明确用于 R&D 和人才;VCBeat 还提到公司在建设自有实验室。后续披露又增加了更多运营承诺:香港 InnoHub、五年支持 50 个项目的 BioX 加速器、苏州高通量实验部署,以及 MegaStream 的 AI 原生 干湿基础设施雄心。这些布局在战略上合理,但不是纯软件经济。它们意味着公司要在模型训练、数据生成、自动化、湿实验能力和生态支持上持续花钱。公开资本可见度仍不完整。Sohu 报道累计融资超过 $200 million,HKIC 宣布战略投资,但 HKIC 支票规模没有公开;本次审阅没有任何来源披露账上现金、月度现金消耗、现金跑道 或债务义务。因此,据报寻求数亿美元募资的保密赴港 IPO,是目前最清晰的公开信号:下一阶段商业化和规模扩张需要更多资本。[CI018, CI019, CI020, CI021, CI022, CI023]
| 项目 | 数值 / 状态 | 截至时间 | 备注 |
|---|---|---|---|
| A 轮融资 | $100M | Jul 2021 | 主要用于研发和人才 |
| 累计融资 | >$200M | Mar 2026 | 据 Sohu;具体构成未公开 |
| HKIC 战略投资 | 已宣布;金额未披露 | Jun 2024 | 增强背书,但披露透明度提升有限 |
| 香港 IPO 目标 | 数亿美元 | Mar 2026 | 最强的公开下一轮融资信号 |
| 计划资金用途 | 研发、人才、InnoHub、BioX、商业化 | 2024-2026 | 多条规模化路径 |
| 现金 / 烧钱率 / 现金跑道 | 未披露 | 2026 | 重大尽调阻碍 |
| 债务 / 项目融资 | 未公开披露 | 2026 | 本章来源未发现明确义务 |
| 下一轮触发因素 | 公开招股书 / 上市执行 | 2026+ | IPO 流程很可能决定资本可见度 |
资本充足性方向上偏正面,但缺少现金、烧钱率和准确轮次规模披露,无法精确量化。
[CI018, CI019, CI020, CI021, CI022, CI024]| 缺失指标 | 影响 | 具体尽调路径 | 优先级 |
|---|---|---|---|
| 经审计收入 / ARR | 无法判断规模或增长 | 要求提供经审计的 2024-2026 收入桥接表 | 关键 |
| 按收入流拆分的收入结构 | 平台、服务与里程碑依赖度不明 | 要求按收入流拆分收入 | 关键 |
| 按工作流拆分的毛利率 | 无法评估软件化质量或服务拖累 | 要求按产品线拆分服务成本 | 关键 |
| 客户集中度 / NRR | 无法判断收入韧性或议价能力 | 要求提供客户分群、续约率和头部客户敞口 | 关键 |
| 现金、烧钱率和现金跑道 | IPO 前无法测算资本充足性 | 要求提供月度现金桥接和跑道情景 | 关键 |
| IPO 招股书 / 募资用途 | 保密递表挡住直接尽调 | HKEX 文件公开后再复核 | 高 |
缺失的数据不是边缘项,而是得出高置信财务结论的核心阻碍。
[CI024, CI029, CI035]定性展示资本可能消耗在哪里,以及哪些地方已有变现信号。
[CI018, CI021, CI022, CI023, CI024, CI034]4.4 财务判断
BioMap 的财务画像强于一个尚未产生收入的 AI 生物技术概念,但弱于公开市场承销通常要求的水平。公司有真实商业化信号:合作伙伴愿意付费、合同数量、项目量、广泛机构用户基础,以及从平台走向管线的持续扩张。但这些信号还未形成可辨识的收入质量图景。投资者仍缺经审计收入、按收入流拆分的组合、集中度、毛利率、现金消耗和现金跑道。因此,核心问题不是 BioMap 有没有业务活动;它显然有。问题是其经济性主要来自经常性平台收入、里程碑较重的合作收入、劳动密集型发现工作,还是某种未来多年仍会渴求资本的混合体。当前合理承销姿态,是把 BioMap 视为平台、服务和资产创造混合公司:需求证明有前景,但财务透明度不足。任何估值论证都应以三件事为条件:积压订单转收入的转化、按工作流测算的交付成本,以及香港上市进程隐含的资本需求。[CI026, CI027, CI028, CI029, CI031, CI032]
4.5 要点展示
05产品与技术
5.1 BioMap 交付什么
BioMap 并不像单一用途药物设计 API,也不是单一内部管线。公开材料把 BioMap OS 描述为干湿闭环发现系统,可作为模型、软件或完整系统交付,拥有近 100 个可组合模块,并支持云端和本地化部署。产品叙事围绕四个垂直系统展开:抗体与创新蛋白、创新疗法与精准医学、合成生物学和前沿研究。这个框架很关键,因为它说明 BioMap 的商业供给以工作流为中心。平台要帮助客户从数据解读走到参数优化、de novo 设计和实验反馈,而不是只跑一个预测任务。产品家族还包括 xTrimo 基础模型线,这是 BioMap OS 底层引擎。公开证据支持一个判断:BioMap 卖的是生命科学 R&D 的操作层,把模型、编排和服务捆在一起,而不是一个窄口径点状工具。[CE001, CE002, CE003, CE012, CE037]
| 模块 / 资产 | 主要用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| xTrimo 基础模型 | BioMap + 研究合作伙伴 | 活跃 | 大型多模态生物学基础模型 | 需要逐版本核对基准测试 |
| BioMap OS | 企业研发团队 | 活跃 | 干湿结合的闭环发现编排 | 需要实施和可用性证明 |
| AIGP 平台 | 蛋白设计客户 | 活跃 | 蛋白设计与优化工作流 | 需要 SKU / 合同细节 |
| 苏州高通量实验室 | 内部 + 合作伙伴项目 | 活跃 | 闭环验证和数据生成 | 需要吞吐量 / 利用率指标 |
| MegaStream 技术栈 | 复杂生物制剂项目 | 早期成形 | 数据集 × 定制模型 × 实验室闭环 | 需要所有权和生产就绪度细节 |
| 研究模型家族(RNA / 智能体 / 结构) | 研究用户 | 不一 | 路线图已扩到蛋白之外 | 需要按模块提供商业化证据 |
BioMap 的资产横跨模型、软件、服务和基础设施,而不是单一 SKU。
[CE001, CE003, CE014, CE018, CE023, CE024]| 用户任务 | 当前工作流 | BioMap 方案 | 可衡量收益 | 局限 |
|---|---|---|---|---|
| 抗体 / 蛋白发现 | 筛选 + 设计迭代 | 蛋白模型 + OS + 湿实验室闭环 | 更快发现和优化候选物 | 未披露公开 ROI 指标 |
| 精准医学靶点发现 | 细胞数据解读 + 实验设计 | BioMap OS 虚拟细胞 / 扰动工作流 | 支撑 FIC 靶点发现主张 | 公开结果颗粒度有限 |
| 合成生物学优化 | 酶 / 菌株 / 工艺调优 | AI 驱动的多目标优化 | 宣称提升产量 / 效率 | 公开案例偏少 |
| 单细胞抗体筛选 | 手工或碎片化筛选栈 | Optoseeker + BioMap AI 智能体系统 | 更高通量闭环,带实时模型反馈 | 部署仍由合作伙伴主导 |
| 复杂生物制剂管线创建 | 传统串行研发 | MegaStream 一体化 AI 原生技术栈 | 宣称效率提升 500%+ | 仍偏前瞻 |
| 药企生物制剂发现 | 合作伙伴专属模型开发 | Sanofi + BioMap AI 模块 | 显示适配大型药企工作流 | 未公开详细模块清单 |
产品和合作伙伴材料对收益有方向性支撑,但往往缺少公开分母或 ROI 基线。
[CE001, CE003, CE015, CE016, CE017, CE018]BioMap 已公开的架构:从基础模型延伸到实验室和基础设施层。
[CE001, CE003, CE012, CE013, CE014]5.2 架构、工作流与依赖
BioMap 公开资料包描述的架构足够具体,已经超过泛泛营销框。BioMap OS 把知识聚合、预测与设计、智能实验、数据采集和模型训练整合成闭环反馈系统。资料包还描述了全球多云生物计算引擎、动态调度、硬件抽象层,以及自主诊断和自恢复;苏州高通量实验室提供物理验证能力。合作伙伴证据进一步说明,这不只是概念。Optoseeker 合作把 AI 模型、高通量单细胞筛选和实时反馈循环连起来;Kexing 与 MegaStream 公告则把同一模式延伸到药物开发和复杂生物药工作流。因此,这套栈不只依赖模型质量:算力可得性、合作伙伴数据、湿实验吞吐、筛选硬件和数据标准化都很重要。这让产品比裸模型 检查点更可防守,但也让运营更复杂,并更容易被实验室执行和基础设施瓶颈拖住。[CE013, CE014, CE015, CE016, CE017, CE018]
| 层级 / 流程 | 作用 | 依赖 | 风险 |
|---|---|---|---|
| 基础模型 | 核心生物推理与生成 | 训练数据 + 算力 | 模型主张可能跑在外部验证之前 |
| 知识聚合 | 上下文组装与检索 | 数据权利 / 质量 | 垃圾输入风险 |
| 预测 / 设计 | 候选物生成与优化 | 模型校准 | 在新模态上泛化不足 |
| 智能实验 | 跨硬件工作流编排 | 实验室集成 | 运营复杂性 |
| 数据采集 / 标准化 | 生成 AI 可用反馈数据 | 仪器化 + 标准 | 数据漂移 / 瓶颈 |
| 模型训练 / 微调 | 客户专属适配 | GPU + 数据可用性 | 算力成本和吞吐量 |
| 全球多云引擎 | 部署和可扩展性 | 云服务商 / 网络 | 可靠性和安全暴露 |
公开架构显示 BioMap 走的是全栈运营模型,但多层仍依赖 BioMap 自控基础设施。
[CE012, CE013, CE014, CE015, CE035]BioMap 栈如何把数据和假设推进到验证与迭代。
[CE001, CE012, CE013, CE016, CE018]决定 BioMap 闭环架构能否真正跑通的关键依赖。
[CE014, CE015, CE016, CE017, CE035]5.3 技术证明、成熟度与路线图
对一家私人公司而言,BioMap 拥有异常强的公开技术证明,但这些证明集中在研究资产,而不是企业运营指标。xTrimoPGLM 代码库和论文披露了模型家族、训练细节、示例推理流程和基准测试结果,包括 100B 规模、1T 训练 token,以及在蛋白基准测试上的强表现。PFMBench 又提供了一套覆盖 38 项任务和 17 个模型的公开基准测试。更广的 GitHub 组织显示,ProteinSage、scFoundation、xTrimoMultimer、抗体设计工具和评估框架仍有持续活动;ProteinReasoner、RNAGenesis 和 BioLab 等较新论文则显示,路线图正从蛋白语言模型走向推理、RNA 疗法和自主多智能体研究。这种广度是真实差异化,但也暴露了成熟度偏斜。BioMap 发表代码、基准测试和论文的地方,公开证明最强;买方真正想看的生产部署、可用性、实施负担,以及蛋白中心工作流之外的验证回报,证据更弱。[CE004, CE005, CE006, CE007, CE008, CE009]
| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 含义 | 来源 |
|---|---|---|---|---|
| 2024 | xTrimoPGLM 论文 / 仓库 | 已发布 | 核心蛋白模型家族已有公开技术细节 | arXiv + GitHub |
| 2024-2025 | xTrimo V4 / BioMap OS 公开资料包 | 活跃 | 商业叙事转向 268B 平台规模 | BioMap 资料包 + Yicai |
| Jun 2025 | PFMBench / InverseFoldingEvaluation | 已发布 | BioMap 向公众开放基准测试工具 | GitHub + arXiv |
| Jul 2025 | ProteinReasoner | 已发布 | 增加以推理为中心的蛋白设计能力 | bioRxiv |
| Jul 2025 | RNAGenesis | 已发布 | 将路线图延伸到 RNA 疗法 | bioRxiv |
| Oct 2025 | BioLab | 已发布 | 推向自主多智能体研究 | bioRxiv |
| 2026 | scFoundation / ProteinSage / xTrimoMultimer 活跃动态 | 活跃仓库 | 释放持续公开研究迭代信号 | GitHub 组织 |
代码和论文存在的地方,路线图成熟度最强;商业化成熟度则因模块而异。
[CE008, CE009, CE010, CE020, CE021, CE022]BioMap 公开证据显示哪些能力最成熟,哪些证明还偏早期。
[CE020, CE021, CE022, CE023, CE024, CE029]5.4 信任、安全与质量控制
BioMap 公开暴露得最清楚的信任证据偏信息安全。隐私政策片段称,BioMap 持有 ISO/IEC 27001:2022 认证,传输中的数据用 TLS 1.2 加密,静态数据用 AES-256 加密,并每年开展第三方渗透测试。同一政策还描述了网站可能通过表单和邮件工作流收集的个人信息类别。对共享敏感生物数据的企业买方来说,这些披露有意义,但不能回答所有尽调问题。审阅来源中没有公开状态页、可用性 档案、事故账本、临床级运营标准,或可与 GxP 要求高的场景买方要求相当的受监管实验室认证组合。因此,BioMap 的信任姿态在信息安全层看起来可信,但在受监管运营层仍不完整。BioMap 的差异化叙事依赖敏感数据、多云基础设施和物理实验,所以缺失的运营信任细节是真正的尽调缺口,而不是文档小瑕疵。[CE029, CE030, CE031, CE032, CE034]
| 控制项 / 指标 | 状态 | 范围 | 缺口 |
|---|---|---|---|
| ISO/IEC 27001:2022 | 称 2023 年取得 | 信息安全 | 需要证书范围和续期证据 |
| TLS 1.2 加密 | 称已具备 | 传输中数据 | 未公开更深入架构细节 |
| AES-256 加密 | 称已具备 | 静态数据 | 未公开密钥管理细节 |
| 年度第三方渗透测试 | 称已具备 | 安全保障 | 未公开供应商或摘要结果 |
| 隐私政策 | 已在政策文本中发布 | 网站个人数据 | 不等同于临床 / 受监管数据政策证明 |
| 运营状态 / 事故档案 | 未找到 | 可靠性透明度 | 未发现公开状态页或事故历史 |
BioMap 披露了具体信息安全控制,但面向重监管工作流,公开信任包仍比买方希望看到的更薄。
[CE029, CE030, CE031, CE032]5.5 要点展示
06客户
6.1 客户分层与采用广度
对一家私人 AI 生物技术公司而言,BioMap 的公开客户基础异常宽,但这种广度是分层披露的,不是一个干净 KPI。HKIC 和 HKSTP 描述了 200 多名签约 AIGP 用户,覆盖国际药企、CDMO、创新药开发商、合成生物学公司、绿色科技企业和科研机构。Yicai 与 PR Newswire 又把视角拉宽,引用 800 多家机构用户、30 多家领先企业和 60 多个已验证项目。合在一起,这些证据支持一个真实买方 / 用户图谱,横跨多个垂直领域和地理区域,而不是单一旗舰药企关系。它也说明,BioMap 当前触达了直接付费方和战略用户:企业 R&D 团队、研究组织和生态项目,但它们未必都会转化成同等收入。因此,核心客户问题不是 BioMap 有没有触达;它显然有。问题是其中有多少处在持久付费生产关系里,多少只是探索性、低强度或由伙伴中介的使用。另一个细微点是,公开指标混合了非常不同的账户单位,所以任何投资者若把 800 家机构解读成 800 个经济等价账户,都会实质性高估变现质量。因此,广度叙事首先应被视为分层和管线强度信号,其次才是收入质量信号。[CU001, CU002, CU003, CU004, CU005, CU006]
| 细分市场 | 买方 / 用户 / 付款方 | 使用场景 | 规模信号 | 收入 / 战略价值 | 缺口 |
|---|---|---|---|---|---|
| 跨国药企 | 研发平台团队 | 生物药发现 | Sanofi 具名;Lilly 有相邻线索 | 战略价值高 | 需要续约 / 收入证据 |
| 中国生物制药企业 | 药物开发团队 | 大分子 / 抗体项目 | Kexing 具名;CSPC 为声称客户 | 潜在价值较高 | 需要客户侧案例 |
| CDMO / 创新药开发商 | 平台用户 | 蛋白设计与优化 | 包含在 HKIC 用户组合中 | 管线价值覆盖面广 | 未公开具名部署 |
| 研究机构 | 科研用户 | 药物发现与模型使用 | 包含在 HKIC 组合中;属于 800 家机构之一 | 覆盖面 / 数据飞轮 | 付费状态不清楚 |
| 合成生物学 / 绿色科技 | 应用生物学用户 | 设计与流程优化 | HKIC 组合;LanzaTech / Syngenta 相邻线索 | 支撑 TAM 宽度 | 部署证据弱 |
| 仪器合作伙伴 | 工作流运营方 | 单细胞筛选与抗体工作流 | Optoseeker 具名 | BioMap 延伸到湿实验室边缘 | 经济条款未公开 |
BioMap 的客户分层异常宽,但各细分的收入权重仍未披露。
[CU001, CU002, CU012, CU013, CU029]| 指标 | 数值 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| 已签约 AIGP 用户 | 200+ | Jun 2024 | HKIC / HKSTP | 高 | 说明已有付费或签约核心用户 | 无 ACV 或续约率 |
| 机构用户 | 800+ | 2026 | Yicai / PR Newswire / BioMap 资料包 | 高 | 说明触达面广 | 活跃用户与历史用户占比未知 |
| 头部企业 | 30+ | 2026 | Yicai | 中 | 已有具名企业层 | 无行业拆分或单账户支出 |
| 已验证项目 | 60+ | 2026 | Yicai / BioMap 资料包 | 中 | 指向重复使用 | 付费生产化转化未知 |
| 有试用兴趣的客户 | 多个 | 2025 | Pharmcube | 中 | 暗示具名客户之外还有管线 | 无数量或转化数据 |
| 平台迁移 | 旧平台退役;用户迁至 AIGP | 2024 | scFoundation API 示例 | 中 | 暗示存量用户群仍活跃 | 未披露用户留存 |
这些公开指标混合了合同、机构、项目和迁移信号,因此更能说明采用面宽, 而不是持久变现。
[CU002, CU003, CU004, CU005, CU014, CU015]典型 BioMap 关系看起来从发现开始,再进入更深的工作流集成。
[CU002, CU007, CU010, CU014, CU023, CU024]公开证明从广泛机构覆盖,收窄到少量可由外部核验的具名战略部署。
这是证明质量漏斗,不是转化漏斗;它展示公开覆盖面如何收敛成小得多的一组具名、可核验部署案例。
[CU002, CU003, CU004, CU005, CU016]6.2 具名客户证明与部署质量
具名证明存在,但不均衡。Sanofi 是最高质量证明,因为多个独立来源描述了具体生物药发现合作,并给出明确交易经济条款。Harbour BioMed 次之,因为 MegaStream 让 BioMap 的平台角色成为复杂生物药建设的核心。Kexing 与 Optoseeker 把证明延伸到中国生物医药和与仪器联动的抗体工作流,说明 BioMap 不局限于一个跨国伙伴。但当公开来源从明确合作转向类似客户标识 或纯媒体客户名单时,证明质量立刻下降。Sohu 提到 CSPC 和 Dabeinong 有方向性价值,但缺少同等客户侧佐证。同样,来自 Lilly、Syngenta 和 LanzaTech 的生态接近性,更多说明分部相关性,而不是主动付费部署。结果是,客户证明栈的顶层可信且有商业意义,但下层仍需要客户侧案例、续约证明和结果指标,才能被视为强持久性证据。这一点很重要,因为 BioMap 的客户标识组合在完全支撑一个可重复、独立验证、经续约检验的商业质量叙事之前,已经足以支撑战略叙事。[CU007, CU008, CU009, CU010, CU011, CU012]
| 客户 / 交易对手 | 细分 | 部署 / 使用场景 | 生产化 vs 试点 | 结果 | 限制 |
|---|---|---|---|---|---|
| Sanofi | 全球药企 | 生物药发现 AI 模块 | 类生产合作 | 经济条款已披露,是质量最高的具名证据 | 未公开续约或结果 KPI |
| Harbour BioMed / MegaStream | 全球生物药公司 | AI 原生复杂生物药平台建设 | 战略建设 / 扩张 | 有力证明深度工作流信任 | 前瞻性强,带有风投属性 |
| Kexing Biopharm | 中国生物制药企业 | 肿瘤 / 自身免疫疾病大分子药物 | 合作 | 说明国内生物科技公司开始采用 | 财务 / 部署深度未披露 |
| Optoseeker | 生命科学仪器 | AI 增强的高通量抗体发现 | 试点到部署路径 | 说明湿实验室工作流已接入,且有试用兴趣 | 合同规模未公开 |
| CSPC / Dabeinong 相关声称 | 制药 / 农业生物 | 出现在媒体客户名单中 | 不清楚 | 可作为尽调线索 | 客户侧佐证弱 |
独立来源描述具体工作流时,具名证据质量最强;媒体只是列出客户名称时,质量最弱。
[CU007, CU008, CU009, CU010, CU011, CU018]BioMap 在具名工作流且有外部来源佐证时,证明质量最高。
[CU007, CU008, CU009, CU010, CU011, CU017]6.3 持续性、扩张与集中度风险
持续性是公开客户记录变弱的地方。公司没有披露 NRR、GRR、流失率、合同期限、头部客户集中度或客户满意度包。这不意味着客户质量差;它意味着外部记录不够细,无法紧密承销。公开层面,持续性只能从反复平台使用、向 AIGP 的平台迁移、持续合作措辞,以及苏州实验部署或 MegaStream 共创这类更深工作流集成来推断。扩张逻辑比留存逻辑更容易看见:BioMap 可从探索性使用走向签约 AIGP 访问,再进入定制工作流、湿实验服务,最终进入资产级合作。集中度风险正是这个机会的镜像。由于最强具名证明集中在少数标杆伙伴,经济性可能比广泛机构用户数字暗示的更集中。采购摩擦也不小,可能涉及定价不透明、数据治理审查和集成投入。另一个实际问题是可被引用程度:公开证据仍很少显示 BioMap 客户是否愿意公开谈论结果、实施周期或续约决定。客户视角对广度是正面的,对持续性则是中等信心。换句话说,BioMap 看起来已足够商业相关,值得尽调;但透明度还不足以支撑高信心客户质量承销。[CU020, CU021, CU022, CU023, CU024, CU025]
| 指标 | 数值 / null | 细分 | 置信度 | 尽调要求 |
|---|---|---|---|---|
| 净收入留存率 | null | 公司整体 | 低 | 要求按客户细分提供 NRR |
| 总收入留存率 | null | 公司整体 | 低 | 要求提供 GRR 和客户 logo 留存历史 |
| 流失率 | null | 公司整体 | 低 | 要求提供 logo 流失和项目流失 |
| 合同期限 | null | 企业账户 | 低 | 要求提供标准期限和续约结构 |
| 重复项目信号 | 60+ 个已验证项目 | 平台用户 | 中 | 区分重复付费项目与一次性试点 |
| 平台迁移连续性 | 用户从旧界面迁至 AIGP | AIGP 用户 | 中 | 量化迁移后的活跃用户和留存 |
公开耐久性证据基于代理信号,而不是直接指标;这里的 null 反映真实披露缺口。
[CU020, CU021, CU034, CU035]| 扩张驱动 / 风险 | 证据 | 影响 | 尽调路径 |
|---|---|---|---|
| 从平台走向定制工作流 | Optoseeker / Sanofi / Harbour 集成 | 若成功,可提高钱包份额 | 要求按账户提供增购路径 |
| 从平台走向资产创建 | MegaStream 架构 | 可能提高单客户战略价值 | 要求说明经济权益归属和管线治理 |
| 跨垂直扩张 | HKIC 组合 + 绿色科技相邻线索 | 将 TAM 扩展到制药之外 | 要求提供细分收入拆分 |
| 标杆客户集中 | 具名证据集中在少数 logo 上 | 可能掩盖收入集中度 | 要求提供前 10 大客户收入结构 |
| 定价不透明 | 未公开合同金额 | 采购和续约风险更难判断 | 要求提供匿名版条款清单 |
| 数据 / 隐私审查摩擦 | 敏感生物数据与企业数据共享 | 可能拖慢销售周期 | 要求提供安全问卷输赢数据 |
公开覆盖面很宽,但经济集中度和采购摩擦仍大多未量化。
[CU022, CU023, CU024, CU025, CU026, CU027]用紧凑视图看当前公开记录中覆盖广度与留存质量的对比。
[CU013, CU020, CU022, CU025, CU026, CU035]6.4 要点展示
07风险
7.1 监管、法律与治理风险
BioMap 的监管和法律风险,根源在于它不只是发表论文;公司还在处理客户关系,横跨北京、苏州和香港运营,并明确把香港的数据与 IP 姿态用作跨国信任叙事的一部分。PDPO 表明香港是严肃隐私监管体系;BioMap 自身政策文本显示,公司会通过网页表单和邮件工作流收集可识别个人与组织数据。即使还没进入更敏感的科学数据共享问题,这已经带来真实合规负担。更大的治理风险是披露。BioMap 保密递表赴港 IPO,意味着投资者仍拿不到招股书级别的结构、控制、募资用途或集中度细节。公开媒体和数据库对融资历史的描述方向一致,但并未完全调和;AInvest 还围绕治理和估值风险给出明确负面框架。这些都不能证明存在不当行为;但确实说明,法律和治理表面比公司雄心规模理应匹配的厚度更薄。对一家向跨国科学买方出售信任的公司而言,薄披露本身就是运营风险,因为它拖慢尽调,让边缘法律问题更难关闭,也让外部人只能依赖叙事碎片,而不是完整控制证据。[CR001, CR002, CR003, CR004, CR005, CR006]
| 规则 / 问题 | 司法辖区 | 状态 | 可能性 | 严重性 | 缓解措施 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|---|---|
| 隐私与个人数据处理 | 香港 / Web 运营 | 政策已发布;义务已生效 | 中 | 高 | 政策、ISO 27001、加密、渗透测试 | 中 | 要求提供 DPA、跨境传输图和处理方清单 |
| 跨境信任 / 数据治理缺口 | 中国 / 香港 / 全球客户 | 叙事可见;细节有限 | 中 | 高 | 香港定位和安全控制 | 中-高 | 要求提供客户数据流图和区域控制措施 |
| 保密 IPO 披露缺口 | 香港资本市场 | 招股书未公开 | 高 | 高 | 管理层可在文件公开后补足 | 高 | 要求提供招股书草稿或 IPO 尽调资料室权限 |
| 融资历史对账缺口 | 公司治理 | 媒体与数据库不一致 | 中 | 中 | IPO 清理可能完成对账 | 中 | 要求提供股权结构表和逐轮融资明细 |
| 合作中的 IP / 数据权利不清 | 多方研发 | 未公开细节 | 中 | 高 | 仅有香港 IP 叙事 | 中-高 | 要求提供合作所有权和模型权利条款 |
当前法律风险更多来自披露不完整和数据治理可见度不足,而不是已知执法事件。
[CR001, CR002, CR003, CR006, CR007, CR035]最高风险集中在运营复杂、公开披露有限的交汇处。
[CR003, CR006, CR010, CR012, CR018, CR022]7.2 运营、技术与依赖风险
BioMap 的运营模型远比纯软件公司复杂。公开技术栈把基础模型、实时 AIGP 界面、多云部署、高通量实验、苏州实验室、硬件联动工作流和伙伴特定集成结合起来。这种复杂性是护城河主张的一部分,也是失败模式的集中点。苏州实验室和干湿反馈循环显然位于产品交付核心;如果吞吐、可重复性、数据质量或实验室运营滑坡,价值会很快被侵蚀。同样,缺少公开状态页或事故档案,意味着外部观察者几乎没有可靠性直接证据。技术侧,xTrimoPGLM 与 PFMBench 等公开基准测试显示了真实科学能力,但无法解决从基准测试领先转化为药物开发结果或买方 ROI 的问题。最后,合作伙伴依赖不是可选项,而是结构性因素:Harbour 贡献数据集和开发能力,Sanofi 锚定旗舰证明,Optoseeker 这类集成把技术栈系在第三方仪器和工作流互操作性上。BioMap 越靠深度工作流集成取胜,任何一个依赖方的失败越可能沿着科学、交付时点和客户信心扩散,而不是被隔离在单个软件模块里。[CR009, CR010, CR011, CR012, CR013, CR014]
| 失效模式 | 可能性 | 严重性 | 缓解成熟度 | 剩余敞口 | 未解决缺口 |
|---|---|---|---|---|---|
| 湿实验室吞吐量或验证瓶颈 | 中 | 高 | 中 | 高 | 无公开吞吐量 / 利用率指标 |
| 多云或基础设施可靠性故障 | 中 | 高 | 低-中 | 高 | 无公开状态或事故历史 |
| 模型基准无法转化为客户 ROI | 高 | 高 | 低-中 | 高 | 公开生产结果证据有限 |
| 数据质量 / 反馈回路漂移 | 中 | 高 | 中 | 中-高 | 未按工作流公开 QA 流程细节 |
| 安全控制与受监管买家要求不匹配 | 中 | 中-高 | 中 | 中 | 信息安全证据强于受监管工作流证据 |
| 路线图横跨蛋白、RNA 和智能体,范围蔓延 | 中 | 中 | 低-中 | 中 | 未公开资源配置拆分 |
运营风险偏高,因为 BioMap 的护城河逻辑本身比标准模型 API 依赖更多活动环节。
[CR004, CR005, CR010, CR011, CR012, CR013]| 依赖 | 交易对手 | 角色 | 集中度 | 失败情景 | 严重性 | 缓解措施 | 剩余敞口 |
|---|---|---|---|---|---|---|---|
| 旗舰药企证据 | Sanofi | 验证企业药物发现适配性 | 高 | 合作停滞或无法扩张 | 高 | 扩大具名证据 | 中-高 |
| 复杂生物药栈 | Harbour BioMed / MegaStream | 数据集、抗体平台、开发能力 | 高 | 联合平台交付不及预期 | 高 | 系统化拓展伙伴与治理 | 高 |
| 筛选硬件工作流 | Optoseeker | 单细胞功能筛选集成 | 中 | 集成延迟或硬件不匹配 | 中 | 支持多条硬件路径 | 中 |
| 国内生物医药工作流 | Kexing | 大分子药物开发合作 | 中 | 试点未转化为持续使用 | 中 | 补强证据和案例 | 中 |
| 政策 / 生态支持 | HKIC / HKSTP | 香港信任背书、资源聚合、本地地位 | 中 | 支持转弱或预期抬高 | 中 | 拓宽政策与伙伴基础 | 中 |
| 算力和数据栈 | 云 + 数据提供商 | 支撑模型训练和推理 | 高 | 成本或访问冲击拖累性能 | 高 | 优化利用率,分散技术栈 | 高 |
依赖风险集中在 BioMap 声称拥有战略差异化的同一批环节。
[CR015, CR016, CR017, CR018, CR021, CR028]BioMap 的主要风险如何层层传导到收入质量、利润率、融资与估值。
[CR003, CR012, CR018, CR022, CR023, CR036]关键依赖包括政策支持、旗舰合作伙伴、算力以及实验室执行。
[CR012, CR015, CR017, CR018, CR021, CR028]7.3 财务、执行与缓释视角
财务和执行风险直接来自 BioMap 的混合模型。公司同时在建模型、采数据、跑实验并推进大型企业合作,资本强度自然高于纯软件业务。公开来源仍指向公司在保密 IPO 进程中有明显融资需求,而现金、现金消耗和现金跑道仍未披露。客户广度有帮助,但还不足以中和集中度风险,因为最强具名证明仍聚集在少数高知名度伙伴周围,更深客户持续性指标也缺席。好消息是,一些缓释因素可见。HKIC 和 HKSTP 支持降低生态与政策风险,公开伙伴多样性减少单一分部依赖,开放科学足迹降低黑箱技术不透明。即便如此,缓释成熟度并不均衡:BioMap 在安全控制和生态支持上的公开证据强于运营 KPI、客户续约或融资韧性。实际风险评级应保持较高,直到管理层能证明重复付费生产使用、纪律性的资本可见度和干净的 IPO 级披露。实质上,BioMap 现在处在尴尬但常见的区间:战略动能跑在证明质量前面;证据足够支持严肃尽调,但还不足以支持轻松放心。[CR019, CR020, CR022, CR023, CR024, CR025]
