Qiming Venture Partners Backs CataX AI Seed Round to Rebuild Clinical Drug Development
Fundraising

Qiming Venture Partners Backs CataX AI Seed Round to Rebuild Clinical Drug Development

Much of the attention in AI-driven pharmaceuticals has gone to discovering new targets and designing molecules. A young Shanghai company is betting that the bigger opportunity now sits further down the pipeline, in clinical trials.

CataX AI has completed a seed financing round with Qiming Venture Partners as the sole investor, according to a report by Investment Circle republished by 36Kr. The size of the round was not disclosed. Founded in July 2026, the company describes itself as an AI-native biotech that uses software agents to accelerate the early clinical development of new drugs.

A Founder Who Moved Between Medicine and Computing

CataX AI was founded by Wang Zilong, born in 1991 in Qinhuangdao, Hebei. As a high school student he won top national prizes in biology, physics and informatics, but chose to enroll in Fudan University's eight-year clinical medicine program, eventually earning a doctorate in oncology.

He reasoned that computer science could be self-taught while medicine required formal training, and he audited computing courses alongside his medical studies. After graduating, he became CTO and Chief Medical Officer of AI medical company VoxelCloud at 28, leading a prospective study across 155 centers in China with more than 47,000 subjects and securing one of the country's early NMPA Class III medical AI device certificates.

He was named to the Forbes 30 Under 30 and Hurun U30 lists, and in 2023 joined Microsoft Research Asia as a senior researcher focused on medical foundation models and autonomous agents. His related research earned more than 20,000 stars on GitHub.

Why Agents Prompted a New Company

The decision to start CataX AI followed a series of agent products released earlier this year that changed how Wang worked, to the point where he rarely reviewed his own code line by line. He began asking what a drug developer would look like if agents were built in from the start.

Clinical development today relies on more than ten functions, including medical, regulatory, operations, data and biostatistics, which must coordinate even on routine documents. Wang summed up his approach this way: "Medicine determines why we set off, and AI determines how we should proceed."

Qiming's Thesis on the Clinical Bottleneck

Wang was introduced to Chen Kan, Partner and Co-Head of Healthcare Innovation at Qiming Venture Partners, through a Fudan alumnus this spring, before he had even finished a business plan. Qiming was an early backer of AI drug discovery in China, including Insilico Medicine.

Chen's view is that AI is now relatively mature in target discovery and molecular design, and the industry's constraint has moved downstream. Early discovery accounts for roughly 20 percent of a new drug's cost and time, with clinical development taking the remaining 80 percent. China now produces 44 percent of the world's preclinical candidate molecules, creating demand for a cheaper route to human proof of concept.

Chen argued that rebuilding trials around agents is the way to lower costs, and that starting fresh is easier than reforming established pharmaceutical organizations. Qiming co-incubated the company and helped recruit co-founder and Chief Development Officer Jiang Zhemin, who has more than a decade of clinical operations experience at Tigermed and BeiGene.

The Clinical-of-One Operating Model

CataX AI plans to acquire rights to preclinical and early clinical assets through in-licensing or co-development, push them to early proof of concept, and then out-license promising programs to partners able to handle later development and commercialization.

Each project is run under what the team calls a Clinical-of-One model: a single clinical development lead coordinates agents and outside experts. Agents assist with literature research, document drafting and data processing, while humans make the key decisions.

The company acknowledges that AI cannot compress biological timelines such as a year-long patient follow-up. Its efficiency claims rest instead on sharper asset selection and less internal friction.

Scaling Beyond a Handful of Experts

The main near-term challenge is growth. Wang has noted he cannot replicate himself, and project leads need both deep clinical knowledge and fluency with AI, a rare combination.

His answer is to capture organizational experience in a shared system, including expert feedback, the reasoning behind protocol changes and go or no-go decisions, and the eventual outcomes. The goal is to let new leads build on that record and reduce reliance on a few cross-disciplinary specialists. The team is also still reworking processes around technical systems, risk control and regulatory requirements.

What the Deal Signals for AI Biotech

The round reflects a broader shift in where investors see AI adding value in pharma, from finding molecules to moving them efficiently through human trials. Chen also pointed to growing overseas interest in the quality of Chinese drug assets, with some promising early programs stalled for lack of funding.

For founders, CataX AI is an early test of whether agent-first organizations can work in a heavily regulated field. Its success will depend less on the technology alone than on proving that lean, AI-assisted teams can deliver credible clinical evidence and connect with hospitals, partners and regulators.

Source: eu.36kr.com

Topics:

JeffkomStory Team

Editorial, JeffKom Story

The JeffKom Story newsroom covers startups, founders, funding and the technology shaping what’s next.

Keep reading

Malaysia Studies Mandatory WhatsApp Identity Checks as Scammers Shift Tactics
Fintech

Malaysia Studies Mandatory WhatsApp Identity Checks as Scammers Shift Tactics

More in Fundraising