In 2025, more than one dollar in every three dollars of potential transaction value in the global biopharmaceutical industry was related to Chinese enterprises.

近日，在大国新药全球会议上，安永发布的《中国创新药全球合作与交易观察》白皮书，尽显中国创新药的崛起之势。而当商务拓展交易越来越多，新问题也逐渐凸显。

比如，潜在大额交易越来越多，生物科技公司实际拿到的款项也在增加吗？借助人工智能研发平台，生物药开发合作交易也越来越多，未来还会更多吗？人工智能平台重新定义了新药研发，是否也会影响生物药开发合作交易本身的规则？

针对这些问题，《每日经济新闻》记者（下称每经记者）采访了安永大中华区咨询服务主管合伙人、大中华区生命科学与医疗健康行业联席主管合伙人吴晓颖。

吴晓颖  受访者供图

BD交易总金额突破了千亿美元大关，但行业依然存在实际兑现率较低的问题。

China’s innovative pharmaceuticals have already established themselves at the core of the global transaction stage.

White paper reveals that the global biopharmaceutical partnership authorization transaction potential total amount reached a record high in 2025, with the proportion of Chinese-related transactions in the potential amount of U.S. and European cooperation transactions surging from 8% in 2020 to 34%; in 2025, the total scale of Chinese enterprises' outward authorization transactions reached 135.7 billion U.S. dollars, and the compound annual growth rate of the potential amount of related transactions over the past five years reached 71%.

Future, this figure will keep growing. As the core buyers, the world's leading pharmaceutical enterprises are facing severe patent cliff pressures, and it is expected that by 2032 the top 25 biopharmaceutical companies will form a growth gap of 370 billion U.S. dollars. The industry holds 2.1 trillion U.S. dollars of M&A cooperation funds, continuously increasing the acquisition of external innovative assets. Tumor, central nervous system, metabolism GLP-1, ADC (antibody drug conjugate), bispecific antibody, and AI R&D platform have become the core arenas in the capital competition.

2016—2025年中国生物制药合作交易潜在金额与交易数量 图片来源：白皮书

But the reporter from Every Economic News noticed that beneath the prosperous BD data many biopharmaceutical companies are facing the actual dilemma of heavy signing and light realization: after numerous domestic innovative drugs complete overseas licensing a large number of them the overseas clinical advancement is slow and the milestone payments land less than expected with paper high transaction value difficult to convert into real benefits.

In the industry this situation does exist. Wu Xiaoying stated that after an asset transaction it simply enters the pipeline pool of the multinational pharmaceutical enterprise. Behind this step generally requires competition for budget clinical resources and management attention from other internal projects of the counterparty. If the strategic priority of the partner is adjusted or better data on similar products emerge then the project may slow down or even be terminated.

而且，还有一些问题在签约前就埋下了。比如部分项目国内早期数据看起来不错，但在试验设计、生产工艺或者数据标准等方面，与全球开发的要求还有距离。合作方接手以后，如果需要补试验、改方案，那就会让推进速度受到影响。另外，双方如果对适应证优先级、投入节奏、成功标准等方面的理解不一致的话，也会导致后面的摩擦变多。

In the view of Wu Xiaoying, when enterprises select partners, they should not only compare which one offers a higher upfront payment but also assess the significance of the pipeline to the counterparty, as well as whether the partner possesses matching capabilities in clinical development and commercialization. The contract should also clearly outline key terms such as development investment, critical milestones, joint decision-making, data sharing, and mechanisms for equity recovery, thereby reducing the risk of long-term underperformance after asset delivery.

AI-related transaction investments have grown 256%, and the appeal of one-off deals will decline.

From the perspective of the therapeutic field structure, the most fiercely contested therapeutic fields by leading biopharmaceutical enterprises in recent years have been long dominated by tumor indications. Since 2023, the competition in the endocrine/metabolic therapy field, including GLP-1 and other anti-obesity products, has also heated up rapidly.

