前OpenAI首席产品官兼科学团队负责人凯文·韦尔创办Warp Sciences，目标是把人工智能与机器人引入药物研发实验室，帮助生物科技和制药公司更快制造、测试候选化合物。韦尔宣布，公司已完成2.15亿美元融资，领投方为Lightspeed；公司估值据称达到8.65亿美元。IVP、Premji Invest、Mayfield、Lux Capital和Menlo Ventures等也参与投资。
Warp Sciences计划在美国建立化学合同研究机构，为药物研发团队提供分子合成和测试服务。公司首阶段聚焦“先导化合物优化类似物合成”，也就是围绕有潜力的分子制造不同变体，再通过实验比较其性质，帮助研究人员筛选更适合继续开发的候选药物。公司表示，客户决定研究哪些分子，并保留对药物项目的控制权；Warp不打算开发自有药物管线。
为训练机器人完成实验室工作，化学研究人员将佩戴摄像设备和传感手套，示范并记录实验操作。公司希望逐步用机器人执行这些任务，让实验室能够持续运行；每轮实验产生的数据再用于改进AI模型和后续实验方案，形成“设计、执行、分析、再设计”的循环。自动化程度和实际效率仍有待公司后续建设和验证。
这家初创公司的业务也瞄准美国生物科技产业对海外研发服务的依赖。药物研发中，批量合成数千种分子变体属于耗时且劳动密集的环节，相关工作长期由中国等地的合同研究机构承接。Warp希望在美国提供替代服务，减少供应链地缘政治风险，并回应药企对知识产权保护的顾虑。
Warp Sciences由韦尔与三名联合创始人共同创立：曾任Terremoto Biosciences药物化学负责人索尔·赖斯伯格、Meta人工智能研究实验室FAIR联合创始人兼前高管马诺哈尔·帕卢里，以及曾任美国首席数据科学家的DJ·帕蒂尔。公司称，团队将结合AI、机器人和药物化学经验，建设本土实验室基础设施，让实体实验数据反过来推动AI辅助药物研发。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Kevin Welle, former Chief Product Officer and Head of Scientific Team of Open AI, founded Warp Science with the goal of introducing artificial intelligence and robotics into drug research and development laboratories and helping biotechnologies and pharmaceutical companies to manufacture and test candidate compounds more quickly. Wehr announced that the company had completed its financing of $215 million, with the recipient being Lightseed; the company ' s valuation allegedly amounted to $865 million. IVP, Premji Invest, Mayfield, Lux Capital and Menlo Ventures are also involved.
Warp Science plans to establish a chemical contract research institute in the United States to provide molecular synthesis and testing services for drug research and development teams. The company's first focus on “preparating the synthesis of analogues” means the manufacture of different variants around potential molecules and then, by comparing their nature with experiments, helping researchers to screen candidates for more suitable drugs for further development. The company indicated that the client decided which molecules to study and retained control over the drug project; Warp did not intend to develop its own drug pipeline.
To train robots to complete laboratory work, chemical researchers will wear camera equipment and sensory gloves to demonstrate and document experimental operations. The company hopes to gradually use robotics to carry out these tasks, allowing the laboratory to operate on a continuous basis; the data generated from each round of experiments are then used to improve AI models and follow-up experimental programmes, creating a cycle of “design, implementation, analysis, re-engineering”. The degree of automation and practical efficiency still need to be followed up and validated by companies.
The start-up company also targets the dependence of the United States biotech industry on overseas R & D services. In drug research and development, mass synthesis of thousands of molecular variants is a time-consuming and labour-intensive process that has long been carried out by contracted research institutes in China and elsewhere. Warp hopes to provide alternative services in the United States to reduce geopolitical risks in the supply chain and to respond to the concerns of pharmaceutical companies about intellectual property protection.
Warp Sciences was created by Weil together with three co-founders: Saul Riceberg, former head of drug chemistry at Termo Biosciences; Manohar Paluri, co-founder and former senior director of the FAIR laboratory of Artificial Intelligence Research at Meta; and DJ Pattil, former chief data scientist in the United States. According to the company, the team will build indigenous laboratory infrastructure in conjunction with AI, robotics and pharmaceutical chemistry experience, so that physical experimental data will in turn drive AI-aided drug development.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.