| 角色 / 职能 | 依赖或缺口 | 可能性 | 严重性 | 缓释措施 | 尽调路径 |
|---|---|---|---|---|---|
| 领导层 / 治理纪律 | 私营公司披露仍然单薄 | 中 | 高 | IPO 流程可能改善纪律 | 审阅董事会和报告流程 |
| 商业化梯队 | 研究证据强于公开的销售质量证据 | 中 | 高 | 扩充客户案例包 | 访谈 GTM 负责人和头部客户 |
| 科学广度管理 | 蛋白、RNA、智能体和平台都会争夺注意力 | 中 | 中高 | 给路线图排序 | 要求资源分配计划 |
| 实施 / 支持能力 | 用户迁移和定制工作流支持可能挤压运营 | 中 | 中 | AIGP 迁移和服务团队 | 要求支持 SLA 和待办数据 |
| 人才留存 / 招聘 | 复杂技术栈需要跨学科人才 | 中 | 中 | 香港 / 中国生态支持 | 审阅组织架构和流失数据 |
执行风险的核心,是能否把出色科学转成可重复的企业交付,并建立有纪律的披露。
[CR020, CR025, CR026, CR027, CR038, CR039]| 风险 | 可监测触发点 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 融资风险 | IPO 进度 / 资本能见度 | 申报停滞,或没有明确的资金桥 | 暂停或重估尽调 |
| 伙伴集中度 | Sanofi / Harbour 扩张证据 | 项目未扩张,或被明显弱化 | 重新评估客户质量假设 |
| 数据治理风险 | 客户安全 / 法务尽调 | 数据条款造成实质摩擦或企业成交延后 | 升级合规审查 |
| 运营可靠性 | 实验室 / 平台支持证据 | 出现实施延迟或验证闭环失败 | 下调对护城河的信心 |
| 聚焦被稀释 | 路线图广度对照执行指标 | 新研究方向过多,却缺少客户证据 | 给创新溢价打折 |
| 披露质量 | 招股书 / 尽调资料室完整度 | 管理层无法勾稽融资、客户或烧钱速度 | 视为击穿投资逻辑 |
终止标准强调可衡量证据,用来区分可管理的复杂度和真正的执行失败。
[CR022, CR024, CR036, CR037, CR039, CR040]7.4 要点展示
08估值
8.1 投资论点与反论点
多头论点是,BioMap 正在 AI 基础设施与真实世界生物学执行的交汇处,打造一种罕见战略资产。它把基础模型深度、干湿闭环发现操作系统、香港战略支持、广泛机构触达,以及来自 Sanofi 和 Harbour 的蓝筹伙伴证明结合在一起。如果即将到来的申报文件显示,这些信号能转化为经常性企业收入、体面的毛利率和可重复客户扩张,BioMap 可能理应相对成熟软件服务可比公司享有溢价,因为公开市场很少能买到广度如此高的纯 AI 生物学平台敞口。反论点是,BioMap 目前仍比“可承销”更容易让人欣赏。公开证据重在里程碑、用户、项目和伙伴叙事,轻在已确认收入、留存、客户集中度和利润率质量。按这种解读,BioMap 可能是一家里程碑重、服务重、资本密集的混合体,战略叙事跑在公开证明之前;一旦招股书经济性到来,激进估值标记容易失望。[CV001, CV002, CV003, CV004, CV005, CV006]
| 维度 | 多头逻辑 | 空头反逻辑 |
|---|---|---|
| 市场 | AI 生物学平台,带来组合里稀缺的赛道敞口 | 赛道热度跑在已验证变现前面 |
| 产品 | xTrimo + BioMap OS + 干湿闭环形成差异化 | 复杂技术栈可能服务占比高,难以规模化 |
| 客户 | Sanofi / Harbour / 800+ 机构触达 | 广度未必等于持久收入 |
| 财务 | 战略股东和伙伴付费意愿 | 没有经审计的收入、利润率或留存证据 |
| 竞争 | 私有战略资产稀缺 | 公开可比公司显示,透明平台已经可投 |
| 治理 | IPO 可能释放验证和披露 | 保密申报保留太多不确定性 |
| 估值 | 若经济性强,招股书可能打开上行 | 当前叙事可能把估值打高、低估隐性风险 |
同一组证据既能撑起溢价故事,也能撑起折价故事,关键看 IPO 文件最终披露什么。
[CV001, CV002, CV003, CV004, CV005, CV006]战略质量与缺失的分母如何共同导向跟踪建议。
该流程为定性呈现,展示决策逻辑,而非加权评分模型。
[CV001, CV002, CV007, CV009]8.2 投资建议、置信度与立场
我们对 BioMap 的建议是观察,置信度中等,风险评级高,估值立场偏高;总分约 5.8/10。BioMap 的战略价值太强, 不能轻易忽略:科学技术栈、合作伙伴组合和市场位置都值得持续跟踪、认真准备尽调。但它的财务信息也太不透明, 不能只靠叙事就支付溢价。这里的投资建议明确受价格影响。若最终监管文件或尽调资料室能证明平台收入具备经常性、 利润率结构可接受、老客户持续扩张,判断可以明显上调。反过来,如果披露出来的是项目制经济、高集中度, 或估值诉求已经高过公开证据集能支撑的水平,正确动作就是守住纪律。现有证据支持关注和准备, 不支持激进承诺。[CV007, CV008, CV009, CV010, CV038, CV039]
| 维度 | 评估 | 依据 |
|---|---|---|
| 建议 | 持续跟踪 | 战略质量真实;承销依据仍不完整 |
| 信心 | 中 | 未披露经审计的收入、利润率或留存 |
| 风险评级 | 高 | 混合经济模型、集中度和融资不确定性 |
| 估值立场 | 偏高 | 公开证据尚不足以稳妥支撑溢价 |
| 整体评分 | 5.8 / 10 | 资产有意思,但当下分母纪律弱 |
| 入场纪律 | 等待申报文件或更深入尽调 | 需要分母、利润率和集中度证据 |
该建议有意锚定证据和价格,而不是给一个泛泛的质量分。
[CV007, CV008, CV009, CV010]围绕当前跟踪判断的主要维度,给出可供投委会使用的评分。
[CV001, CV007, CV019, CV031, CV039, CV040]8.3 融资背景与入场纪律
BioMap 的融资背景给投资逻辑提供了方向性支撑,但信息仍不完整。公开资料能支撑 $100 million Series A、累计融资超过 $200 million、HKIC 战略背书,以及数亿美元 IPO 目标。Sanofi 带来 $10 million 首付款和超过 $1 billion 的里程碑选择权;Harbour/MegaStream 还可能带来首付款、里程碑和版税。这些事实重要, 因为它们说明成熟伙伴和投资人已经给平台赋予战略价值。但它们没有回答后期承销最关键的问题:市场究竟在为多大收入基数付钱, 经济条款中有多少更像里程碑而不是经常性收入,稀释和优先股堆叠长什么样,以及如果上市窗口转弱, 公司还需要多少资本。因此,入场纪律应围绕分母和条款展开,而不是围绕标题叙事。[CV011, CV012, CV013, CV014, CV027]
8.4 乐观、基准与悲观情景
我们的基准情景假设 BioMap 确实是有战略价值的平台,但仍是混合型平台;一旦披露经济数据, 它更可能落在上市 TechBio 和软件工具可比公司的区间,而不是前沿 AI 热度估值。按这个判断, 约 $1.5-2.5 billion 的公允价值区间站得住。乐观情景假设监管文件显示平台收入经常性强、 企业客户重复使用、集中度可控、利润率路径可信;届时约 $3-5 billion 的估值就有可能成立。 悲观情景假设公开经济数据低于预期,业务比想象中更偏项目和里程碑,或合作伙伴集中度过高; 届时约 $0.7-1.2 billion 的估值,或融资路径推迟,都有可能。区间很宽,因为 BioMap 的质量信号是真实的, 但这些信号如何转成财务结果,大部分仍藏在水面下。[CV015, CV016, CV017, CV018, CV035, CV036]
| 情景 | 概率信号 | 关键假设 | 价值区间 | 何事会改变判断 |
|---|---|---|---|---|
| 乐观 | ~25% | 平台收入可重复、利润率可接受、企业客户持续扩张、申报材料强 | 3-5B USD | 招股书证明收入质量高,集中度可控 |
| 基准 | ~45% | 平台具备战略价值,但混合模式下可重复收入能见度只露出一部分 | 1.5-2.5B USD | 公开可比公司和披露仍是主要锚 |
| 悲观 | ~30% | 项目制经济性、客户集中、利润率弱或 IPO 延迟 | 0.7-1.2B USD | 申报材料不及预期或资本路径转弱 |
概率只是定性信号,不是组合权重;区间很宽,反映财务分母缺失。
[CV015, CV016, CV017, CV018, CV035, CV036]基于公开证据的 BioMap 宽口径情景区间,单位为十亿美元。
情景区间不是预测;它们是承销判断区间,锚定公开可比公司和 BioMap 特有的战略期权价值。
[CV019, CV035, CV036, CV037]8.5 可比公司组
上市可比公司能提供有用但不完美的估值锚。截至 2026 年 7 月,Recursion、Absci、Schrödinger 和 Certara 的市值大致都在 $1.0-1.75 billion 区间,但收入基数和隐含倍数差异很大。Recursion 约 $1.75 billion 的估值对应约 $65.73 million 收入,隐含约 26.6x 销售额;Absci 约 $1.72 billion 的估值对应约 $2.8 million 收入,体现极端期权价值定价;Schrödinger 约 $1.22 billion 的估值、 约 $250 million 收入和 ~4.79x 销售额,是更均衡的软件科学参照;Certara 约 $1.05 billion 的估值、 约 $410 million 收入和 ~2.54x 销售额,则是低倍数软件服务锚。这组可比公司不能给 BioMap 一个唯一答案, 但能说明市场对透明度、收入质量和平台可选性的定价差异很大。在 BioMap 没有公开分母之前, 使用精确的表观倍数只是伪精确。[CV019, CV020, CV021, CV022, CV023, CV024]
| 可比对象 | 指标 | 估值 / 状态 | 为何可比 | 局限 |
|---|---|---|---|---|
| Recursion | 基于约 $65.73M 收入,隐含收入倍数约 26.6x | 约 $1.75B 市值 | 公开市场 TechBio,带平台期权 | 临床 / 管线画像不同 |
| Absci | 基于约 $2.8M 收入,隐含倍数极高 | 约 $1.72B 市值 | 体现市场愿意为 AI 生物药叙事付期权价值 | 收入分母很小 |
| Schrödinger | 基于约 $250M 收入,收入倍数约 4.79x | 约 $1.22B 市值 | 最接近的软件科学公开市场锚 | 物理软件组合与 BioMap 不同 |
| Certara | 基于约 $410M 收入,收入倍数约 2.54x | 约 $1.05B 市值 | 风险更低的软件服务锚 | 前沿模型期权更少 |
| Sanofi 交易 | >$1B 里程碑付款 + $10M 首付款 | 战略验证点 | 显示伙伴愿意付费 | 不是公司估值 |
| MegaStream 经济性 | 首付款 + 里程碑付款 + 版税 | 期权价值参照 | 捕捉资产上行逻辑 | 未披露入账价值 |
可比对象有意混合:用公开市场锚保持纪律,用战略交易衡量期权价值。
[CV019, CV020, CV021, CV022, CV023, CV024]不同收入与倍数假设下的示意估值结果,单位为十亿美元。
敏感性仅作示意,因为 BioMap 尚未公开披露清晰的收入分母;它展示的是投资者估算收入与质量时,价值会多快变化。
[CV025, CV026, CV035, CV036]8.6 退出准备度与最终尽调
最强的可行退出路径仍是计划中的 HKEX 上市;如果 BioMap 平台对更大的生物制药或 TechBio 参与者变得不可替代, 战略收购是次一级选项。但两条路都要求证据质量快速提高。打破投资逻辑的触发条件很简单: 监管文件无法证明收入经常性;集中度高;利润率和烧钱速度不好看;或估值诉求显著高于证据集能支撑的水平。 最高优先级的尽调问题同样清楚:经审计收入、收入结构、毛利率、集中度、现金跑道、优先权结构和客户续约证据。 在这些问题得到回答前,可提交 IC 的正确姿态是观察并有纪律地准备,而不是提前押注式兴奋。[CV029, CV030, CV031, CV032, CV039, CV040]
| 触发因素 | 阈值 / 事件 | 对投资逻辑的传导 | 行动含义 |
|---|---|---|---|
| 招股书分母偏弱 | 收入 / 利润率 / 留存显著不及预期 | 削弱平台溢价 | 放弃或大幅重估 |
| 客户集中度高 | 头部客户主导经济性 | 客户广度含金量下降 | 要求集中度折价 |
| IPO 延迟且融资不清晰 | 上市窗口后移且没有过桥资金 | 抬高融资风险 | 仅保留观察名单 |
| 项目制占比高 | 可重复平台收入低于预期 | 叙事变成混合服务故事 | 下调倍数锚 |
| 数据治理摩擦 | 企业销售卡在法务 / 安全审查 | 拖慢扩张并推高销售成本 | 下调情景权重 |
| 溢价诉求高于证据 | 估值跑在已披露证据前面 | 压缩预期回报 | 不追价 |
终止触发因素聚焦会击穿战略平台承销逻辑的证据。
[CV030, CV031, CV034, CV037, CV040]| 主题 | 缺失证据 | 重要性 | 负责人 / 路径 |
|---|---|---|---|
| 经审计收入 / ARR | 缺少可勾稽的公开分母 | 决定倍数基数 | 公司 / 申报文件 |
| 收入结构 | 平台、服务与里程碑收入占比未知 | 决定持久性和质量 | 公司 / 财务 |
| 毛利率路径 | 未公开披露毛利率 | 区分软件型经济性与混合经济性 | 公司 / 财务 |
| 客户集中度 | 头部客户结构未披露 | 关键下行驱动因素 | 公司 / RevOps |
| 烧钱速度与资金续航 | 缺少公开现金流能见度 | 决定融资风险 | 公司 / 财务 |
| 优先权 / 稀释结构 | 条款细节未公开 | 会实质改变回报结果 | 公司 / 法务 |
| 续约与留存 | 没有 NRR / 流失数据 | 检验客户持久性 | 公司 / 销售 |
这些是把“跟踪”升级为溢价投资观点前的门槛尽调项。
[CV013, CV032, CV039]8.7 展示材料
免责声明
本报告仅供参考,基于截至 2026-07-14 的公开来源,并非投资建议。公开财务与估值证据仍不完整,任何决策前都应独立核验。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | BioMap was formally established on September 25, 2020. | 中 | SO002, SO003 |
| CO002 | Robin Li served as BioMap’s lead initiator and public founder-chairman. | 中 | SO001, SO003 |
| CO003 | Wei Liu is BioMap’s co-founder and current chief executive officer. | 中 | SO001, SO002, SO004 |
| CO004 | BioMap describes itself as a global pioneer in AI foundation models for life sciences. | 中 | SO001 |
| CO005 | BioMap OS is a dry-wet closed-loop life-science discovery system that combines knowledge aggregation, predictive design, experimental control, and model training. | 中 | SO001, SO009, SO014 |
| CO006 | BioMap publicly lists offices in Beijing, Suzhou, Hong Kong, and Silicon Valley. | 中 | SO001 |
| CO007 | Independent profiles and Chinese reporting frame BioMap as Beijing-based and Chinese in origin. | 中 | SO002, SO009 |
| CO008 | HKSTP’s June 2024 release described BioMap as headquartered in the US. | 中 | SO006 |
| CO009 | BioMap’s public executive roster includes Xiaoming Zhang, Xiaoyue Sun, Ziyao Xu, Patrick Zhang, and Stan Z. Li in addition to the founders. | 中 | SO001 |
| CO010 | Wei Liu previously served as CEO of Baidu Ventures and as a Baidu group vice president. | 中 | SO001, SO003 |
| CO011 | Stan Z. Li is publicly positioned as BioMap’s chief scientist for AI large models and a Westlake University chair professor. | 中 | SO001 |
| CO012 | BioMap announced a $100 million Series A on July 30, 2021. | 中 | SO007, SO008 |
| CO013 | GGV Capital led the Series A, with Baidu, Legend Capital, BlueRun Ventures, Zhenzhi Capital, Xiang He Capital, and additional Robin Li participation also reported. | 中 | SO007, SO008 |
| CO014 | Public Series A coverage said the funds would be used primarily for R&D and talent recruitment. | 中 | SO007, SO008 |
| CO015 | HKIC signed a strategic partnership with BioMap in June 2024 and publicly stated it had invested in the company. | 高 | SO004, SO005 |
| CO016 | BioMap launched its first international innovation hub in Hong Kong in 2024 to expand global operations. | 高 | SO004, SO005 |
| CO017 | BioMap BioX aims to support more than 50 early-stage life-science R&D projects in Hong Kong over five years. | 中 | SO004 |
| CO018 | HKSTP said BioMap is one of the second batch of OASES strategic enterprises setting up operations in Hong Kong. | 高 | SO005, SO006 |
| CO019 | HKSTP said BioMap was collaborating with over 10 commercial partners and more than 200 academic institutions by June 2024. | 中 | SO005 |
| CO020 | HKIC said BioMap had secured contracts with over 200 AIGP users by June 2024. | 中 | SO004 |
| CO021 | HKIC said the Hong Kong summit included senior executives from Sanofi, Lilly China, LanzaTech, and Syngenta. | 中 | SO004 |
| CO022 | BioMap and Sanofi announced a strategic collaboration in October 2023 to co-develop AI modules for biotherapeutic drug discovery. | 中 | SO012, SO013 |
| CO023 | The Sanofi collaboration included a $10 million upfront payment and milestone potential above $1 billion. | 中 | SO012, SO013 |
| CO024 | Public coverage said the Sanofi work would target biologic discovery in areas including immunology, neurology, oncology, and rare diseases. | 中 | SO012, SO013 |
| CO025 | Yicai, Sohu, and Tencent reported that BioMap confidentially filed for a Hong Kong IPO in March 2026 to raise several hundred million US dollars. | 中 | SO009, SO010, SO011 |
| CO026 | Sohu and Tencent said the reported IPO process involved CICC, Morgan Stanley, and UBS as advisers. | 中 | SO010, SO011 |
| CO027 | Yicai and Tencent both referenced Liu Wei’s June 2024 statement that BioMap planned to seek a Hong Kong IPO within roughly 18 months. | 中 | SO009, SO011 |
| CO028 | Yicai reported that Robin Li and related Baidu entities had exited BioMap’s direct domestic shareholder list by September 2025 as part of pre-IPO equity optimization. | 中 | SO009 |
| CO029 | Yicai reported that Robin Li and related entities held about 40 percent of BioMap in its early stages. | 中 | SO009 |
| CO030 | BioMap’s current website describes xTrimo V4 as a 268-billion-parameter life-science foundation model. | 中 | SO001, SO014 |
| CO031 | BioMap claims xTrimo V4 has delivered 300-plus state-of-the-art results across more than 20 life-science fields. | 中 | SO001, SO014 |
| CO032 | HKIC’s June 2024 profile described xTrimo as a 100Bn-plus biology foundation model, indicating that BioMap scaled model size materially after the earlier Hong Kong launch. | 中 | SO001, SO004 |
| CO033 | Yicai said BioMap OS had been validated in more than 60 projects by March 2026. | 中 | SO009 |
| CO034 | Yicai reported that BioMap had served more than 800 global institutions and over 30 leading enterprises by March 2026. | 中 | SO009 |
| CO035 | Harbour BioMed’s June 2026 release said BioMap currently supports over 800 institutional users worldwide. | 中 | SO014 |
| CO036 | Harbour BioMed and BioMap launched MegaStream TechBio in June 2026 as an AI-native complex biologics venture. | 中 | SO014 |
| CO037 | Harbour BioMed and BioMap projected that the MegaStream dry-wet laboratory could deliver more than 500 percent efficiency gains, over 10-fold data accumulation gains, and more than 5 petabytes of data in five years. | 中 | SO014 |
| CO038 | BioMap’s official 2025 news archive publicly highlighted launches or disclosures for a generative discovery system, RNAGenesis, ProteinReasoner, PFMBench, and BioLab. | 中 | SO001 |
| CO039 | The BioLab preprint describes an end-to-end autonomous life-science research system built around multi-agent orchestration and biological foundation models. | 中 | SO021 |
| CO040 | RNAGenesis and ProteinReasoner show that BioMap’s model portfolio is expanding beyond generic protein language modeling into RNA therapeutics and multimodal reasoning. | 中 | SO019, SO020 |
| CO041 | An external March 2026 market commentary argued that BioMap’s confidential IPO route delayed public validation of valuation and governance questions. | 低 | SO026 |
| CM001 | The narrowest public market category that matches BioMap is AI-driven drug discovery platforms rather than total pharma R&D. | 中 | SM002, SM021, SM022 |
| CM002 | Precedence defines AI-driven drug discovery platforms to include software SaaS, platform-plus-wet-lab partnerships, CDMO/CRO integrations, data services, and related professional services sold to pharma, biotech, CROs, and research institutions. | 中 | SM002 |
| CM003 | BioMap’s disclosed product set fits the platform-plus-model-plus-wet-lab category more closely than a pure software point tool. | 中 | SM016, SM021 |
| CM004 | Global Market Insights estimated the AI drug discovery market at $3.1 billion in 2025 and $4.0 billion in 2026, growing to $43.9 billion by 2035. | 中 | SM001 |
| CM005 | Precedence said North America held the largest AI-driven drug-discovery-platform share in 2025. | 中 | SM002 |
| CM006 | Precedence said Asia Pacific is expected to post the fastest CAGR for AI-driven drug discovery platforms from 2026 to 2035. | 中 | SM002 |
| CM007 | MarketsandMarkets sized the broader drug discovery technologies market at $30.58 billion in 2025 and $51.51 billion by 2030. | 中 | SM003 |
| CM008 | MarketsandMarkets said the broader discovery-tech market is being pulled by advanced screening platforms and demand for biologics, cell and gene therapies, and RNA-based drugs. | 中 | SM003 |
| CM009 | MarketsandMarkets’ accessible excerpt cited a drug discovery informatics market growing from $2.2 billion in 2020 to $3.5 billion by 2025. | 低 | SM003 |
| CM010 | MarketsandMarkets’ accessible excerpt cited a life-science analytics market of $40.03 billion in 2025 and $68.81 billion by 2030. | 低 | SM003 |
| CM011 | Precedence said lead optimization and multi-parameter optimization held the largest market share by workflow in 2025. | 中 | SM002 |