然而根据白皮书并购支出并不局限于这些热点比如受2025年1月宣布的全年最大生物制药交易拉动2025年中枢神经系统CNS领域的收购金额已超过肿瘤领域的投资规模

2021—2025年生物制药并购支出的治疗领域分布 图片来源：白皮书

更值得关注的是，2025年生命科学行业涉及AI资产的潜在交易金额同比上涨了约90%。其中，约90%的生物制药AI投资流向了研发平台。为了获取AI技术平台，各企业大幅增加了对合作授权交易的投入，AI相关交易投入同比增长256%，达到496亿美元。

吴晓颖告诉每经记者，这类交易的估值逻辑比传统的药物授权更加复杂。传统交易通常会看单个资产的临床阶段、成功概率、适应症空间和未来销售潜力；AI平台还需要看平台的持续产出能力，比如它会不会持续产出有潜力的候选分子，预测结果和实验结果是否一致，研发周期能缩短多少等。

Additionally, in the contract design for AI platforms, besides upfront payments, research and development funds, and milestone payments, data usage rights, the attribution of model improvement results, and the allocation of newly generated intellectual property are typically also involved.

吴晓颖认为在未来几年这类合作还会继续增加但合作要求和交易方式会发生变化在前一阶段药企更多的是广泛试用平台和工具在经过前期验证以后接下来的合作会越来越集中到比较明确的治疗领域和具体的项目上药企会要求AI平台对候选分子的质量和实验验证效率方面承担更清晰的责任。

另外，交易方式也会发生变化。以往那种一次性购买软件或者单纯支付服务费的模式，吸引力会逐渐下降；而围绕着靶点以及候选化合物展开的共同研发、分阶段授权等方式，预计会逐步增加。吴晓颖解释称，药企会希望先验证平台能力并控制风险，再根据项目的进展去决定要不要继续增加投入。

In the future, AI may well increase clinical success rates for new drugs, but its current effects on BD transactions have yet to be validated.

The data above do not reflect the actual effect of AI on clinical success rates for new drugs. However, it is altering BD transactions.

The data do not reflect the actual effect of AI on clinical success rates for new drugs. However, it is altering BD transactions.

The data do not reflect the actual effect of AI on clinical success rates for new drugs. However, it is altering BD transactions.

The data do not reflect the actual effect of AI on clinical success rates for new drugs. However, it is altering BD transactions.

目前，AI资产的BD交易非常火热，生物制药行业密切关注AI缩短新药研发周期、降低研发成本的潜力，但吴晓颖认为市场对AI短期提高新药临床成功率的期待偏高，需要更多长期数据验证。

相比之下，AI对于企业数据积累和组织学习将产生重要影响，每一个研发项目、每一次实验以及每一笔交易都可能反过来训练企业的判断体系，时间越长，这个价值就会越大，但这还没有得到足够的重视。

白皮书还提醒称，从短期价值看，更直接的机会或许在于利用人工智能来改进交易过程本身。

The AI process to enhance M&A efficiency and execution quality (see figure). Source: white paper.

吴晓颖也观察到这一变化。她告诉每经记者，AI较早应用的场景之一是标的筛选。过去团队会主要靠行业数据库、人脉和人工检索去找标的，但现在可以同时分析论文、专利、临床试验、监管记录、公司公告和历史交易，更快地找到技术路线相近、或者尚未被市场充分关注的资产。

到了尽调阶段，AI直接的价值在于它能够帮助团队快速形成问题清单。通过将临床数据、专利边界、竞品进度、关键人员经历以及公开披露的信息放在一起进行交叉核对，它可以找出前后不一致的数据缺口以及其他潜在风险。这样专家就可以把时间集中在真正影响交易判断的问题上。

In the valuation stage, AI will help improve the speed of scenario calculations. Scenario calculations also require estimating various variables, such as clinical success rates, launch timelines, market shares, pricing and milestone realization paces, and AI models can quickly assist in simulating changes in different variables, thereby updating asset values under different scenarios and making it easier to compare with similar transactions and comparable pipelines.

吴晓颖认为，AI短期比较直接的作用在于改善效率，而长期则有望逐步影响企业发现资产、判断风险以及配置资源的方式。这部分价值需要随着数据和项目经验的积累，逐渐显现出来。

“当然，最终的判断还是需要医学、研发、商业和财务团队共同完成。”吴晓颖补充道。

封面图片来源：每经媒资库

来源：创新药BD签约容易，拿钱难？对话安永大中华区吴晓颖：AI正在重构BD交易全流程 | 每日经济新闻