More Than 3,500 Reasons Why Roche is Winning the Race for AI in Drug Discovery
Pharmaceutical giant “Roche” is making a major push into artificial intelligence, investing heavily in cutting-edge Computing power to speed up how new drugs are discovered and developed. The company recently revealed plans to build a large-scale “AI factory” powered by thousands of NVIDIA’s advanced chips, a move that could significantly reshape AI Drug Discovery.
A Massive Investment in AI Infrastructure
At the center of this effort are 2,176 of NVIDIA’s latest Blackwell GPUs, which will be installed across Roche’s operations in the United States and Europe. With this addition, Roche now has more than 3,500 GPUs in total, something it says is the largest AI (Artificial Intelligence) Computing capacity in the Pharmaceutical industry, further strengthening its capabilities in AI Drug Discovery.
The announcement was made alongside Nvidia’s GTC conference in San Jose, highlighting just how important AI has become in the race to develop new treatments. Roche hopes this Computing power will help streamline everything from early-stage discovery to manufacturing and commercialization of therapies, especially as AI Drug Discovery continues to evolve.
Genentech’s Role and Real-World Impact
Although Roche is headquartered in Switzerland, much of its research strength comes from its U.S. subsidiary, Genentech. The Biotechnology firm is known for innovative drugs such as Avastin (a medication for lung and colon cancer) and Herceptin (a medication for breast cancer), which have been widely used in Cancer care. Overall, Roche generated about $78 billion in revenue last year.
The company has not disclosed exactly how much it’s spending on this AI expansion, but it’s clearly significant. NVIDIA CEO Jensen Huang has previously stated that Blackwell chips cost between $30,000 and $40,000 each. That would put Roche’s investment at roughly $65 million or more just for the hardware.
Roche isn’t alone in betting big on AI. Earlier this year, Eli Lilly announced a $1 billion partnership with NVIDIA to build an AI innovation lab in California. Across the industry, both major drugmakers and startups are increasingly convinced that AI could transform how medicines are created, with AI Drug Discovery at the center of this shift.
How AI Is Changing Drug Discovery
The potential payoff is huge. Developing a novel drug can take over a decade and cost around $1 billion. If AI can shorten that timeline, even by a small margin, it could mean faster access to life-saving treatments and a significant financial advantage for companies investing in AI Drug Discovery.
Roche says it’s already seeing results. In one case, its AI tools helped Scientists design a complex Cancer-related molecule 25% faster than usual, and with a structure that wouldn’t have been possible using the traditional methods. In another significant project, AI was used to predict whether a treatment might trigger an unwanted immune response, allowing Researchers to adjust it early before human trials.
Collaborations and the Future of AI in Pharma
Beyond its in-house work, Roche is also collaborating with startups. Last year, it partnered with Manifold Bio in a deal worth up to $2 billion to explore new ways of delivering drugs to the brain using AI.
According to Aviv Regev, who leads Research and early development at Genentech, the biggest shift isn’t just using AI in one part of the process; it’s integrating it across the entire process. