StoneWise (望石智慧)
Beijing AI small-molecule company: a 3D molecular-generation platform sold to pharma, plus an in-house HPK1 inhibitor in Phase I.
1. Core Product / Service
StoneWise was founded in 2018 and is based in Beijing [1]. Its core is a 3D small-molecule generation model. Around it the company has built:
- a multi-agent system for early-stage R&D;
- a "digital compass" data system;
- an integrated dry/wet R&D platform [4].
It sells AI-driven molecular design, custom model training and compound screening to pharma companies and research institutes. It also runs its own pipeline [3].
Pipeline.
- SWA1211. An oral HPK1 inhibitor designed with the company's generative model. The company says it received clinical clearance in both China and the US in June 2025 [3]. China's trial registry (CTR20251571, first posted 19 May 2025) lists a first-in-human, open-label, dose-escalation Phase I in advanced solid tumors, planning 60 patients [5][6].
- The company says it is advancing more than ten internal projects, some of them clinical [3]. Only SWA1211 is identified publicly.
Wet-lab model. StoneWise says it runs its own standardized chemistry lab, which synthesizes AI-generated molecules, tests their activity and drug-likeness, and feeds the results back into its models [3]. It also cites a physical compound library and a joint lab with Peking University, set up in 2022 and later designated a Beijing key laboratory for intelligent drug R&D [3]. Lab size, headcount and CRO use are not disclosed. These are company statements.
2. Target Users & Pain Points
- Pharma and biotech medicinal-chemistry teams that want faster hit finding and lead optimization. The company claims hit cycles of 1–6 months and 70–80% lower R&D cost [3] (self-reported). It cites more than 100 partners, including several top-20 global pharma companies, Qilu and Chipscreen [3].
- Domestic-compute buyers. In May 2026 StoneWise signed a three-way agreement with Guangzhou Pharmaceutical Holdings and Huawei. Huawei supplies a private deployment on domestic compute, and the pharma partner supplies real R&D use cases [4].
3. Competitive Landscape
| Player | Positioning | Vs. StoneWise |
|---|---|---|
| xtalpi | Chinese AI + robotics discovery platform, listed | Much larger wet-lab automation footprint |
| insilico-medicine | Chinese AI-native pipeline company, listed | More clinical assets; also licenses molecules |
| schrodinger | Physics-based design software plus partnered pipeline | Similar "software + own assets" hybrid, without a wet lab |
| iambic-therapeutics | US AI small-molecule biotech with automated loops | Pipeline-first; StoneWise is platform-first |
4. Unique Observations
- It has moved toward platform, not away from it. By 2026 StoneWise calls itself a "pharma AI infrastructure provider" offering foundation models and full-chain agents [4]. Many peers instead pivot from services to pipeline. SWA1211 serves as a showcase that the platform can reach the clinic.
- Funding has been quiet since 2021. The only public large round is the US$100M Series B/B+ in April 2021 [1]. Searches found no later disclosed financing, which may explain the platform-revenue emphasis.
- SWA1211 sits in a crowded HPK1 field. BeiGene, Treadwell and Glenmark all have HPK1 programs in early clinical trials [5].
5. Financials / Funding
| Round | Date | Amount | Lead |
|---|---|---|---|
| Early rounds | 2019–2020 | not disclosed ("four rounds in the past year" as of Apr 2021) [1] | — |
| Series A | Mar 2020 | ~US$10M | Longhill Capital; Linear Capital participated [2] |
| Series B | Apr 2021 | B + B+ combined US$100M | Legend Capital (B) [1] |
| Series B+ | Apr 2021 | (combined above) | Lightspeed China Partners and Greater Bay Area Homeland Development Fund (B+) [1] |
- Total raised (disclosed): about US$110M or more.
- Revenue and valuation: not disclosed.
6. People & Relationships
- Founder & CEO: 周杰龙 (Zhou Jielong). He holds AI bachelor's and master's degrees from Beijing Institute of Technology, has over 20 years in AI algorithms, and was formerly a chief architect at Baidu [1][3].
- Academic partner: Peking University, through a joint lab; a 2023 Nature Machine Intelligence paper came out of it [3].
- Investors: Legend Capital, Lightspeed China Partners, Longhill Capital, Linear Capital.
- Peers: mindrank, metis-techbio, helixon-therapeutics, revir-therapeutics.
Sources
- Yicai Global (Series B/B+) — https://www2.yicaiglobal.com/news/lightspeed-greater-bay-area-fund-pour-usd100-million-into-drug-ai-startup-stonewise
- TechNode (Series A) — https://technode.com/2020/03/16/ai-drug-development-platform-raises-10-million-in-series-a/
- Pharnexcloud (company interview) — https://www.pharnexcloud.com/zixun/shiye/zjgd_293471
- 36Kr (2026 positioning) — https://www.36kr.com/p/3866854221698313
- AllSci (SWA1211 Phase I) — https://allsci.com/news/chinas-stonewise-puts-novel-hpk1-inhibitor-to-first-in-human-test/
- Pharnexcloud trial record CTR20251571 — https://www.pharnexcloud.com/data/lcsy_070f7147a77170ce87027c15995602e1.html