| CM012 | Precedence said target identification and validation is the fastest-growing workflow segment. | 中 | SM002 |
| CM013 | Precedence said small-molecule support held the largest modality share in 2025. | 中 | SM002 |
| CM014 | Precedence said biologics is expected to be the fastest-growing modality. | 中 | SM002 |
| CM015 | Precedence said generative models are expected to be the fastest-growing AI stack inside AI-driven drug-discovery platforms. | 中 | SM002 |
| CM016 | Precedence said oncology led the AI-driven drug-discovery-platform market by therapeutic area in 2025. | 中 | SM002 |
| CM017 | Precedence said rare diseases and orphan indications are expected to be the fastest-growing therapeutic-area focus. | 中 | SM002 |
| CM018 | HKIC said BioMap had secured contracts with more than 200 users across pharma, CDMOs, innovative drug developers, synthetic biology, green technology, and research institutions by June 2024. | 中 | SM015 |
| CM019 | Harbour BioMed said BioMap supports institutional users across antibody/protein, innovative therapies and precision medicine, synthetic biology, and frontier scientific research. | 中 | SM016 |
| CM020 | Global Market Insights cited data issues and a lack of skilled resources as major restraints on AI drug discovery adoption. | 中 | SM001 |
| CM021 | IQVIA said biopharma R&D funding remained high in 2025 even though it slowed versus 2024. | 中 | SM005 |
| CM022 | IQVIA said growing scientific complexity and longer development timelines increased pressure on clinical productivity. | 中 | SM005 |
| CM023 | IQVIA said AI increasingly enabled R&D and provided early evidence of stronger success rates for AI-driven programs. | 中 | SM005 |
| CM024 | McKinsey argued that generative AI could transform nearly all parts of pharma but only if companies address industry-specific scaling challenges. | 中 | SM004 |
| CM025 | McKinsey framed generative AI as a large value opportunity but not one that can be captured through consumer-software-style deployment shortcuts. | 中 | SM004 |
| CM026 | Precedence’s market definition implies that the buyer base for AI-discovery platforms sits inside scientific and discovery organizations rather than generic enterprise IT alone. | 中 | SM002 |
| CM027 | Global Market Insights said Asia-Pacific growth is being accelerated by government support for AI-powered drug discovery and biotech funding. | 中 | SM001 |
| CM028 | Global Market Insights said North America leads because of high healthcare spending, advanced AI infrastructure, and dense biopharma hubs. | 中 | SM001 |
| CM029 | Recursion positions itself as a clinical-stage TechBio company advancing wholly owned and partnered pipeline programs, illustrating that AI-discovery buyers and competitors often monetize through pipeline economics as well as software. | 中 | SM006, SM007 |
| CM030 | Absci positions itself as a generative-AI biologics company with both internal and partnered programs. | 中 | SM008, SM009 |
| CM031 | Schrödinger positions itself as a physics-based software platform for molecular discovery and design. | 中 | SM010 |
| CM032 | Certara said it serves more than 2,400 biopharma companies, academia, and regulatory agencies, showing that adjacent discovery-platform buyers extend beyond drug sponsors alone. | 中 | SM011, SM012 |
| CM033 | Insilico markets both software platforms and pipeline programs from target identification through Phase II, showing how category boundaries blur between model vendors and drug developers. | 中 | SM013 |
| CM034 | BioMap’s practical market sits at the intersection of AI-discovery platforms, broader discovery technologies, and discovery-software adjacencies rather than a single clean external category. | 中 | SM001, SM002, SM003, SM021 |
| CM035 | Because BioMap is concentrated in biologics, generative models, and Asia-Pacific expansion, it is pointed toward the faster-growing segments of the market rather than the slowest-growing ones. | 中 | SM001, SM002, SM015, SM016 |
| CM036 | Public evidence from HKIC and Harbour shows that BioMap’s opportunity extends beyond pharma into synthetic biology and green-technology applications. | 中 | SM015, SM016 |
| CM037 | BioMap’s likely buying centers are discovery biology leaders, platform heads, computational-science teams, innovation offices, and business-development leaders rather than only CIOs. | 低 | SM002, SM015, SM016 |
| CM038 | The typical adoption path for a platform like BioMap runs from a scoped pilot into workflow integration and then into strategic co-development or pipeline ventures. | 中 | SM002, SM016 |
| CM039 | The main adoption constraints for BioMap are data quality, talent scarcity, wet-lab validation, proof-of-ROI demands, and the challenge of converting pilots into scaled recurring contracts. | 中 | SM001, SM004, SM005, SM015, SM016 |
| CM040 | Available market lenses are not directly comparable because some measure the AI-specific discovery platform category while others measure the full discovery-technology stack or analytics adjacency. | 中 | SM001, SM002, SM003 |
| CM041 | No accessible public source isolates BioMap’s exact SAM or SOM by customer class, geography, or contract size. | 低 | |
| CP001 | BioMap competes in a mixed field of AI-native TechBio platforms, discovery-software vendors, and internal-build substitutes. | 中 | SP001, SP015, SP017, SP022 |
| CP002 | Insilico markets an AI stack spanning target identification through Phase II, combining software and pipeline narratives. | 中 | SP017 |
| CP003 | Recursion positions itself as a clinical-stage TechBio with wholly owned and partnered pipeline assets. | 中 | SP001, SP019 |
| CP004 | Absci positions itself as a generative-AI biologics company with internal and partnered programs. | 中 | SP014, SP020 |
| CP005 | Owkin frames itself around biological artificial superintelligence and an autonomous AI scientist. | 中 | SP002 |
| CP006 | Valo pitches AI-enabled human causal biology and closed-loop chemistry for drug discovery. | 中 | SP003 |
| CP007 | insitro pitches itself as a machine-learning drug company built on data at scale. | 中 | SP004 |
| CP008 | BenevolentAI pitches life-science intelligence for complex R&D decisions. | 中 | SP005 |
| CP009 | Schrödinger positions itself as a physics-based molecular discovery and design software platform. | 中 | SP018 |
| CP010 | Certara positions itself as predictive drug-development software and services rather than a frontier biology foundation-model company. | 中 | SP015, SP016 |
| CP011 | As of July 2026, CompaniesMarketCap listed Recursion at roughly $1.75 billion, Absci at $1.72 billion, Schrödinger at $1.22 billion, and Certara at $1.05 billion of market capitalization. | 中 | SP006, SP007, SP008, SP009 |
| CP012 | Listed adjacent comparables publish annual-report histories through SEC EDGAR, giving them materially better public disclosure than BioMap. | 高 | SP010, SP011, SP012, SP013 |
| CP013 | BioMap’s main differentiation claims are xTrimo, BioMap OS, cross-sector biology use cases, and a dry-wet closed-loop architecture. | 中 | SP022, SP023 |
| CP014 | The Harbour-MegaStream venture shows BioMap is trying to compete for shared pipeline economics, not only workflow-software budgets. | 中 | SP021, SP024 |
| CP015 | BioMap’s closest direct peer set is better framed around Insilico, Recursion, Absci, Owkin, insitro, Valo, and BenevolentAI than around general-purpose software vendors. | 中 | SP001, SP002, SP003, SP004, SP005, SP014, SP017 |
| CP016 | Schrödinger and Certara are adjacent competitors because they sit in similar budgets and workflows even though their product philosophy differs from BioMap’s. | 中 | SP015, SP016, SP018 |
| CP017 | The status-quo substitute for BioMap is an internal discovery stack assembled from pharma scientists, point tools, CROs, and service vendors. | 中 | SP001, SP015, SP017, SP018 |
| CP018 | A buyer can also solve slices of the same job through separate docking, informatics, analytics, simulation, and lab-service vendors rather than one integrated platform. | 中 | SP010, SP015, SP016, SP018 |
| CP019 | Public websites provide almost no usable posted pricing for BioMap or its major peers, implying an enterprise-led rather than self-serve market. | 中 | SP014, SP015, SP016, SP017, SP018 |
| CP020 | Certara is the closest public analogue for scaled recurring software-plus-services economics among the compared companies. | 中 | SP015, SP016 |
| CP021 | Recursion and Insilico compete partly through pipeline economics and partnership upside rather than software-seat pricing alone. | 中 | SP001, SP017, SP019 |
| CP022 | Schrödinger’s strength is physics-based molecular software rather than protein-foundation-model specialization. | 中 | SP018 |
| CP023 | Absci’s strongest overlap with BioMap is generative-AI biologics discovery. | 中 | SP014, SP020 |
| CP024 | Owkin’s public narrative is more centered on autonomous scientific systems and multimodal biology intelligence than on protein design alone. | 中 | SP002 |
| CP025 | Valo’s public narrative is more centered on human causal biology and chemistry than on BioMap-style protein foundation models. | 中 | SP003 |
| CP026 | BioMap’s moat claims depend on xTrimo scale, dry-wet loop integration, Hong Kong trust, and cross-sector customer coverage. | 中 | SP022, SP023, SP024 |
| CP027 | Foundation-model commoditization is a high risk because BioMap competes in a category where model capability can diffuse rapidly. | 中 | SP005, SP017, SP022 |
| CP028 | Better-disclosed public comparables may look safer than BioMap to some customers and investors because BioMap reveals less governance, pricing, and operating detail. | 中 | SP010, SP011, SP012, SP013 |
| CP029 | Internal build and incumbent workflow vendors can blunt BioMap differentiation even without matching every model capability. | 中 | SP015, SP016, SP017, SP018 |
| CP030 | Multi-homing risk is high because buyers can often evaluate or deploy several AI-biology vendors in parallel. | 中 | SP014, SP015, SP017, SP018 |
| CP031 | BioMap gains real lock-in only when it captures proprietary wet-lab data loops, workflow integration, or shared-economics partnerships. | 中 | SP022, SP023, SP024 |
| CP032 | Distribution power in this market often sits with companies already embedded in discovery workflows or enterprise service relationships. | 中 | SP015, SP016, SP018, SP019 |
| CP033 | Because public pricing detail is scarce, capability and trust often matter more than advertised unit prices in competition. | 中 | SP014, SP015, SP016, SP017, SP018 |
| CP034 | The competitive field contains both pipeline-owning TechBio firms and software/service vendors, so one generic comparison source is insufficient. | 中 | SP001, SP010, SP015, SP017, SP018 |
| CP035 | BioMap’s white space is a protein-heavy, Asia-linked, cross-sector biology platform that still wants a route to shared pipeline economics. | 中 | SP022, SP023, SP024 |
| CP036 | Opaque pricing and limited public production-usage data make it hard to prove that any one vendor has durable competitive dominance today. | 中 | SP014, SP015, SP016, SP017, SP018 |
| CP037 | Public adjacent comparables clustered around roughly $1-2 billion of market cap in July 2026, which frames BioMap’s unicorn narrative competitively but not irrationally. | 中 | SP006, SP007, SP008, SP009 |
| CI001 | BioMap monetizes through a hybrid mix of collaboration economics, platform contracts, project delivery and future asset-upside structures rather than a single software subscription model. | 中 | SI001, SI007, SI009, SI013 |
| CI002 | The Sanofi collaboration starts with a $10 million upfront payment and includes more than $1 billion of potential milestones. | 高 | SI007, SI008 |
| CI003 | HKIC and HKSTP both described BioMap as having over 200 contracted AIGP users by mid-2024. | 高 | SI001, SI010 |
| CI004 | Yicai reported that BioMap OS had already been validated across more than 60 projects. | 中 | SI002 |
| CI005 | Yicai reported that BioMap had served over 800 global institutions and more than 30 leading enterprises. | 中 | SI002 |
| CI006 | PR Newswire likewise described BioMap as supporting more than 800 institutional users worldwide, corroborating that the company has meaningful top-of-funnel commercial reach. | 高 | SI009, SI002 |
| CI007 | Sohu described BioMap as having potential orders totaling about $2 billion, but that figure is pipeline-like opportunity rather than recognized revenue. | 中 | SI003 |
| CI008 | The MegaStream structure gives BioMap possible upfronts, milestones and royalty-sharing economics beyond simple software fees. | 中 | SI009 |
| CI009 | Partnership announcements with Kexing and Optoseeker show BioMap commercializing via bespoke co-development and workflow integration, not public self-serve pricing. | 中 | SI011, SI012, SI013 |
| CI010 | Direct public list pricing for BioMap remains unavailable, so monetization must be inferred from deal structures, contract counts and partner narratives. | 中 | SI001, SI007, SI008, SI011, SI024 |
| CI011 | Public traction is expressed mainly through users, institutions, projects and partnership logos rather than disclosed revenue or ARR. | 中 | SI001, SI002, SI009 |
| CI012 | HKIC described AIGP customers spanning pharma, CDMOs, innovative drug developers, synthetic biology, green technology enterprises and research institutions. | 中 | SI001 |
| CI013 | Public reports place BioMap with multinational pharma, Chinese biopharma and research users, indicating a long enterprise-sales motion across multiple verticals. | 中 | SI001, SI002, SI003 |
| CI014 | MegaStream and Optoseeker-style deployments imply BioMap is moving toward platform-plus-service and platform-plus-pipeline revenue, not just software enablement. | 中 | SI009, SI013 |
| CI015 | BioMap’s own-lab strategy means some delivery cost sits in wet-lab and high-throughput execution rather than purely digital inference. | 中 | SI005, SI009, SI013 |
| CI016 | The Optoseeker collaboration explicitly points to deployment at BioMap’s Suzhou high-throughput experimental center to provide experimental services for global customers. | 中 | SI013 |
| CI017 | Sales-efficiency proxies are directionally positive but noisy because BioMap discloses mixed units such as projects, institutions, enterprises and contracts. | 中 | SI001, SI002, SI003, SI009 |
