Although the announcement was subtle and was withdrawn during a panel discussion at Davos, its impact sent tremors through the startup ecosystem. OpenAI is exploring a “shared value” model, demanding a share of the intellectual property (IP) customers generate using its AI technology for scientific breakthroughs. The proposal moves beyond common API usage fees and directly toward profit sharing, signaling a fundamental shift in the way basic AI labs monetize their vast computational power.
OpenAI Chief Financial Officer Sarah Friar, speaking on a panel moderated by The Information CEO Jessica Lessin at Davos, suggested that OpenAI could potentially “license medicines discovered using OpenAI’s technology” in areas such as drug discovery. This concept of trading compute for equity, or a portion of the ultimate revenue stream, represents OpenAI’s active attempt to capture the exponential financial benefits of the innovations that OpenAI’s model enables, rather than simply being satisfied with linear revenue growth tied to token consumption. This is central to reimagining generative AI providers as strategic, high-stakes investors rather than public utilities.
The cost of running large-scale AI models for round-the-clock research is astronomical. Drug discovery in particular requires large-scale, specialized computing (often referred to as “agents”) to analyze huge data sets and accelerate clinical timelines. Traditionally, biotech companies lacking internal resources have raised large sums of money from venture capitalists and public investors and given up equity to acquire the computing power needed to hire these agents. OpenAI’s shared value model completely removes this intermediary. By providing compute directly in exchange for a profit share, OpenAI converts its AI capabilities into a form of capital, a non-dilutive (cash) investment, with returns tied directly to discovery success.
The initial reaction of many in the technology community was skepticism, wondering which companies would agree to such a request. However, this model is not without precedent, especially in areas where speed and resources are paramount. The video references a 2018 partnership between pharmaceutical giants GSK and 23andMe, where genetic insights are being harnessed for new drug development. Furthermore, proactively acquiring IP in exchange for access to resources is already commonplace, although often controversial, in deep research environments. Elite academic institutions, like Stanford University, often have strict patent policies, claiming ownership of “potentially patentable inventions created in a process that goes beyond individual university responsibilities, participation in Stanford research projects, or incidental use of Stanford resources.” This institutional model establishes a clear precedent for centralizing intellectual property rights around the providers of core resources, in this case computing power.
For drug discovery companies facing long timelines, overwhelming regulatory hurdles, and high failure rates, the ability to use cutting-edge proprietary AI to accelerate research can outweigh the cost of giving up some future profits. Alternatives may see slower growth, rely on traditional slow funding cycles, and be overtaken by competitors adopting AI-driven acceleration. In the high-stakes world of life sciences, acting quickly to cure disease is often the best and safest bet, despite the capital cost.
This strategy is closely related to the concept of artificial general intelligence (AGI) and the scarcity of computing. The race to use scarce computing to secure high-value IP sets the stage for a dramatic concentration of wealth and power. This concern is not limited to OpenAI. “Rivals such as Anthropic, Google DeepMind, and Isomorphic Labs, an Alphabet subsidiary focused on using AI for drug discovery, are also in talks with early-stage biotech startups about data licensing and partnerships.”With the current scarcity of advanced GPUs and the massive capital needed to train and run frontier models, entities that control computing infrastructure have enormous influence. By leveraging this scarcity to acquire intellectual property rights, these fundamental laboratories are positioned to capture the full economic value of AGI-driven breakthroughs across multiple industries, cementing their dominant position.
The immediate impact of this shared value model is clear. Intellectual property is rapidly becoming the currency exchanged for access to the most powerful discovery engines. For early-stage biotech companies lacking the large capital reserves needed to participate in the computer arms race, this agreement could be a strategic necessity to accelerate efforts. But in the long run, innovation and economic power will be concentrated in a small number of organizations that can provide and control the world’s most advanced computational resources.
