The U.S. is cutting health funding at precisely the moment when artificial intelligence is transforming drug discovery, and industry leaders say this timing could cost America its scientific edge. At Fortune Brainstorm Tech in Aspen, executives from companies building AI drug discovery systems warned that while Washington pulls back, competitors worldwide are doubling down on AI-powered medicine research.

What's Driving the Urgency Around AI Drug Discovery Funding?

The AI drug discovery market is growing fast. Today it's worth $3.25 billion, expanding at roughly 26% annually and projected to reach $10 billion by 2031. Major funding rounds underscore the momentum: Isomorphic Labs, backed by DeepMind founder Demis Hassabis, raised $2.1 billion in a Series B earlier this year. Lila Sciences, a Flagship Pioneering spinout, has raised $550 million to build what it calls scientific superintelligence, AI systems that run the scientific method continuously across materials, chemistry, and life sciences.

The concern from industry leaders is straightforward: other countries are investing aggressively while the U.S. retreats. "If we defund now, while the rest of the world leans in, which Europe is leaning in, which a lot of the Asian countries are leaning in, we will be left behind," stated Kimberly Powell, vice president of healthcare at NVIDIA. Geoffrey von Maltzahn, co-founder and CEO of Lila Sciences, framed it as a matter of national competitiveness: "Falling below the scientific intelligence of one's adversary at a corporate level, at a sovereign level, is almost like an unimaginable competitive disadvantage".

How Are AI Systems Actually Changing Drug Discovery?

The structural similarity between how AI agents work and how the scientific method operates is central to the pitch. Both follow the same pattern: pose a question, gather context, observe, reason, and act. This alignment suggests that AI isn't just a tool for drug discovery, it's a natural extension of how science itself works.

Real-world results are already emerging. Lila Sciences' AI agents recently identified catalysts for splitting water into hydrogen and oxygen that outperform the precious metals the industry currently relies on. Notably, one-third of the AI's suggestions initially made no sense to the company's Caltech-trained chemists. Those counterintuitive suggestions turned out to be the highest-performing catalysts on record. This pattern suggests AI can discover solutions that human intuition alone might miss.

Von Maltzahn described this moment as transformative: "I think to many people this Claude Code-esque moment, when a new intelligence gets injected into science and changes it forever thereafter, may not feel imminent. But it is right around the corner".

Von Maltzahn

What Infrastructure Does AI Drug Discovery Actually Need?

Building the foundation for AI drug discovery requires more than just funding. NVIDIA's role illustrates the infrastructure challenge. The company invests in open-source biology foundation models, antibody design models, and multimodal models that companies like Lila adapt for specific scientific goals. Powell emphasized why this matters: "If it's only closed models that exist, there's not a lot of ability to create the conditions that all of these applications across life sciences can thrive in the age of AI".

Powell

The infrastructure needs include:

• Open-Source Foundation Models: Publicly available AI models trained on biology data that companies can customize for their own drug discovery pipelines without building from scratch.
• Multimodal AI Systems: Models that can process multiple types of data simultaneously, such as protein structures, chemical properties, and clinical outcomes, to make more informed predictions.
• Computational Resources: Massive computing power to run AI agents continuously, simulating thousands of experiments in silico before any physical lab work begins.

What's the Biggest Remaining Obstacle to FDA Approval?

One critical unresolved question looms: how do regulators approve a drug designed by AI when the training data and reasoning behind the AI's decision can't be fully traced? This is not a minor issue. It's a regulatory bottleneck that could slow AI-designed medicines from reaching patients.

Powell outlined a potential path forward: a future where digital twin models of biology become precise enough that regulators accept in silico evidence, meaning computational simulations would carry the same weight as physical experiments. "We're just not there yet," she acknowledged. This suggests the field is still years away from a regulatory framework that fully embraces AI-designed therapeutics.

Powell

How Can Organizations Prepare for the AI Drug Discovery Era?

For companies, research institutions, and policymakers, several practical steps emerge from the current landscape:

• Invest in Foundation Model Access: Ensure your organization has access to open-source biology AI models rather than relying solely on proprietary systems, which limits flexibility and innovation.
• Build Computational Infrastructure: Develop or partner for the computing resources needed to run AI agents continuously, as this is foundational to modern drug discovery workflows.
• Prepare Regulatory Strategies: Begin conversations with regulators now about how AI-designed molecules will be evaluated, rather than waiting until a candidate drug is ready for approval.
• Recruit Interdisciplinary Teams: Hire scientists who understand both traditional chemistry and AI systems, since the highest-performing discoveries come from humans and machines working together.

The broader message from industry leaders is clear: the window for the U.S. to maintain leadership in AI drug discovery is closing. The technology is advancing rapidly, capital is flowing globally, and the scientific method itself is being automated. Cutting health funding now, they argue, is betting against the very moment when that investment would pay the highest dividends.