| CI018 | Multiple outlets reported that the confidential Hong Kong IPO sought several hundred million dollars, implying material forward capital need. | 高 | SI002, SI003, SI004 |
| CI019 | BioMap’s July 2021 Series A was reported at roughly $100 million and earmarked mainly for R&D and talent. | 中 | SI005, SI006 |
| CI020 | Sohu reported that cumulative financing had exceeded $200 million by March 2026. | 中 | SI003 |
| CI021 | HKIC announced a strategic investment and resource-support package, but not the check size, leaving capital adequacy only partially visible. | 高 | SI001, SI010 |
| CI022 | The Hong Kong InnoHub and BioX accelerator commit BioMap to support more than 50 projects over five years, which implies ongoing opex and ecosystem-spend obligations. | 高 | SI001, SI010 |
| CI023 | Building proprietary models, running dry-wet loops and operating high-throughput centers makes BioMap more capital-intensive than a pure software vendor. | 中 | SI005, SI009, SI013 |
| CI024 | Public cash balance, monthly burn, runway, debt and capex commitments are not disclosed in any source reviewed for this chapter. | 中 | SI001, SI002, SI003, SI014, SI024, SI025 |
| CI025 | The confidential HKEX filing is the clearest public next-round trigger because it is the only disclosed path to large-scale new capital. | 高 | SI002, SI003, SI004 |
| CI026 | Revenue quality is mixed: partnership economics are visible, recurring platform revenue is implied, but audited revenue is absent. | 中 | SI001, SI007, SI008, SI009 |
| CI027 | Public TechBio and drug-software comparables monetize at very different scales, showing that similar narratives can map to very different revenue outcomes. | 中 | SI016, SI017, SI018, SI019 |
| CI028 | CompaniesMarketCap listed TTM revenue of about $65.73 million for Recursion, $2.8 million for Absci, $250 million for Schrödinger and $410 million for Certara. | 中 | SI016, SI017, SI018, SI019 |
| CI029 | SEC 10-Q availability for those comparables illustrates the disclosure standard investors can use for public peers but do not yet have for BioMap. | 高 | SI020, SI021, SI022, SI023 |
| CI030 | Tracxn still surfaces only one disclosed Series A round, highlighting how third-party databases lag or conflict with later media coverage of the HKIC round. | 中 | SI014, SI001, SI003 |
| CI031 | An adverse outside framing exists: AInvest’s headline cast the confidential filing as a valuation and governance-risk narrative-control exercise. | 中 | SI015 |
| CI032 | Public evidence supports real commercialization, but it does not support an underwriteable revenue figure. | 中 | SI001, SI002, SI003, SI007, SI008, SI009 |
| CI033 | Any margin path remains unproven because the split among compute cost, data generation, wet-lab cost and services cost is undisclosed. | 中 | SI005, SI009, SI013, SI024 |
| CI034 | Capital adequacy is directionally stronger than an early-stage AI-biotech startup because BioMap has a $100 million Series A base, strategic-government backing and an IPO process underway, but exact runway is unknown. | 高 | SI001, SI003, SI005, SI006, SI018 |
| CI035 | The most important diligence blockers are audited revenue, backlog conversion, concentration by top customer, and burn-to-runway visibility. | 中 | SI002, SI003, SI007, SI008, SI024, SI025 |
| CI036 | BioMap should be underwritten as a hybrid platform, services and asset-creation business rather than as a clean SaaS multiple story. | 中 | SI007, SI008, SI009, SI011, SI013 |
| CI037 | The Sanofi upfront and milestone structure proves partner willingness to pay for BioMap’s platform, but it is not equivalent to recurring recognized revenue. | 高 | SI007, SI008 |
| CE001 | BioMap OS is presented as a foundation-model-driven dry-wet closed-loop life-science discovery system that helps users with data insight, parameter optimization and de novo design. | 中 | SE002 |
| CE002 | BioMap says the product can be deployed as model, software or complete system, with nearly 100 combinable modules, global cloud access, localized deployment and professional services. | 中 | SE002 |
| CE003 | BioMap OS is publicly framed around four vertical solution systems: antibody and innovative protein, innovative therapy and precision medicine, synthetic biology, and frontier research. | 中 | SE002 |
| CE004 | BioMap’s bundle and Yicai both describe xTrimo V4 at 268 billion parameters. | 高 | SE002, SE006 |
| CE005 | The bundle states that xTrimo V4 achieves more than 300 SOTA results across many life-science tasks. | 中 | SE002 |
| CE006 | BioMap’s about-page strings say BioMap OS has served more than 800 institutional users and achieved experimental validation in more than 60 FIC discovery projects. | 中 | SE002 |
| CE007 | Yicai corroborated the 60-plus project validation claim while describing BioMap OS as an LLM-driven discovery system. | 中 | SE006 |
| CE008 | HKIC described xTrimo as a 100Bn-plus parameter foundation model in June 2024, showing that BioMap’s public model-size narrative evolved across versions. | 高 | SE004, SE006, SE007 |
| CE009 | The xTrimoPGLM GitHub repository exposes open-source 1B, 3B, 10B masked models, 1B, 3B, 7B causal models and 100B INT4 inference artifacts. | 中 | SE013 |
| CE010 | The xTrimoPGLM paper says the model was trained at 100B parameters and 1 trillion training tokens. | 高 | SE014, SE013 |
| CE011 | The xTrimoPGLM paper says the model outperformed advanced baselines across 18 protein-understanding benchmarks and supports both understanding and generation tasks. | 中 | SE014 |
| CE012 | BioMap’s core product description says the platform integrates knowledge aggregation, predictive design, experimental control and model training. | 中 | SE002 |
| CE013 | The product bundle describes five operating units: knowledge, prediction and design, intelligent experiment, data, and model training. | 中 | SE002 |
| CE014 | BioMap publicly claims a large-scale high-throughput Suzhou laboratory that closes the loop between model hypotheses and experimental validation. | 高 | SE002, SE011 |
| CE015 | The bundle attributes the infrastructure layer to a global multi-cloud bio-computing engine with dynamic scheduling, hardware abstraction, autonomous diagnosis and closed-loop feedback. | 中 | SE002 |
| CE016 | Optoseeker and BioMap described an AI-agent workflow that combines high-throughput single-cell screening with BioMap’s models and real-time feedback. | 中 | SE011 |
| CE017 | The Kexing partnership described AI-exclusive large models and AI intelligent laboratories spanning the full development process for macromolecular drugs. | 中 | SE012 |
| CE018 | The MegaStream announcement describes a stack of exclusive datasets, purpose-built large models and an integrated dry-wet closed-loop discovery laboratory. | 中 | SE008 |
| CE019 | MegaStream projects more than 500% efficiency gains and more than 5 petabytes of AI-ready life-science data within five years. | 中 | SE008 |
| CE020 | The BioMap research organization shows continued public engineering and research activity through July 2026 across multiple repositories. | 高 | SE019, SE021, SE022, SE023, SE027 |
| CE021 | PFMBench gives BioMap a public benchmark suite covering 38 downstream tasks and 17 pre-trained models, which is stronger technical proof than simple marketing claims. | 高 | SE015, SE016 |
| CE022 | ProteinReasoner extends BioMap’s protein stack into multi-modal reasoning and chain-of-thought style design assistance. | 中 | SE017, SE002 |
| CE023 | RNAGenesis shows that BioMap’s model roadmap is expanding beyond proteins into RNA therapeutics. | 中 | SE018, SE002 |
| CE024 | BioLab indicates a roadmap toward autonomous, multi-agent life-science research rather than only static prediction models. | 中 | SE020, SE002 |
| CE025 | xTrimoMultimer extends the stack into monomer and multimer structure prediction on GPU clusters. | 中 | SE023 |
| CE026 | ProteinSage adds a structurally constrained protein foundation model to the public BioMap research footprint. | 中 | SE022 |
| CE027 | The scFoundation repository shows BioMap-adjacent work in large-scale single-cell modeling, broadening the underlying data and modeling toolchain, and its API example points users to a newer AIGP surface with online inference and CLI tools. | 中 | SE021, SE027, SE026 |
| CE028 | De-novoVHH and InverseFoldingEvaluation show public workflow assets around antibody design and evaluation rather than only a generic protein model story. | 中 | SE024, SE025 |
| CE029 | BioMap’s security statement says it obtained ISO/IEC 27001:2022 certification in 2023, uses TLS 1.2 in transit, AES-256 at rest and annual third-party penetration tests. | 中 | SE003 |
| CE030 | BioMap’s privacy policy says the site may collect name, email, postal address, phone number, job title and organization name through forms, inputs and email correspondence. | 中 | SE003 |
| CE031 | Public trust evidence is information-security centric; the reviewed sources do not show clinical-grade, GxP, HIPAA or regulated-lab certifications. | 中 | SE003, SE001, SE004 |
| CE032 | No public status page, uptime archive or incident log was identified in the sources reviewed for this chapter. | 中 | SE001, SE002, SE003 |
| CE033 | BioMap’s public product proof is strongest for protein and antibody discovery; other verticals such as RNA, synthetic biology and autonomous agents look more roadmap-heavy or research-oriented. | 中 | SE002, SE017, SE018, SE020 |
| CE034 | The differentiation case depends on combining data, models, wet-lab execution and workflow orchestration rather than on a single checkpoint or single benchmark. | 中 | SE002, SE008, SE010, SE011 |
| CE035 | Critical technical dependencies include compute/cloud infrastructure, partner datasets, wet-lab throughput, screening hardware and continual high-quality data capture. | 中 | SE002, SE008, SE011, SE013 |
| CE036 | The public developer footprint is meaningful, but it is fragmented across research repositories and papers rather than a single externally documented production platform. | 中 | SE013, SE015, SE019, SE021, SE022, SE023 |
| CE037 | Overall, BioMap is delivering a discovery operating system plus a research-model family, not a single turnkey drug product. | 高 | SE002, SE004, SE006, SE008 |
| CU001 | BioMap’s customer base is segmented across multinational pharma, Chinese biopharma, CDMOs, research institutions, synthetic-biology groups and green-tech enterprises. | 高 | SU001, SU005 |
| CU002 | HKIC said BioMap had more than 200 contracted AIGP users including international pharma companies, leading CDMOs, innovative drug developers, synthetic-biology companies, green-tech enterprises and research institutions. | 高 | SU001, SU002 |
| CU003 | Yicai reported that BioMap had served more than 800 global institutions and more than 30 leading enterprises. | 中 | SU003 |
| CU004 | PR Newswire likewise described BioMap as supporting over 800 institutional users across multiple verticals. | 高 | SU005, SU003 |
| CU005 | BioMap’s own bundle strings also say the platform has served over 800 institutional users and validated more than 60 FIC discovery projects. | 中 | SU021 |
| CU006 | The public adoption story is stronger on breadth metrics than on named production deployments. | 中 | SU001, SU003, SU005, SU021 |
| CU007 | Sanofi is the strongest publicly named enterprise proof because both independent media sources describe a live discovery collaboration with explicit platform scope. | 高 | SU006, SU007 |
| CU008 | Harbour BioMed is a second major named biopharma proof because the two parties are co-founding MegaStream around BioMap’s platform and Harbour’s datasets and development capabilities. | 高 | SU005, SU011 |
| CU009 | Kexing Biopharm shows a domestic Chinese biopharma customer / partner segment focused on tumor and autoimmune macromolecular drug development. | 高 | SU008, SU013 |
| CU010 | Optoseeker shows an instrumentation-led customer / partner segment where BioMap’s value is AI-driven screening and workflow integration rather than direct drug-asset ownership. | 高 | SU009, SU010, SU014 |
| CU011 | Sohu named Sanofi, CSPC and Dabeinong as clients, but customer-side corroboration is weaker for the latter two than for Sanofi. | 中 | SU004, SU012, SU018, SU026 |
| CU012 | HKIC’s summit attendee list involving Sanofi, Lilly China, LanzaTech and Syngenta indicates BioMap’s ecosystem reaches beyond pure pharma, but attendance is weaker proof than deployment. | 中 | SU001, SU015, SU016, SU017 |
| CU013 | The named-customer set spans biologics discovery, precision medicine, instrumentation, agriculture-adjacent life science and green-tech applications. | 中 | SU001, SU005, SU012, SU015, SU016 |
| CU014 | The existence of a live AIGP/xTrimo Explorer surface and a documented migration path from an older platform implies an active user base beyond one-off bespoke projects. | 高 | SU020, SU025 |
| CU015 | More than 60 validated projects indicates repeat technical use, but it does not disclose how many of those projects are paid production deployments. | 中 | SU003, SU021 |
| CU016 | The public record separates a narrow monetized core of 200-plus contracts from a much broader halo of 800-plus institutions. | 中 | SU001, SU002, SU003, SU005 |
| CU017 | Named customer proof is mostly derived from launch and collaboration announcements rather than independent case studies or procurement records. | 中 | SU005, SU006, SU007, SU008, SU009, SU010 |
| CU018 | Sanofi-level proof is strongest because it includes explicit task scope and disclosed deal economics, not just a logo. | 高 | SU006, SU007 |
| CU019 | Harbour, Kexing and Optoseeker provide useful named proof, but public outcome metrics remain limited to qualitative workflow descriptions. | 中 | SU005, SU008, SU009, SU010 |
| CU020 | Public retention metrics such as NRR, GRR, churn, renewal rate and contract length are not disclosed. | 中 | SU001, SU003, SU021, SU023 |
| CU021 | Because retention metrics are absent, durability must be inferred from continued platform migration, ongoing collaborations and repeated project counts. | 中 | SU005, SU006, SU010, SU014, SU025 |
| CU022 | The public data do not reveal top-customer revenue concentration, leaving open the possibility that a few marquee relationships dominate economics. | 中 | SU003, SU004, SU019, SU023 |
