ByteDance has found a way to own a drug company without owning a drug. Anew Labs, the AI drug-discovery unit the TikTok parent spun out in June, has closed a $290 million round at a valuation of roughly $1.5 billion, according to Reuters, which cited two people familiar with the deal. ByteDance keeps 56 percent. Outside investors, led by HSG, the firm formerly known as Sequoia China, now hold the rest of a business that has four preclinical candidates, one open-source protein model, and, as of this week, no product in a human being.
The round was co-led by IDG Capital and Hillhouse Investment, with 5Y Capital as co-lead. Gaorong Ventures, Primavera Venture Partners and Boyu Capital also participated, alongside the state-backed Shanghai Future Industries Fund and SBP Group as a strategic investor. Neither Anew Labs nor ByteDance has publicly confirmed the figures. The South China Morning Post reported that Anew "did not immediately respond to a request for comment," and none of the lead investors have issued a statement. Everything known about the deal comes from unnamed sources.
What the money buys is a research operation that started inside ByteDance in 2021 under Liu Kai, who spent seven years in venture capital at IDG Capital and Fire Stone Investment before joining the company. IDG is now one of his lead investors. Roughly 50 staff moved to the new entity with the algorithms, platforms and pipeline assets, according to the 36Kr publication Intelligent Emergence, which first reported the spin-off. Anew's website listed 36 core team members in May. Volcano Engine, ByteDance's cloud arm, will keep supplying the compute.
The Stack
Anew's platform has four named pieces. AnewFold predicts the structure of proteins and molecular complexes. AnewSampling handles molecular dynamics. AnewDesign generates and optimizes antibodies. AnewMind is a large language model pitched as a reasoning assistant for drug-discovery decisions. The company also maintains Protenix, an open-source reproduction of Google DeepMind's AlphaFold 3 released under an Apache 2.0 license, and published a March preprint on AnewOmni, a generative framework trained on more than five million biomolecular complexes that claims to design functional molecules across small-molecule, peptide and nanobody scales.
The pipeline is narrower than the model catalog. Anew's disclosed programs target IL-17, the IL-4 receptor and two undisclosed targets, all oral small molecules aimed at immunology targets that today are served by injectable antibodies. The lead program is a pan-IL-17 inhibitor that Chris Li, Anew's head of biology, presented at the American Association of Immunologists meeting in Boston in April. Pharmaphorum reported that the company describes the IL-17F homodimer as a target that "has long been regarded as notoriously challenging for small molecules." That program is in lead optimization, which in drug-development terms means it has not yet been dosed in animals under formal safety studies, let alone filed for a clinical trial.
Why It Matters
The most useful way to read this deal is as an accounting event. Finimize put it plainly: a minority round like this "is less about headlines and more about price discovery." ByteDance has been operating a drug-discovery group for five years with no external mark on it. Now it has one. A 56 percent stake in a $1.5 billion company is worth about $840 million on paper, which is trivial next to ByteDance's own valuation of more than $600 billion but is not trivial as a proof point for the company's AI-for-science ambitions. The sources cited by Reuters said the separation was meant to support the unit's long-term development because drug discovery follows a different industry logic and management approach from ByteDance's core businesses. Translated: a consumer internet company's budget cycle does not fit a business where the first revenue may be a decade away.
The second reading is about the Chinese funding wave. SCMP reported the same week that Nutshell Therapeutics, a Shanghai AI small-molecule company, raised tens of millions of dollars in a C+ round backed by Citic Capital, Decheng Capital and the state-backed Fortune Venture Capital, and that AIPher, incubated at Tsinghua University, closed an angel round of roughly 100 million yuan. XtalPi, the sector's Hong Kong-listed bellwether, has been signing platform deals for years. Investors are explicitly framing these companies as domestic answers to Isomorphic Labs and Anthropic, which paid $400 million for Coefficient Bio earlier this year and this month landed Novo as a Claude customer for drug discovery. The state's presence on Anew's cap table, through the Shanghai Future Industries Fund, signals that the local government sees this as strategic rather than speculative.
The third reading is the skeptical one, and it is the one TNW landed on in May: "What it does not yet have is clinical data." More than 173 AI-discovered programs are now in clinical development worldwide, and the industry's roughly 90 percent clinical failure rate has not demonstrably moved. Insilico's rentosertib has positive Phase IIa data. Recursion discontinued its lead AI-discovered candidate after long-term data failed to hold up. Anew is behind both of them on the development curve. A $1.5 billion valuation for four preclinical programs and a model suite is a bet on the platform, not the pipeline, and platform bets in biotech have a long history of looking cheap right up until the first molecule fails in a body.
What to Watch
The near-term signals are procedural rather than scientific. Watch for an IND filing on the IL-17 program, which would be the first hard evidence that Anew can move a molecule from a conference poster to a regulated trial; the company's presentation calendar, which already ran through BIO in San Diego and the Free Energy Workshop in Barcelona, suggests it is courting pharma partners, so a licensing deal with a named counterparty would be the next credibility marker. Watch also whether ByteDance's 56 percent holds through the next round or gets diluted below control, which would tell you whether the parent views Anew as a core asset or a financial one. And watch the compute arrangement: Anew's dependence on Volcano Engine is a subsidy that outside investors are implicitly pricing in, and any change to that relationship would change the math on a company that, for now, is valued on what it might build rather than what it has.
"What it does not yet have is clinical data."— Allison Steffens Herrera, Reporter, The Next Web