OpenAI Researcher Miles Wang Planning $2B AI Drug Discovery Startup

What happened

Miles Wang, a prominent researcher at OpenAI, is reportedly in the process of departing the artificial intelligence leader to establish a new venture focused on the intersection of generative AI and biotechnology. The new startup aims to leverage advanced machine learning models to accelerate the drug discovery process, a field that has seen increasing overlap with large-scale AI research. According to sources familiar with the matter, the startup is currently in discussions to raise significant capital, aiming for an initial valuation that would immediately place it among the most valuable early-stage firms in the AI-biotech space. This move marks another high-profile departure from OpenAI as researchers seek to apply foundational model breakthroughs to specific industrial domains.

What's new in this update

Current reports indicate that Wang is in active negotiations to secure approximately $200 million in funding for the venture, which is expected to carry a valuation of around $2 billion. Venture capital firm Lightspeed is purportedly in discussions to lead this substantial investment round. While Wang has disputed the exact financial figures and certain descriptions of the company’s focus, the move signals a high level of investor confidence in his technical expertise. Furthermore, the startup is expected to draw additional talent directly from OpenAI’s current research roster, suggesting a concentrated effort to port foundational AI methodologies into the specialized domain of life sciences, specifically targeting the pharmaceutical industry’s long-standing research and development bottlenecks.

Key details

One of the strategic pillars of the new startup may involve using AI models to identify novel applications for existing pharmaceutical compounds. By focusing on drugs that have already received FDA approval or those that previously stalled in clinical trials for reasons unrelated to safety, the company could significantly shorten the timeline to commercialization. Developing new drugs from scratch is notoriously time-consuming and expensive; repurposing existing ones offers a faster path to revenue since much of the safety and toxicity testing is already completed. Wang’s previous work at OpenAI involved co-authoring research on how AI can automate scientific discovery, providing a strong theoretical framework for this practical application in pharmacology. His background as a Harvard computer science student before joining OpenAI further underscores the trend of venture capital betting on young, technical founders.

Background and context

The emergence of Wang’s startup follows a pattern of high-profile AI researchers pivoting to the biotechnology sector. Recently, Chai Discovery—co-founded by fellow former OpenAI researcher Josh Meier—announced a $400 million funding round at a $3.8 billion valuation. Similarly, Google DeepMind’s spinout, Isomorphic Labs, secured a $2.1 billion Series B earlier this year. These developments reflect a broader industry trend where the architectural breakthroughs found in large language models are being adapted to predict molecular interactions and protein folding. Investors are increasingly comfortable backing technical founders who possess deep expertise in machine learning architectures, even if they lack traditional pharmaceutical backgrounds, provided they can disrupt the traditional drug discovery lifecycle through computational speed.

What to watch next

As negotiations with Lightspeed and other potential investors continue, the final structure and valuation of the deal remain subject to change. The industry will be closely monitoring the specific technological approach Wang’s team takes, particularly whether they will release open-source models or maintain a proprietary drug-development pipeline. The potential for a larger migration of talent from general-purpose AI labs like OpenAI toward specialized vertical startups is also a key dynamic to watch. If successful, this venture could validate the thesis that the next frontier for generative AI lies in highly regulated, high-stakes industries where the ability to parse complex biological data provides a distinct competitive advantage over traditional trial-and-error laboratory methods.