| CU023 | The customer journey likely runs from experimental exploration to contracted AIGP use, then into custom workflows, wet-lab services and potentially asset-level partnerships. | 中 | SU001, SU005, SU010, SU020, SU025 |
| CU024 | The largest public expansion vector is moving from platform use into deeper workflow integration and joint program creation, as seen in Sanofi, Harbour and Optoseeker. | 中 | SU005, SU006, SU010 |
| CU025 | A second expansion vector is segment expansion from pharma into green tech, agriculture and industrial biology where proof is still earlier-stage. | 中 | SU001, SU015, SU016 |
| CU026 | Procurement friction likely includes opaque pricing, data-governance review and integration burden because BioMap does not publish customer-facing contract terms or standard ROI metrics. | 中 | SU001, SU022, SU023 |
| CU027 | BioMap’s privacy disclosures and security claims may help enterprise procurement, but they do not substitute for customer-reference or implementation data. | 中 | SU022, SU001 |
| CU028 | Counterparty homepages confirm that Sanofi, Harbour, Kexing and Optoseeker are real operating organizations in the exact verticals BioMap claims to serve. | 高 | SU011, SU012, SU013, SU014 |
| CU029 | Syngenta, LanzaTech and Lilly fit BioMap’s cross-vertical buyer map, but current public evidence is better for ecosystem proximity than for active paid deployment. | 中 | SU001, SU015, SU016, SU017 |
| CU030 | CSPC’s and Dabeinong’s homepages confirm they are plausible customer categories, but customer-side proof that BioMap is actively deployed there remains weak. | 中 | SU004, SU018, SU026 |
| CU031 | The public complaint / churn surface is unusually quiet, which may reflect strong enterprise focus but also limited public customer-review exposure. | 中 | SU019, SU023 |
| CU032 | The absence of public complaint boards or review marketplaces means customer-quality underwriting is limited more by missing disclosure than by explicitly negative evidence. | 中 | SU019, SU023 |
| CU033 | Overall adoption appears real and broad, but the gap between “institutional reach” and “verifiable durable production deployment” remains large. | 中 | SU001, SU003, SU005, SU006, SU021 |
| CU034 | BioMap’s customer proof is strongest at the named-enterprise and strategic-partner layer, weaker at the usage-retention and satisfaction layer. | 中 | SU006, SU007, SU008, SU009, SU010, SU020 |
| CU035 | The customer chapter therefore supports a positive adoption view but a medium-confidence durability view. | 中 | SU001, SU003, SU005, SU020, SU023 |
| CR001 | BioMap’s Hong Kong operations bring it within a serious privacy-regulation environment because the PDPO is a mature comprehensive data-protection regime. | 高 | SR001, SR003, SR004 |
| CR002 | BioMap’s own privacy-policy text confirms that it collects identifiable contact and organization data through forms, inputs and emails. | 中 | SR002 |
| CR003 | Because BioMap uses Hong Kong positioning and multinational-client trust as a selling point, any gap in cross-border privacy controls or documentation could become a commercial as well as legal risk. | 中 | SR001, SR002, SR003 |
| CR004 | BioMap’s public trust package is strongest at the infosec layer—ISO/IEC 27001, TLS 1.2, AES-256 and annual third-party penetration tests. | 中 | SR002 |
| CR005 | Those controls do not by themselves prove regulated-life-science readiness such as GxP, clinical-lab or other workflow-specific compliance. | 中 | SR002, SR003 |
| CR006 | The confidential Hong Kong IPO filing creates a disclosure risk because public investors still lack the prospectus-level detail needed to assess governance, use of funds and concentration. | 高 | SR005, SR006, SR007 |
| CR007 | Media and database coverage of BioMap’s financing history are not perfectly reconciled, which is a governance and diligence risk in itself. | 中 | SR005, SR006, SR007 |
| CR008 | AInvest’s framing adds a specifically adverse outside view that the confidential filing may be partly about narrative control and valuation management. | 中 | SR007 |
| CR009 | Public scientific proof is real: xTrimoPGLM and PFMBench expose model families, benchmark suites and open artifacts rather than only marketing copy. | 高 | SR016, SR017, SR018, SR019 |
| CR010 | The translation risk is that strong protein-model benchmarks do not automatically translate into better drug programs, customer ROI or clinical outcomes. | 中 | SR017, SR019, SR020, SR021 |
| CR011 | BioMap’s roadmap toward RNA therapeutics and autonomous multi-agent research expands technical upside but increases execution complexity and scope risk. | 中 | SR021, SR022, SR013 |
| CR012 | The public stack depends on a large-scale high-throughput Suzhou laboratory and dry-wet closed-loop operations, making facility and throughput reliability material. | 高 | SR011, SR013 |
| CR013 | Multi-cloud deployment, hardware abstraction and autonomous diagnosis imply a complex operational footprint that is harder to secure and operate than a simple model API. | 中 | SR013, SR014, SR015 |
| CR014 | No public uptime archive, status page or incident ledger was identified in the retained pack, leaving reliability risk materially under-documented. | 中 | SR002, SR013, SR014 |
| CR015 | Optoseeker integration makes BioMap dependent on high-throughput screening hardware and workflow interoperability, not just software quality. | 中 | SR011, SR025 |
| CR016 | The Kexing collaboration adds value but also underlines that BioMap’s delivery model can extend into more operationally demanding AI-lab workflows. | 中 | SR012, SR026 |
| CR017 | MegaStream concentrates critical dependency on Harbour BioMed’s datasets, antibody platform and global clinical-development capabilities. | 高 | SR008, SR024 |
| CR018 | The Sanofi collaboration creates upside but also partner concentration and milestone-dependence risk because it is the clearest public flagship relationship. | 中 | SR009, SR010, SR023 |
| CR019 | Public customer evidence is broad but economically opaque, so marquee-partner concentration may be much higher than the 800-institution figure suggests. | 中 | SR003, SR006, SR007, SR008 |
| CR020 | The product migration from older surfaces to AIGP shows continuity, but also creates implementation and service-transition risk if user experience or support slips. | 中 | SR014, SR015 |
| CR021 | Public counterparty pages confirm that BioMap is serving serious organizations, but they do not remove dependency risk because those counterparties are much larger and often control the surrounding workflow. | 中 | SR023, SR024, SR025, SR026 |
| CR022 | Funding dependence remains material because public sources still point to several-hundred-million-dollar IPO targets after prior large funding rounds. | 高 | SR005, SR006, SR007 |
| CR023 | Building models, generating data and running wet-lab systems makes BioMap more capital-intensive than a pure software company. | 中 | SR008, SR011, SR013 |
| CR024 | The absence of public cash, burn and runway data means financing risk cannot be bounded tightly from public evidence. | 中 | SR005, SR006, SR007 |
| CR025 | Cross-sector customer ambitions across pharma, green tech, synthetic biology and frontier research create focus-dilution risk. | 中 | SR001, SR003, SR008, SR021 |
| CR026 | The breadth of BioMap’s research repos and papers reduces black-box risk but increases reputational risk if scientific claims do not convert into commercial outcomes. | 中 | SR016, SR018, SR020, SR022 |
| CR027 | The developer footprint is an execution advantage for hiring and transparency, but it also exposes BioMap’s methods to closer external scrutiny. | 中 | SR016, SR018, SR020 |
| CR028 | Hong Kong support from HKIC and HKSTP mitigates ecosystem and policy risk by improving local credibility, partner access and strategic-enterprise standing. | 高 | SR003, SR004 |
| CR029 | BioMap’s named partner diversity across Sanofi, Harbour, Kexing and Optoseeker mitigates the risk of being dependent on only one customer segment. | 中 | SR009, SR011, SR012 |
| CR030 | The Suzhou lab and high-quality data loop can be a moat if executed well, but they are also a single-point-of-failure risk if utilization or data quality disappoints. | 中 | SR011, SR013 |
| CR031 | Public competitor disclosure surfaces such as Schrödinger’s investor SEC-filings page show how much more transparent a public comp can be than BioMap today. | 高 | SR027, SR005 |
| CR032 | Broken or shifting IR reference paths on some public comps are a reminder that external web evidence can be brittle, reinforcing the need for direct diligence packs rather than web-only comfort. | 中 | SR028, SR029, SR030 |
| CR033 | There is no public evidence in this pack of enforcement actions, major litigation or safety incidents against BioMap itself. | 中 | SR001, SR002, SR005 |
| CR034 | The absence of disclosed incidents should not be read as proof of low risk because BioMap remains private and public disclosure obligations are lighter than for listed peers. | 中 | SR005, SR027 |
| CR035 | BioMap’s trust pitch explicitly relies on globally aligned IP and data regulations in Hong Kong, which means any regulatory change or compliance stumble could hit both sales and narrative. | 中 | SR003, SR004, SR001 |
| CR036 | Because BioMap’s strongest public proof is still collaboration-driven, commercialization failure risk remains tied to whether those partners expand into durable recurring usage. | 中 | SR008, SR009, SR010, SR011 |
| CR037 | The customer chapter’s missing retention and concentration data convert directly into risk because investors cannot yet tell whether breadth equals durability. | 中 | SR003, SR008, SR014 |
| CR038 | The right current risk view is not “red flag disaster” but “high-complexity, medium-transparency platform” where execution mistakes would transmit quickly into customer, margin and financing outcomes. | 中 | SR003, SR006, SR013, SR024 |
| CR039 | Mitigation maturity is strongest in security controls and ecosystem support, weaker in public operating metrics, disclosure discipline and customer-durability evidence. | 中 | SR002, SR003, SR013, SR027 |
| CR040 | Thesis-break triggers should therefore focus on failed IPO execution, material partner non-expansion, evidence of data-governance friction, or inability to show repeat paid production usage. | 中 | SR005, SR007, SR009, SR014 |
| CV001 | BioMap has a credible strategic-asset narrative because it combines foundation models, dry-wet workflows, 800-plus institutional users, flagship partners and Hong Kong strategic backing. | 中 | SV003, SV004, SV005, SV026, SV027 |
| CV002 | The anti-thesis is that BioMap remains financially opaque, milestone-heavy and operationally complex, which makes any premium valuation difficult to underwrite. | 中 | SV001, SV002, SV006, SV024, SV030 |
| CV003 | Sanofi’s $10 million upfront and more-than-$1 billion milestone structure proves strategic willingness to pay, but it does not establish recurring revenue quality. | 高 | SV006, SV007 |
| CV004 | Harbour/MegaStream adds asset-upside optionality through upfronts, milestones and royalties, which can justify upside but also make valuation more scenario-dependent. | 中 | SV005, SV029 |
| CV005 | The confidential IPO process and HKIC support suggest BioMap is aiming to be valued as a strategic platform company rather than a narrow services vendor. | 高 | SV001, SV003, SV004 |
| CV006 | AInvest’s adverse framing means downside investors can plausibly argue the current narrative is ahead of the public proof set. | 中 | SV024 |
| CV007 | We rate BioMap track with medium confidence, a high risk rating and a stretched valuation stance. | 中 | SV001, SV002, SV024, SV030 |
| CV008 | A reasonable overall score from public evidence is about 5.8 out of 10: interesting enough to track closely, not transparent enough to pay up aggressively. | 中 | SV001, SV003, SV024 |
| CV009 | The recommendation is price-sensitive because BioMap’s quality signals are real but the absence of audited revenue, margin and retention data caps conviction. | 中 | SV001, SV002, SV026, SV030 |
| CV010 | Absent IPO-grade disclosure, investors should demand entry discipline rather than rely on the unicorn narrative alone. | 中 | SV001, SV024, SV025 |
| CV011 | Publicly disclosed financing supports a chronology of a $100 million Series A, cumulative funding above $200 million, HKIC strategic investment and a several-hundred-million-dollar IPO target. | 高 | SV001, SV002, SV003, SV025 |
| CV012 | Potential order value of $2 billion and more than 800 institutional users show demand strength, but not recognized revenue or gross-margin quality. | 中 | SV002, SV026 |
| CV013 | The absence of disclosed preference stack, dilution terms and cap-table cleanup remains a meaningful entry-risk for late-stage private investors. | 中 | SV001, SV024, SV025 |
| CV014 | HKIC is strategically valuable capital because it can improve local policy support and international signaling, but it is not a substitute for revenue quality. | 中 | SV003, SV004 |
| CV015 | Our base case assumes BioMap is a real but still hybrid platform whose verified economics settle closer to public biotech-software comps than to frontier-AI premium narratives. | 中 | SV008, SV009, SV010, SV011, SV012, SV013, SV014, SV015 |
| CV016 | Our bull case assumes the IPO prospectus reveals strong recurring platform revenue, acceptable margins, repeat enterprise usage and durable flagship-partner expansion. | 中 | SV001, SV003, SV026, SV027 |
| CV017 | Our bear case assumes commercialization is thinner than the user metrics imply, partner concentration is high, and the market discounts BioMap toward lower-end biotech software valuations. | 中 | SV002, SV024, SV030 |
| CV018 | The scenario range is necessarily wide because the same company can be read as a strategic AI-biotech platform or as an opaque services-and-milestones hybrid. | 中 | SV001, SV002, SV006, SV024 |
| CV019 | A public-comp anchor is useful because Recursion, Absci, Schrödinger and Certara all trade around roughly $1.0-1.75 billion of market value as of July 2026. | 中 | SV008, SV009, SV010, SV011 |
| CV020 | Recursion’s combination of about $1.75 billion market cap and $65.73 million revenue implies a valuation multiple around 26.6x sales. | 中 | SV008, SV012 |
