The Departure
<cite index="1-4">Miles Wang, an OpenAI researcher whose work includes using artificial intelligence (AI) to accelerate scientific and biological discovery, is leaving the company to launch a new startup focused on developing AI models for drug discovery, according to four people with knowledge of his plans.</cite> <cite index="1-1,1-2">Several other OpenAI researchers are expected to join the new company, and Wang is in talks to raise about $200 million at a $2 billion valuation.</cite> <cite index="1-3">Lightspeed Venture Partners is in discussions to lead the funding round.</cite>
<cite index="1-8,1-9">Talks are ongoing and the deal may not be final; Wang disputed the story's funding figures and description of the company but did not specify the correct numbers or details.</cite> <cite index="1-10">Lightspeed did not respond to a request for comment.</cite>
Background on Wang
<cite index="1-17">Wang joined OpenAI in 2024 after dropping out from Harvard, where he was working on a bachelor's degree in computer science.</cite> <cite index="7-19,7-20">At OpenAI, he sat on the reinforcement-learning team, with work spanning alignment and safety as well as AI for science and biology.</cite> <cite index="7-21,7-22">He co-authored a wet-lab study in which GPT-5 suggested adjustments to a cloning protocol while humans ran the experiments; OpenAI later reported a 79-fold efficiency gain in that narrow setup, while stressing it produced no drug.</cite>
Proposed Product Focus
<cite index="1-15">Wang's new startup may be working on AI models that will help find new uses for existing drugs and possibly those that previously failed in trials, according to sources.</cite> <cite index="1-16">Finding new uses for Food and Drug Administration (FDA)-approved drugs can result in significantly faster time to revenue than developing new drugs from scratch, as these medicines have already been tested for safety.</cite> <cite index="12-4">The company is expected to apply transformer architectures — the same technology underlying large language models (LLMs) such as ChatGPT — to biological data and molecular prediction.</cite>
Sector Context
Wang's exit arrives amid a surge of capital into AI-driven life sciences. <cite index="1-12,1-13">Chai Discovery, a two-year-old startup developing AI models that can predict molecular interactions to identify new drugs, announced it raised $400 million at a $3.8 billion valuation; co-founder Josh Meier also passed through OpenAI as a researcher.</cite> <cite index="1-14">Meanwhile, Google DeepMind spinout Isomorphic Labs, which also develops AI models for drug discovery, raised a $2.1 billion Series B in May.</cite>
<cite index="11-17,11-18">Wang's $2 billion starting valuation stands out: most AI biotech startups launch at valuations between $200 million and $500 million, even with strong founding pedigrees.</cite> <cite index="11-19">The premium likely reflects both Wang's OpenAI credentials and Lightspeed's conviction that LLM techniques can transfer directly to protein modeling and molecular prediction.</cite>
Talent Exodus Pattern
<cite index="18-1">At least 12 senior executives left OpenAI in 2025 alone.</cite> Wang's departure is the latest instance of a well-documented pattern. <cite index="2-2">The most notable prior departure was a group of safety researchers who founded Anthropic, which has raised more than $7 billion.</cite> <cite index="11-3">The exodus reflects both opportunity and tension — OpenAI's focus on general AI leaves domain experts eager to pursue vertical applications where near-term commercial traction is clearer.</cite>
<cite index="6-9,6-10">While the valuation reflects high investor confidence, the drug discovery sector carries inherent risks: AI startups in this space often face long timelines before generating steady revenue, as their products must still undergo rigorous clinical trials and regulatory approvals regardless of the speed of initial discovery.</cite> <cite index="12-10,12-11">The funding discussions remain ongoing and the final terms are not yet set; Wang has not publicly named the company or detailed its product roadmap.</cite>