The 'OpenAI Moment' for Pharma Labs: C12.ai Raises ¥50M as Robot Scientists Step onto the Lab Bench
Inside BeOne Medicines' lab, C12.ai's robot system Talos has been running for 50 days, completing over 3,000 experiment operations with zero errors. C12.ai closes a new ¥50M round led by Yonghua Capital, bringing AI from the digital world into real laboratories.
Inside the laboratory of global leading innovative pharmaceutical company BeOne Medicines, a robot has been running for 50 days. During this time, it completed over 3,000 experiment operations — from molecular purification to complex experimental workflows — maintaining zero errors throughout, demonstrating proficiency approaching that of a senior medicinal chemist. This is an industry first.
This is Talos, a robot system developed by Yunqi-backed C12.ai. Through proprietary sensors and a custom VLA (Vision-Language-Action) architecture, Talos can understand experimental workflows, autonomously complete complex operations in real laboratory environments, and perform real-time reasoning and error correction during continuous experiments.
Recently, C12.ai completed a new ¥50M funding round led by Yonghua Capital, with participation from Weihao Chuangxin and individual investors. Yunqi Partners led the company's initial round in August 2022.
Founded in 2022, C12.ai is working to bring AI from the digital world into real laboratories — enabling robots not only to "understand science" but to actually perform experiments hands-on.
Breaking "Rigidity": From Repetitive Labor to Intelligent Substitution
For medicinal chemists, molecular purification is a formidable challenge: it consumes approximately 50% of their working time and demands extensive operational experience and "feel." Laboratory automation has long been stuck in a rigid "one machine, one task" phase — equipment is expensive, cannot be reused across scenarios, and fails at the slightest process variation.
C12.ai's Talos is effectively breaking through these limitations. As a general-purpose dual-arm mobile robot, Talos can operate existing instruments like a human — without any lab modifications.
The core breakthrough lies in its proprietary VLA (Vision-Language-Action) architecture, which gives the robot a "brain" — no longer mechanically executing commands, but understanding objectives. In real-world deployment at BeOne Medicines, Talos demonstrated proficiency approaching that of a medicinal chemist with 10 years of experience. With the robot handling 80% of routine purification work, scientists can finally step away from tedious physical operations and focus on more creative research decisions.
A Pharma Veteran's Cross-Industry Exploration: Software-Defined "Physical Scientists"
This precision targeting of industry pain points stems from the technical depth and industry insight of C12.ai founder and CEO Chen Zhigang.
As an industry veteran, Chen previously served as Chief Architect at Alibaba Health, founded Tencent's Medical Big Data Lab, and was WuXi AppTec's first Chief Digital Officer (CDO). He once used algorithms to raise traditional production equipment utilization from 40% to 60%, but through hands-on experience realized that pharma's efficiency bottleneck lies not in computing power, but in connecting to the physical world.
"Traditional automation solved standardization, but flexibility has always been its weakness," Chen said. "What we're building is a general-purpose, software-defined robot scientist."
This "software-defined hardware" approach gives C12.ai exceptional scenario migration capabilities. Beyond pharmaceuticals, Talos shows strong deployment potential in "small-batch, multi-lot" physical scenarios such as flexible cosmetics manufacturing — from scanning a photo to identify formulations to overnight production and direct delivery, physical AI is reshaping the closed loop of production relationships.
The Future: Making Research More Efficient
With this funding round complete, C12.ai plans to continue product iteration and gradually expand application scenarios. The company's technology roadmap extends along the drug development pipeline: from current purification experiments, further into ADME (pharmacokinetic) experiments, and eventually covering GxP-compliant manufacturing and analytical processes.
"We're not building a single-point tool — we want robots to gradually participate in more stages along the 'life of a drug,'" said Chen. By having AI handle large volumes of repetitive experimental operations, C12.ai aims to improve overall drug development efficiency, allowing researchers to focus more on innovation and decision-making.