| CV021 | Absci’s roughly $1.72 billion market cap on only about $2.8 million of revenue shows how option value can produce extreme multiples for speculative AI-biologics stories. | 中 | SV009, SV013 |
| CV022 | Schrödinger’s roughly $1.22 billion market cap, about $250 million of revenue and around 4.79x sales provide a more software-like public benchmark. | 中 | SV010, SV014, SV018 |
| CV023 | Certara’s roughly $1.05 billion market cap, about $410 million of revenue and around 2.54x sales provide a lower-risk, lower-multiple software-services reference point. | 中 | SV011, SV015, SV019 |
| CV024 | Those public comps show that the market is willing to assign anywhere from low-single-digit to very high sales multiples depending on platform optionality, revenue quality and transparency. | 中 | SV008, SV009, SV010, SV011, SV018, SV019 |
| CV025 | Because BioMap’s revenue denominator is not public, any attempt to apply a headline multiple directly is false precision. | 中 | SV001, SV002, SV025 |
| CV026 | The more defensible way to value BioMap today is to underwrite scenario bands and require diligence triggers that would justify moving up or down the band. | 中 | SV024, SV030 |
| CV027 | Biopharma partnerships like Sanofi and MegaStream are better read as valuation-supporting proof points and option value than as clean recurring-revenue comparables. | 中 | SV005, SV006, SV007 |
| CV028 | Public SEC filing surfaces for Recursion, Absci, Schrödinger and Certara highlight how much more detailed public-comparable underwriting can be than BioMap underwriting today. | 高 | SV020, SV021, SV022, SV023 |
| CV029 | BioMap’s strongest exit narratives are an HKEX IPO or strategic acquisition by a large pharma, techbio or data-platform actor, but both require cleaner economics and governance proof. | 中 | SV001, SV003, SV028, SV029 |
| CV030 | A failed or delayed IPO would be a serious negative signal because it would extend financing uncertainty without adding disclosure. | 中 | SV001, SV024 |
| CV031 | The core thesis-break triggers are weak prospectus economics, heavy customer concentration, no evidence of recurring revenue durability, or a valuation ask far above what the proof set supports. | 中 | SV001, SV024, SV026, SV030 |
| CV032 | The highest-priority final diligence asks are audited revenue, revenue mix, top-customer concentration, gross margin, burn, runway and the preference stack. | 中 | SV001, SV002, SV024, SV025 |
| CV033 | The market-supportive case is that BioMap may deserve a premium to mature software-services comps because of model IP, strategic partners and option value. | 中 | SV003, SV005, SV006, SV026 |
| CV034 | The discount case is that BioMap may deserve a discount to hype-cycle AI multiples because public evidence still does not show software-like margins, retention or disclosure. | 中 | SV001, SV002, SV024, SV030 |
| CV035 | Base-case fair value is best expressed as a broad $1.5-2.5 billion range, which acknowledges platform upside but still centers on public-comp reality. | 中 | SV008, SV009, SV010, SV011, SV022, SV023 |
| CV036 | Bull-case fair value of roughly $3-5 billion would require prospectus evidence that BioMap is closer to a durable platform-of-record than to a project-heavy hybrid. | 中 | SV001, SV003, SV005, SV026 |
| CV037 | Bear-case fair value of roughly $0.7-1.2 billion would be plausible if economics, concentration and financing risk look materially worse once disclosed. | 中 | SV002, SV024, SV030 |
| CV038 | The recommendation is therefore to track, not pass forever: BioMap is interesting enough that better disclosure could improve the call materially. | 中 | SV001, SV003, SV026 |
| CV039 | The company would become more investable quickly if the public filing clarifies recurring revenue quality, margin path and customer durability. | 中 | SV001, SV025, SV030 |
| CV040 | Until then, the right posture is to avoid pricing BioMap as if the strategic narrative has already been fully proven in public financial form. | 中 | SV001, SV024, SV030 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | BioMap | BioMap website bundle (about, product, technology and news strings) | BioMap is the pioneer in life science AI foundation models. As the world's first 268B-parameter foundation model, BioMap's xTrimo V4... achieved 300+ State-of-the-Art model performances in over 20 fields. |
| SO002 | Tracxn | BioMap company profile | BioMap is a series A company based in Beijing (China), founded in 2020 by Robin Li and Wei Liu. |
| SO003 | BaiduWiki | BioMap | On September 25, 2020, the life science platform company “BioMap” was officially established. |
| SO004 | Hong Kong Investment Corporation | Hong Kong Investment Corporation Limited x BioMap Sign Strategic Partnership Agreement (June 2024) | To date, BioMap has secured contracts with over 200 users based on the AIGP platform. |
| SO005 | HKSTP | HKSTP Congratulates BioMap on Signing Strategic Partnership Agreement with HKIC | BioMap... is collaborating with over 10 commercial partners and more than 200 academic institutions. |
| SO006 | HKSTP | HKSTP congratulations press-release PDF for BioMap / HKIC partnership | BioMap, headquartered in the US, is one of the second batch of strategic enterprises signed by the Office for Attracting Strategic Enterprises. |
| SO007 | VCBeat / VBData | BioMap Completes $100 Million Series A Financing | BioMap today announced the completion of a Series A financing of 100 million dollars, led by GGV Capital, with participation from Baidu, Legend Capital, BlueRun Ventures, Zhenzhi Capital and Xiang He Capital. |
| SO008 | ACN Newswire / Legend Capital | Legend Capital invests in Series A funding round of BioMap, a biological computing platform | BioMap... has recently completed the Series A funding round worth over a hundred million US dollars. |
| SO009 | Yicai Global | Baidu-Backed BioMap Reportedly Files for Hong Kong IPO | BioMap has served over 800 global institutions and more than 30 leading enterprises, it said. |
| SO010 | Sohu | 李彦宏创立的百图生科冲刺港股:中金、大摩、瑞银联手,去年CEO刘维表示未来一年半内寻求香港上市 | 百图生科已以保密方式向香港联合交易所递交上市申请,拟募集资金规模达数亿美元。 |
| SO011 | Tencent News | 传百度支持的医药AI公司「百图生科」已秘密申请香港IPO,拟筹数亿美元,选定中金、大摩和瑞银合作 | 百图生科正在与中金公司、摩根士丹利和瑞银集团等投行合作推进上市事宜。 |
| SO012 | pharmaphorum | Sanofi partners BioMap on AI hunt for biologic therapies | The alliance is getting underway with a $10 million upfront payment. |
| SO013 | PMLive | Sanofi partners with AI specialist BioMap in deal worth more than $1bn | BioMap will receive an upfront payment of $10m and will be eligible to receive over $1bn based on the achievement of pre-clinical development, clinical development, regulatory and commercial milestones. |
| SO014 | PR Newswire / Harbour BioMed | Harbour BioMed and BioMap Jointly Initiate MegaStream TechBio | BioMap currently supports over 800 institutional users worldwide. |
| SO015 | GitHub | biomap-research/xTrimoPGLM | xTrimoPGLM is the open-source version of the latest protein language models towards protein understanding tasks and protein design. |
| SO016 | arXiv | xTrimoPGLM: Unified 100B-Scale Pre-trained Transformer for Deciphering the Language of Protein | xTrimoPGLM is a 100B-scale protein language model. |
| SO017 | GitHub | biomap-research/PFMBench | PFMBench is a unified benchmark suite for evaluating Protein Foundation Models across dozens of downstream tasks. |
| SO018 | arXiv | PFMBench: Protein Foundation Model Benchmark | The field lacks a comprehensive benchmark for fair evaluation and in-depth understanding. |
| SO019 | bioRxiv | ProteinReasoner: A Multi-Modal Protein Language Model with Chain-of-Thought Reasoning for Efficient Protein Design | ProteinReasoner introduces a multi-modal protein language model with chain-of-thought reasoning for efficient protein design. |
| SO020 | bioRxiv | RNAGenesis: A Generalist Foundation Model for Functional RNA Therapeutics | RNAGenesis is a generalist foundation model for functional RNA therapeutics. |
| SO021 | bioRxiv | BioLab: End-to-End Autonomous Life Sciences Research with Multi-Agents System Integrating Biological Foundation Models | BioLab is an end-to-end autonomous life sciences research system integrating biological foundation models. |
| SO022 | GitHub | biomap-research organization repositories | The organization shows active repositories including ProteinSage, scFoundation, MorphDiff and other research assets updated through 2026. |
| SO023 | Fineline Insight | Kexing Biopharm Partners with BioMap on AI-Driven Macromolecular Drug Development | Kexing Biopharm partners with BioMap on AI-driven macromolecular drug development. |
| SO024 | Fineline Insight | Optoseeker Biotech and BioMap Collaborate to Accelerate Antibody Therapeutics Development | Optoseeker Biotech and BioMap collaborate to accelerate antibody therapeutics development. |
| SO025 | ByDrug / PharmCube | 百图生科(BioMap)与追光生物(Optoseeker Biotech)达成战略合作,以AI大模型赋能高通量单细胞筛选 | 双方将开发结合AI技术和高通量设备技术的抗体筛选智能体系统(AI Agent)。 |
| SO026 | AInvest | BioMap confidential HK IPO filing narrative-control analysis | |
| SM001 | Global Market Insights | Artificial Intelligence in Drug Discovery Market Size, Share – 2035 | The global artificial intelligence in drug discovery market was estimated at USD 3.1 billion in 2025 and is expected to grow from USD 4 billion in 2026 to USD 43.9 billion in 2035. |
| SM002 | Precedence Research | AI-Driven Drug Discovery Platforms Market Size, Report by 2035 | The market includes pure software SaaS/platforms, platform+wet-lab partnerships, CDMO/CRO integrations, data and annotation services, and related professional services sold to pharma, biotech, CROs, and research institutions. |
| SM003 | MarketsandMarkets | Drug Discovery Technologies Market by Product, Technology, Process and Therapeutic Area - Global Forecast to 2030 | The Drug Discovery Technologies market, valued at US$28.61 billion in 2024, stood at US$30.58 billion in 2025 and is projected to reach US$51.51 billion by 2030. |
| SM004 | McKinsey & Company | Generative AI in the pharmaceutical industry: Moving from hype to reality | Generative AI could offer the pharma industry a once-in-a-century opportunity—but only if they learn to scale it and address the industry’s unique challenges. |
| SM005 | IQVIA Institute | Global R&D Trends 2026 | Biopharmaceutical R&D remained resilient in 2025 ... while emerging developments in artificial intelligence-enabled discovery and development offer a tantalizing glimpse of a future in which reduced pipeline attrition dramatically improves R&D productivity. |
| SM006 | Recursion Pharmaceuticals | Investor Relations | Recursion Pharmaceuticals, Inc. | Recursion Pharmaceuticals is a clinical stage TechBio company decoding biology to radically improve lives. |
| SM007 | Recursion Pharmaceuticals | Pioneering AI Drug Discovery | Recursion | Recursion was founded on the idea that AI could understand the vast unknown biological space driving disease. |
| SM008 | Absci | Investor Relations | Absci Corp | Absci is a clinical-stage biopharmaceutical company advancing breakthrough therapeutics designed with generative AI. |
| SM009 | Absci | Home | Absci | We’re unlocking novel biology and creating better biologics with AI. |
| SM010 | Schrödinger | Physics-based Software Platform for Molecular Discovery & Design | Schrödinger’s computational platform, powered by physics, is transforming the way therapeutics and materials are discovered. |
| SM011 | Certara | Home | Certara empowers this evolution with predictive technologies that transform drug discovery and development. |
| SM012 | Certara | Investor Relations | Certara, Inc. | Its clients include more than 2,400 biopharmaceutical companies, academia and regulatory agencies. |
| SM013 | Insilico Medicine | Main | Insilico Medicine | Insilico presents Biology42, Medicine42, Science42 and generative AI software across target ID through Phase II. |
| SM014 | Harbour BioMed | Harbour BioMed - HBM Holdings | Harbour BioMed is a global biopharmaceutical company committed to novel antibody therapeutics in immunology, oncology and other areas. |
| SM015 | Hong Kong Investment Corporation | Hong Kong Investment Corporation Limited x BioMap Sign Strategic Partnership Agreement (June 2024) | BioMap has secured contracts with over 200 users ... including international pharmaceutical companies, leading CDMOs, innovative drug developers, synthetic biology, green technology enterprises, and research institutions. |
| SM016 | PR Newswire / Harbour BioMed | Harbour BioMed and BioMap Jointly Initiate MegaStream TechBio | BioMap currently supports over 800 institutional users worldwide, spanning multiple verticals including antibody and protein, innovative therapies and precision medicine, synthetic biology, and frontier scientific research. |
| SM017 | HKSTP | HKSTP Congratulates BioMap on Signing Strategic Partnership Agreement with HKIC | BioMap ... is collaborating with over 10 commercial partners and more than 200 academic institutions. |
| SM018 | Yicai Global | Baidu-Backed BioMap Reportedly Files for Hong Kong IPO | BioMap’s products include the BioMap OS ... which has been validated in over 60 projects ... BioMap has served over 800 global institutions and more than 30 leading enterprises. |
| SM019 | PMLive | Sanofi partners with AI specialist BioMap in deal worth more than $1bn | This approach enables superior prediction from limited data in immunology, neurology, oncology and rare diseases. |
| SM020 | pharmaphorum | Sanofi partners BioMap on AI hunt for biologic therapies | BioMap has built a biological map of proteins from public and private data sources and will develop AI models and LLMs to design and optimize new biologic drugs with Sanofi. |
| SM021 | BioMap | BioMap website bundle (about, product, technology and news strings) | BioMap OS, powered by xTrimo foundation models, is a dry-wet closed-loop life science discovery system. |
| SM022 | Tracxn | BioMap company profile | It operates as an AI-based precision medicine supporting tool for drug discovery. |
| SM023 | GitHub | biomap-research/xTrimoPGLM | The xTrimoPGLM family models are developed by BioMap and Tsinghua University. |
