Anew Labs has raised $290 million in its first independent funding round, underscoring investor enthusiasm for AI drug discovery despite unresolved clinical challenges.
By Li Xinting
ByteDance founder Zhang Yiming, recently named Asia’s richest person, has another record-breaker on his hands.
Anew Labs, the AI drug discovery business spun out of ByteDance in June, has raised $290 million (2 billion yuan) in its first independent funding round, valuing the company at $1.5 billion. Based on publicly disclosed figures, it is the largest financing round in China’s AI drug discovery sector this year.
ByteDance retains a 56% stake and majority control. The star-studded investor list includes HSG (formerly known as Sequoia China), IDG Capital, Hillhouse Venture Capital, 5Y Capital, Gaorong Ventures and Primavera Capital, alongside pharmaceutical companies including Sino Biopharmaceutical.
The scale of the funding reflects a wider boom. Globally, venture capital firms have invested more than $40 billion in AI drug discovery over the past decade, according to PitchBook. In China, 117 of 164 AI drug discovery companies tracked by CVsource have received investment. By September, funding in the sector had exceeded 15.7 billion yuan this year — more than the previous three years combined.
Yet the excitement comes with a caveat: AI is accelerating drug research, but its ability to deliver effective treatments remains largely unproven.
A predictable spin-off
ByteDance has been dubbed an “AI factory”, with products ranging from the Doubao chatbot to the CapCut video editor. Pharmaceuticals might seem far removed from its core internet businesses, but the company has been developing AI drug discovery capabilities for several years.
Anew Labs originated in the AI for Science (AI4S) team within ByteDance’s Seed research division, established in 2023 to explore new approaches to artificial general intelligence.
The team developed models including Protenix and SeedFold for predicting biomolecular structures, and PXDesign and SeedProteo for designing protein binders. Former team member Gu Quanquan said SeedFold outperformed DeepMind’s AlphaFold 3 across a broad range of benchmarks. ByteDance also reported that SeedProteo outperformed competing systems, including DeepMind’s AlphaProteo and RFdiffusion, developed by a team led by Nobel chemistry laureate David Baker.
The company has begun exploring commercial applications. In April, it disclosed a preclinical small-molecule inhibitor targeting IL-17, a pro-inflammatory cytokine implicated in psoriasis and ankylosing spondylitis. ByteDance described it as the world’s first full-spectrum IL-17 small-molecule inhibitor, with potential applications in autoimmune diseases.
But developing medicines requires a very different business model from running an internet platform. Drug development typically takes around a decade and costs $1 billion, with companies facing expensive experimental work, clinical trials and commercialisation risks.
Spinning off Anew Labs allows ByteDance to pursue the opportunity while keeping control of the business and separating its operations from the parent company.
The move also follows an exodus of talent from the AI4S team. Media reports in June said Gu and colleagues, including Xiao Wenzhi, had left or were preparing to leave ByteDance to establish businesses focused on AI drug discovery and protein design.
Gu, formerly a tenured associate professor of computer science at the University of California, Los Angeles, led work on SeedFold and SeedProteo. On Sept. 10, he announced the launch of Geodesic Intelligence, a company aiming to develop general-purpose AI for drug discovery.
Investor enthusiasm meets clinical reality
The AI drug discovery boom has gathered pace since generative AI transformed text, video and software development. The sector promises to shorten a traditionally lengthy process: identifying a drug target, discovering and optimising compounds, conducting preclinical research and progressing through three phases of clinical trials.
Industry data published in 2023 suggest AI could reduce the time needed to identify a preclinical drug candidate from four or five years to eight to 12 months.
Investors have responded enthusiastically. Xaira Therapeutics, an AI drug discovery company founded by Baker, raised more than $1 billion in seed funding in 2024. Isomorphic Labs, spun out of DeepMind, raised $600 million in its first external round in 2025 and secured a further $2.1 billion in Series B funding in May this year.
OpenAI chief executive Sam Altman has also voiced ambitious expectations. In a 2025 blog post, he suggested that 10 gigawatts of computing power might enable AI to find a cure for cancer. In September, he argued that curing cancer was not ambitious enough a goal for AI.
But a review published in Nature Reviews Drug Discovery in August struck a more cautious note. Its 16 authors, from institutions including Cambridge University, Harvard Medical School and AstraZeneca, highlighted three major obstacles.
First, AI drug discovery remains concentrated in preclinical research. Only a few dozen programs have reached Phase I or Phase II trials, while Phase III projects remain scarce. Clinical efficacy, not the performance of a model, is the ultimate test of a medicine.
Second, reports of success rates approaching 90% in AI-assisted Phase I trials rely on relatively small datasets. Many of the drugs involved build on established disease biology and chemistry, making it difficult to attribute the results to AI alone.
Third, there is a fundamental gap between early decisions based on surrogate data and the clinical effects of medicines, which may take years to emerge. This makes AI’s actual contribution to successful drug development difficult to measure.
Dario Amodei, the co-founder and chief executive of Anthropic, has offered a more measured perspective. In October 2024, he argued that AI could enable personalised cancer treatments and potentially eliminate most cancers. By August 2026, however, he was more sceptical of the rhetoric surrounding the technology.
In an Aug. 15 post on social media platform X, Amodei said the claim that AI would cure cancer had become “more a cliché than it is inspiring, and most people think it is deceptive”.
“The thing that will work is actually curing cancer,” he said.
The message for investors is clear: AI may dramatically accelerate drug discovery, but the industry must still prove that faster research translates into effective treatments.
Source:
chinaventure.com.cn