| SM024 | bioRxiv | RNAGenesis: A Generalist Foundation Model for Functional RNA Therapeutics | RNAGenesis is a generalist foundation model for functional RNA therapeutics. |
| SM025 | bioRxiv | BioLab: End-to-End Autonomous Life Sciences Research with Multi-Agents System Integrating Biological Foundation Models | BioLab is an end-to-end autonomous life sciences research system integrating biological foundation models. |
| SP001 | Recursion Pharmaceuticals | Pioneering AI Drug Discovery | Recursion | Recursion was founded on the idea that AI could understand the vast unknown biological space driving disease. |
| SP002 | Owkin | Owkin | Building Biological Artificial Superintelligence | Owkin is building the autonomous AI Scientist. |
| SP003 | Valo Health | This is Intelligent Health | Valo harnesses AI to find patterns in large-scale human data, identify novel disease targets and rapidly engineer novel small molecules. |
| SP004 | insitro | Making Medicines Differently - insitro | At insitro, we are building a different kind of drug company through the power of machine learning and data at scale. |
| SP005 | BenevolentAI | BenevolentAI | AI Drug Discovery | AI Pharma | Our next generation platform targets the complex decisions that drive R&D: life science intelligence at your fingertips. |
| SP006 | CompaniesMarketCap | Recursion Pharmaceuticals (RXRX) - Market capitalization | As of July 2026 Recursion Pharmaceuticals has a market cap of $1.75 Billion USD. |
| SP007 | CompaniesMarketCap | Absci (ABSI) - Market capitalization | As of July 2026 Absci has a market cap of $1.72 Billion USD. |
| SP008 | CompaniesMarketCap | Schrödinger (SDGR) - Market capitalization | As of July 2026 Schrödinger has a market cap of $1.22 Billion USD. |
| SP009 | CompaniesMarketCap | Certara (CERT) - Market capitalization | As of July 2026 Certara has a market cap of $1.05 Billion USD. |
| SP010 | SEC | EDGAR search results for Recursion 10-K filings | EDGAR lists Recursion annual-report filings including the 2026 filing. |
| SP011 | SEC | EDGAR search results for Absci 10-K filings | EDGAR lists Absci annual-report filings including the 2026 filing. |
| SP012 | SEC | EDGAR search results for Schrödinger 10-K filings | EDGAR lists Schrödinger annual-report filings including the 2026 filing. |
| SP013 | SEC | EDGAR search results for Certara 10-K filings | EDGAR lists Certara annual-report filings including the 2026 filing. |
| SP014 | Absci | Home | Absci | We’re unlocking novel biology and creating better biologics with AI. |
| SP015 | Certara | Home | Certara empowers drug development with predictive technologies. |
| SP016 | Certara | Investor Relations | Certara, Inc. | Its clients include more than 2,400 biopharmaceutical companies, academia and regulatory agencies. |
| SP017 | Insilico Medicine | Main | Insilico Medicine | Insilico presents Biology42, Medicine42, Science42 and generative AI software from target identification through Phase II. |
| SP018 | Schrödinger | Physics-based Software Platform for Molecular Discovery & Design | Schrödinger’s computational platform, powered by physics, is transforming the way therapeutics are discovered. |
| SP019 | Recursion Pharmaceuticals | Investor Relations | Recursion Pharmaceuticals, Inc. | Recursion is advancing a portfolio of differentiated investigational medicines across its wholly owned and partnered pipeline. |
| SP020 | Absci | Investor Relations | Absci Corp | Absci is a clinical-stage biopharmaceutical company advancing breakthrough therapeutics designed with generative AI. |
| SP021 | Harbour BioMed | Harbour BioMed - HBM Holdings | Harbour BioMed is a global biopharmaceutical company committed to novel antibody therapeutics in immunology, oncology and other areas. |
| SP022 | BioMap | BioMap website bundle (about, product, technology and news strings) | BioMap OS is a dry-wet closed-loop life science discovery system. |
| SP023 | Hong Kong Investment Corporation | Hong Kong Investment Corporation Limited x BioMap Sign Strategic Partnership Agreement (June 2024) | BioMap has secured contracts with over 200 users ... including international pharmaceutical companies, leading CDMOs, innovative drug developers, synthetic biology, green technology enterprises, and research institutions. |
| SP024 | PR Newswire / Harbour BioMed | Harbour BioMed and BioMap Jointly Initiate MegaStream TechBio | The alliance aims to launch MegaStream TechBio, a next-generation AI-native pipeline company targeting global markets. |
| SP025 | Tracxn | BioMap company profile | BioMap operates as an AI-based precision medicine supporting tool for drug discovery. |
| SI001 | HKIC | Hong Kong Investment Corporation Limited x BioMap Sign Strategic Partnership Agreement (June 2024) | To date, BioMap has secured contracts with over 200 users based on the AIGP platform. |
| SI002 | Yicai Global | Baidu-Backed BioMap Reportedly Files for Hong Kong IPO | BioMap OS has been validated in over 60 projects and BioMap has served over 800 global institutions and more than 30 leading enterprises. |
| SI003 | Sohu / 新识研究所 | 李彦宏创立的百图生科冲刺港股:中金、大摩、瑞银联手,去年CEO刘维表示未来一年半内寻求香港上市 | 累计融资金额超过2亿美元……潜在订单总额达到20亿美元。 |
| SI004 | Tencent News | 传百度支持的医药AI公司「百图生科」已秘密申请香港IPO,拟筹数亿美元,选定中金、大摩和瑞银合作 | 百图生科拟筹数亿美元,并已秘密申请香港IPO。 |
| SI005 | VCBeat / VBData | BioMap Completes $100 Million Series A Financing | BioMap today announced the completion of a Series A financing of 100 million dollars. The funds raised will mainly be used for technology research and development and talent introduction. |
| SI006 | ACN Newswire | Legend Capital invests in Series A funding round of BioMap, a biological computing platform | BioMap has recently completed the Series A funding round worth over a hundred million US dollars. The funds will be used for R&D and talent recruitment. |
| SI007 | pharmaphorum | Sanofi partners BioMap on AI hunt for biologic therapies | The alliance is getting underway with a $10 million upfront payment. |
| SI008 | PMLiVE | Sanofi partners with AI specialist BioMap in deal worth more than $1bn | BioMap will receive an upfront payment of $10m and will be eligible to receive over $1bn based on the achievement of pre-clinical development, clinical development, regulatory and commercial milestones. |
| SI009 | PR Newswire | Harbour BioMed and BioMap Jointly Initiate MegaStream TechBio | The founding parties will be entitled to potential upfront payments, success-based milestones, and royalty sharing in accordance with industry practice. |
| SI010 | HKSTP | HKSTP congratulates BioMap on signing strategic partnership agreement with HKIC (PDF) | The company has secured contracts with over 200 users based on the AIGP platform. |
| SI011 | Fineline Cube | Kexing Biopharm Partners with BioMap on AI-Driven Macromolecular Drug Development | Financial specifics of the partnership were not disclosed. |
| SI012 | Fineline Cube | Optoseeker Biotech and BioMap Collaborate to Accelerate Antibody Therapeutics Development | The partnership creates a streamlined workflow that can identify promising antibody candidates with unprecedented speed and precision. |
| SI013 | ByDrug / Pharmcube | 百图生科(BioMap)与追光生物(Optoseeker Biotech)达成战略合作 | 目前,该智能体已获得多家行业客户的试用意向,即将在百图生科苏州高通量实验中心部署,为全球客户提供实验服务。 |
| SI014 | Tracxn | BioMap - Funding & Investors | BioMap has raised an undisclosed amount of funding from 1 Series A round on Jul 30, 2021. |
| SI015 | AInvest | BioMap confidential HK IPO filing narrative-control play valuation governance risks | JS-only page retrieved during run; headline frames the filing around valuation and governance risks. |
| SI016 | CompaniesMarketCap | Recursion Pharmaceuticals (RXRX) - Revenue | Revenue in 2026 (TTM): $65.73 Million USD. |
| SI017 | CompaniesMarketCap | Absci (ABSI) - Revenue | Revenue in 2025 (TTM): $2.8 Million USD. |
| SI018 | CompaniesMarketCap | Schrödinger (SDGR) - Revenue | Revenue in 2025 (TTM): $0.25 Billion USD. |
| SI019 | CompaniesMarketCap | Certara (CERT) - Revenue | Revenue in 2025 (TTM): $0.41 Billion USD. |
| SI020 | SEC | EDGAR search results for Recursion 10-Q filings | EDGAR lists quarterly filings for Recursion. |
| SI021 | SEC | EDGAR search results for Absci 10-Q filings | EDGAR lists quarterly filings for Absci. |
| SI022 | SEC | EDGAR search results for Schrödinger 10-Q filings | EDGAR lists quarterly filings for Schrödinger. |
| SI023 | SEC | EDGAR search results for Certara 10-Q filings | EDGAR lists quarterly filings for Certara. |
| SI024 | BioMap | BioMap homepage | BioMap homepage was retrievable, but the live site exposes little readable pricing or revenue detail without JavaScript assets. |
| SI025 | Tech in Asia | Baidu-backed AI biotech BioMap seeks Hong Kong IPO | Fetched page resolved with minimal readable content, reinforcing how little primary IPO detail is publicly disclosed before the confidential filing surfaces. |
| SE001 | BioMap | BioMap homepage | BioMap describes itself as a pioneer in large biology language models. |
| SE002 | BioMap | BioMap portal bundle | BioMap OS is a foundation-model-driven dry-wet closed-loop life science discovery system; xTrimo V4 is described at 268B parameters and 300-plus SOTA results. |
| SE003 | BioMap | BioMap privacy policy bundle chunk | ISO/IEC 27001:2022 certificate in 2023; TLS 1.2 in transit; AES-256 at rest; annual third-party penetration tests. |
| SE004 | HKIC | Hong Kong Investment Corporation Limited x BioMap Sign Strategic Partnership Agreement (June 2024) | With the world’s largest 100Bn+ parameters biology foundation model “xTrimo”, we enable our partners to build their own AI models with limited data. |
| SE005 | HKSTP | HKSTP congratulates BioMap on signing strategic partnership agreement with HKIC (PDF) | The company has secured contracts with over 200 users based on the AIGP platform. |
| SE006 | Yicai Global | Baidu-Backed BioMap Reportedly Files for Hong Kong IPO | The technological backbone of BioMap is the foundational life sciences large language model xTrimo V4, which comprises 268 billion parameters. |
| SE007 | Sohu / 新识研究所 | 李彦宏创立的百图生科冲刺港股 | 2024年发布的生物大模型xTrimo V3,参数规模达2100亿,覆盖蛋白质、DNA、RNA等七大生命科学模态,在200余项任务中达到行业领先水平。 |
| SE008 | PR Newswire | Harbour BioMed and BioMap Jointly Initiate MegaStream TechBio | The system is expected to generate over 5 petabytes of high-quality, AI-ready life science data within five years. |
| SE009 | PMLiVE | Sanofi partners with AI specialist BioMap in deal worth more than $1bn | Matt Truppo said Sanofi combines BioMap’s protein large language models, high-performance computing, and deep understanding of AI with Sanofi’s data and drug-development expertise. |
| SE010 | pharmaphorum | Sanofi partners BioMap on AI hunt for biologic therapies | BioMap has built what is essentially a biological map of proteins using data sets from public and private sources to inform its foundational models. |
| SE011 | ByDrug / Pharmcube | 百图生科与追光生物达成战略合作 | The system will combine AI models and wet experiments into a seamless loop and will deploy at BioMap’s Suzhou high-throughput experimental center. |
| SE012 | Fineline Cube | Kexing Biopharm Partners with BioMap on AI-Driven Macromolecular Drug Development | The collaboration will focus on the full development process of macromolecular drugs, including the construction of AI-exclusive large models and AI intelligent laboratories. |
| SE013 | GitHub | GitHub - biomap-research/xTrimoPGLM | The xTrimoPGLM family includes 1B, 3B, 10B masked models, 1B, 3B, 7B causal models and 100B INT4 weights that can infer on a single 80G A100/800 GPU. |
| SE014 | arXiv | xTrimoPGLM: Unified 100B-Scale Pre-trained Transformer for Deciphering the Language of Protein | xTrimoPGLM was trained at 100 billion parameters and 1 trillion training tokens and outperformed baselines across 18 protein understanding benchmarks. |
| SE015 | GitHub | GitHub - biomap-research/PFMBench | PFMBench covers 38 downstream tasks and 17 pre-trained models, with fine-tuning and zero-shot evaluation support. |
| SE016 | arXiv | PFMBench: Protein Foundation Model Benchmark | PFMBench is a benchmark for protein foundation models. |
| SE017 | bioRxiv | ProteinReasoner: A Multi-Modal Protein Language Model with Chain-of-Thought Reasoning for Efficient Protein Design | The chain-of-thought reasoning embedded within the in-context learning framework is effective in characterizing combinatorial mutation landscapes. |
| SE018 | bioRxiv | RNAGenesis: A Generalist Foundation Model for Functional RNA Therapeutics | RNAGenesis extends BioMap’s foundation-model work into functional RNA therapeutics. |
| SE019 | GitHub | biomap-research organization | biomap-research has active repositories including ProteinSage, scFoundation, PFMBench and xTrimoPGLM, with updates through July 2026. |
| SE020 | bioRxiv | BioLab: End-to-End Autonomous Life Sciences Research with Multi-Agents System Integrating Biological Foundation Models | BioLab points to autonomous multi-agent life-science research integrating biological foundation models. |
| SE021 | GitHub | GitHub - biomap-research/scFoundation | scFoundation is a 100M-parameter pretrained model trained on over 50 million human single cells and published in Nature Methods. |
| SE022 | GitHub | GitHub - biomap-research/ProteinSage | ProteinSage is a protein foundation model built around explicit structural constraints. |
| SE023 | GitHub | GitHub - biomap-research/xTrimoMultimer | xTrimoMultimer optimizes protein-structure prediction for both monomer and multimer on GPU clusters. |
| SE024 | GitHub | GitHub - biomap-research/De-novoVHH | The De-novoVHH pipeline includes preprocessing, Rosetta scoring, filtering and coordinate relax steps for antibody design. |
| SE026 | BioMap | xTrimo Explorer | BioMap operates an xTrimo Explorer / AIGP surface, showing that the platform has a live public product endpoint beyond static marketing pages. |
| SE027 | GitHub | scFoundation apiexample at main · biomap-research/scFoundation | The old platform was discontinued on April 30th, 2024 and users were asked to migrate to the new platform at https://aigp.biomap.com/, which aims to provide online inference service and CLI tools. |
| SE025 | GitHub | GitHub - biomap-research/InverseFoldingEvaluation | This repository benchmarks inverse folding models for antibody CDR sequence design. |
| SU001 | HKIC | Hong Kong Investment Corporation Limited x BioMap Sign Strategic Partnership Agreement (June 2024) | BioMap has secured contracts with over 200 users based on the AIGP platform, including international pharmaceutical companies, leading CDMOs, innovative drug developers, synthetic biology, green technology enterprises, and research institutions. |
| SU002 | HKSTP | HKSTP congratulates BioMap on signing strategic partnership agreement with HKIC (PDF) | The company has secured contracts with over 200 users based on the AIGP platform. |
| SU003 | Yicai Global | Baidu-Backed BioMap Reportedly Files for Hong Kong IPO | BioMap has served over 800 global institutions and more than 30 leading enterprises. |
| SU004 | Sohu / 新识研究所 | 李彦宏创立的百图生科冲刺港股 | 客户包括赛诺菲、石药集团、大北农等行业龙头。 |
| SU005 | PR Newswire | Harbour BioMed and BioMap Jointly Initiate MegaStream TechBio | BioMap currently supports over 800 institutional users worldwide, spanning multiple verticals. |
| SU006 | PMLiVE | Sanofi partners with AI specialist BioMap in deal worth more than $1bn | The partnership combines BioMap’s AI platform with Sanofi’s capabilities to develop cutting-edge AI modules for biotherapeutic drug discovery. |
| SU007 | pharmaphorum | Sanofi partners BioMap on AI hunt for biologic therapies | Working with Sanofi, BioMap will develop AI models and large language models to design and optimise new biologic drugs. |
| SU008 | Fineline Cube | Kexing Biopharm Partners with BioMap on AI-Driven Macromolecular Drug Development | The collaboration aims to develop macromolecular drugs targeting tumors and autoimmune diseases using AI technology. |
| SU009 | Fineline Cube | Optoseeker Biotech and BioMap Collaborate to Accelerate Antibody Therapeutics Development | The partnership combines Optoseeker’s proprietary high-throughput cell screening technology with BioMap’s life-science AI. |
| SU010 | ByDrug / Pharmcube | 百图生科与追光生物达成战略合作 | The intelligent agent has already received trial interest from multiple industry customers and will be deployed in BioMap’s Suzhou experimental center. |
| SU011 | Harbour BioMed | Harbour BioMed - HBM Holdings | Harbour BioMed is a global biopharmaceutical company focused on novel antibody therapeutics. |
| SU012 | Sanofi | R&D-Driven and AI-Powered Biopharma Company | Sanofi | Sanofi presents itself as an AI-powered biopharma company. |
| SU013 | Kexing Biopharm | 科兴生物制药股份有限公司 | Kexing is an operating biopharmaceutical company, corroborating that the named BioMap partner is a real sector buyer. |
| SU014 | Optoseeker | 追光生物科技(深圳)有限公司 | Optoseeker is an operating life-science instrumentation company, corroborating BioMap’s instrumentation-partner segment. |
| SU015 | Syngenta | Syngenta Group | Syngenta is a global agriculture technology company. |
| SU016 | LanzaTech | LanzaTech | LanzaTech develops carbon-management and biotechnology platforms, matching BioMap’s green-tech customer narrative. |
| SU017 | Lilly | Eli Lilly and Company | Lilly is a global pharmaceutical company, matching the multinational pharma segment referenced by HKIC. |
| SU018 | CSPC | CSPC Pharmaceutical Group Limited | CSPC is a large pharmaceutical company, relevant to Sohu’s named-customer list. |
| SU019 | AInvest | BioMap confidential HK IPO filing narrative-control play valuation governance risks | JS-only page retrieved during run; headline frames BioMap’s listing around valuation and governance risks. |
| SU020 | BioMap | xTrimo Explorer | BioMap operates an xTrimo Explorer / AIGP surface, indicating a live user-facing product entry point. |
| SU021 | BioMap | BioMap portal bundle | BioMap OS has served over 800 institutional users and achieved validation in over 60 FIC discovery projects. |
| SU022 | BioMap | BioMap privacy policy bundle chunk | The privacy policy shows BioMap collects contact and organization data through forms and emails. |
| SU023 | Tech in Asia | Baidu-backed AI biotech BioMap seeks Hong Kong IPO | Fetched page resolved with minimal readable detail, underscoring sparse public disclosure before the prospectus. |
| SU024 | Tencent News | 传百度支持的医药AI公司「百图生科」已秘密申请香港IPO | Tencent also reported a confidential Hong Kong IPO targeting several hundred million dollars. |
| SU026 | Dabeinong Group | 大北农集团 | Dabeinong is a real operating agriculture and biotech group, making the Sohu client reference category-plausible even though deployment proof is still weak. |
| SU025 | scFoundation API example | scFoundation/apiexample at main · biomap-research/scFoundation | The old platform was discontinued and users were asked to migrate to the new AIGP platform, implying an active user base and product migration path. |
| SR001 | PCPD | The Personal Data (Privacy) Ordinance | The PDPO is one of Asia’s longest standing comprehensive data protection laws. |
| SR002 | BioMap | BioMap privacy policy bundle chunk | BioMap says it collects personal information through forms and emails, and claims ISO/IEC 27001:2022, TLS 1.2, AES-256 and annual third-party penetration tests. |
| SR003 | HKIC | Hong Kong Investment Corporation Limited x BioMap Sign Strategic Partnership Agreement | BioMap intends to leverage Hong Kong’s globally aligned IP and data regulations to garner greater trust from multinational clients. |
| SR004 | HKSTP | HKSTP congratulates BioMap on signing strategic partnership agreement with HKIC (PDF) | HKSTP described BioMap as a strategic enterprise signing with OASES and expanding in Hong Kong. |
| SR005 | Yicai Global | Baidu-Backed BioMap Reportedly Files for Hong Kong IPO | BioMap confidentially applied for an IPO in Hong Kong and uses xTrimo V4 with 268 billion parameters. |
| SR006 | Sohu / 新识研究所 | 李彦宏创立的百图生科冲刺港股 | 累计融资金额超过2亿美元,拟募集资金规模达数亿美元,未来仍需市场进一步检验。 |
| SR007 | AInvest | BioMap confidential HK IPO filing narrative-control play valuation governance risks | JS-only page retrieved during run; the headline frames the filing around valuation and governance risks. |
| SR008 | PR Newswire | Harbour BioMed and BioMap Jointly Initiate MegaStream TechBio | The system is expected to generate over 5 petabytes of high-quality AI-ready life-science data within five years. |
| SR009 | PMLiVE | Sanofi partners with AI specialist BioMap in deal worth more than $1bn | The Sanofi collaboration ties BioMap’s platform to milestone-heavy biotherapeutic discovery economics. |
| SR010 | pharmaphorum | Sanofi partners BioMap on AI hunt for biologic therapies | BioMap has built a biological map of proteins from public and private sources to inform its foundational models. |
| SR011 | ByDrug / Pharmcube | 百图生科与追光生物达成战略合作 | The Optoseeker workflow will deploy in BioMap’s Suzhou high-throughput experimental center. |
| SR012 | Fineline Cube | Kexing Biopharm Partners with BioMap on AI-Driven Macromolecular Drug Development | The collaboration includes AI-exclusive large models and AI intelligent laboratories across the drug-development process. |
| SR013 | BioMap | BioMap portal bundle | The bundle describes global multi-cloud deployment, autonomous diagnosis and self-recovery, high-throughput intelligent experiments and a Suzhou laboratory. |
| SR014 | BioMap | xTrimo Explorer | BioMap operates an AIGP / xTrimo Explorer surface, indicating a live user-facing product endpoint. |
| SR015 | scFoundation API example | scFoundation/apiexample at main · biomap-research/scFoundation | Users were asked to migrate to a new AIGP platform with online inference service and CLI tools. |
| SR016 | GitHub | GitHub - biomap-research/xTrimoPGLM | The xTrimoPGLM family exposes multiple model sizes and 100B INT4 inference artifacts. |
| SR017 | arXiv | xTrimoPGLM: Unified 100B-Scale Pre-trained Transformer for Deciphering the Language of Protein | xTrimoPGLM outperformed other advanced baselines in 18 protein understanding benchmarks. |
| SR018 | GitHub | GitHub - biomap-research/PFMBench | PFMBench is a unified benchmark suite for evaluating protein foundation models across 38 downstream tasks. |
| SR019 | arXiv | PFMBench: Protein Foundation Model Benchmark | PFMBench is a benchmark for protein foundation models. |
| SR020 | bioRxiv | ProteinReasoner | ProteinReasoner uses chain-of-thought reasoning within an in-context learning framework. |
| SR021 | bioRxiv | RNAGenesis | RNAGenesis extends the roadmap into RNA therapeutics. |
| SR022 | bioRxiv | BioLab | BioLab points to end-to-end autonomous life-sciences research with a multi-agent system. |
| SR023 | Sanofi | Our Science | Sanofi | Sanofi is pioneering a new era in immunology and advancing innovative therapies and vaccines. |
| SR024 | Harbour BioMed | Harbour BioMed - HBM Holdings | Harbour BioMed is committed to novel antibody therapeutics and global development. |
| SR025 | Optoseeker | 关于我们 | Optoseeker was founded in 2023 and is building a world-class functional single-cell analysis platform. |
| SR026 | Kexing | 关于科兴 | Kexing has multiple R&D centers, nearly 200 R&D staff and several technology platforms. |
| SR027 | Schrödinger IR | Schrödinger, Inc. - Financials - SEC Filings | Schrödinger provides a public SEC-filings surface for investors. |
| SR028 | Recursion IR | 404 Not Found | The requested SEC-filings URL returned a 404 page, highlighting how reference paths change and why direct diligence packs matter. |
| SR029 | Absci IR | Page Not Found | Absci Corp | The requested SEC-filings URL returned a page-not-found result. |
| SR030 | Certara IR | Page Not Found | Certara, Inc. | The requested SEC-filings URL returned a page-not-found result. |
| SV001 | Yicai Global | Baidu-Backed BioMap Reportedly Files for Hong Kong IPO | BioMap aims to raise several hundred million US dollars from the IPO. |
| SV002 | Sohu / 新识研究所 | 李彦宏创立的百图生科冲刺港股 | 累计融资金额超过2亿美元;潜在订单总额达到20亿美元。 |
| SV003 | HKIC | Hong Kong Investment Corporation Limited x BioMap Sign Strategic Partnership Agreement | HKIC will aggregate and channel resources to support BioMap’s development in Hong Kong. |
| SV004 | HKSTP | HKSTP congratulates BioMap on signing strategic partnership agreement with HKIC (PDF) | BioMap is one of the strategic enterprises signed with OASES and is expanding in Hong Kong. |
| SV005 | PR Newswire | Harbour BioMed and BioMap Jointly Initiate MegaStream TechBio | The founding parties will be entitled to potential upfront payments, success-based milestones, and royalty sharing. |
| SV006 | PMLiVE | Sanofi partners with AI specialist BioMap in deal worth more than $1bn | BioMap will receive an upfront payment of $10m and will be eligible to receive over $1bn in milestones. |
| SV007 | pharmaphorum | Sanofi partners BioMap on AI hunt for biologic therapies | The alliance is getting underway with a $10 million upfront payment. |
| SV008 | CompaniesMarketCap | Recursion Pharmaceuticals (RXRX) - Market capitalization | As of July 2026 Recursion Pharmaceuticals has a market cap of $1.75 Billion USD. |
| SV009 | CompaniesMarketCap | Absci (ABSI) - Market capitalization | As of July 2026 Absci has a market cap of $1.72 Billion USD. |
| SV010 | CompaniesMarketCap | Schrödinger (SDGR) - Market capitalization | As of July 2026 Schrödinger has a market cap of $1.22 Billion USD. |
| SV011 | CompaniesMarketCap | Certara (CERT) - Market capitalization | As of July 2026 Certara has a market cap of $1.05 Billion USD. |
| SV012 | CompaniesMarketCap | Recursion Pharmaceuticals (RXRX) - Revenue | Revenue in 2026 (TTM): $65.73 Million USD. |
| SV013 | CompaniesMarketCap | Absci (ABSI) - Revenue | Revenue in 2025 (TTM): $2.8 Million USD. |
| SV014 | CompaniesMarketCap | Schrödinger (SDGR) - Revenue | Revenue in 2025 (TTM): $0.25 Billion USD. |
| SV015 | CompaniesMarketCap | Certara (CERT) - Revenue | Revenue in 2025 (TTM): $0.41 Billion USD. |
| SV016 | CompaniesMarketCap | Recursion Pharmaceuticals (RXRX) - P/S ratio | Current and historical P/S ratio charts for Recursion Pharmaceuticals. |
| SV017 | CompaniesMarketCap | Absci (ABSI) - P/S ratio | Current and historical P/S ratio charts for Absci. |
| SV018 | CompaniesMarketCap | Schrödinger (SDGR) - P/S ratio | P/S ratio as of July 2026 (TTM): 4.79. |
| SV019 | CompaniesMarketCap | Certara (CERT) - P/S ratio | P/S ratio as of July 2026 (TTM): 2.54. |
| SV020 | SEC | EDGAR search results for Recursion 10-Q filings | EDGAR lists quarterly filings for Recursion. |
| SV021 | SEC | EDGAR search results for Absci 10-Q filings | EDGAR lists quarterly filings for Absci. |
| SV022 | SEC | EDGAR search results for Schrödinger 10-Q filings | EDGAR lists quarterly filings for Schrödinger. |
| SV023 | SEC | EDGAR search results for Certara 10-Q filings | EDGAR lists quarterly filings for Certara. |
| SV024 | AInvest | BioMap confidential HK IPO filing narrative-control play valuation governance risks | JS-only page retrieved during run; headline frames the filing around valuation and governance risks. |
| SV025 | Tracxn | BioMap - Funding & Investors | BioMap has raised an undisclosed amount from 1 Series A round on Jul 30, 2021. |
| SV026 | BioMap | BioMap portal bundle | BioMap OS has served over 800 institutional users and achieved validation in over 60 FIC discovery projects. |
| SV027 | BioMap | xTrimo Explorer | BioMap operates a live AIGP / xTrimo Explorer surface. |
| SV028 | Sanofi | R&D-Driven and AI-Powered Biopharma Company | Sanofi | Sanofi presents itself as an AI-powered biopharma company. |
| SV029 | Harbour BioMed | Harbour BioMed - HBM Holdings | Harbour BioMed is a global biopharmaceutical company focused on novel antibody therapeutics. |
| SV031 | CompaniesMarketCap | Recursion Pharmaceuticals enterprise value page | The requested enterprise-value page returned a 404, showing the limitations of web comp surfaces for some metrics. |
| SV032 | CompaniesMarketCap | Absci enterprise value page | The requested enterprise-value page returned a 404. |
| SV033 | CompaniesMarketCap | Schrödinger enterprise value page | The requested enterprise-value page returned a 404. |
| SV034 | CompaniesMarketCap | Certara enterprise value page | The requested enterprise-value page returned a 404. |
| SV030 | PCPD | The Personal Data (Privacy) Ordinance | The PDPO is one of Asia’s longest standing comprehensive data protection laws. |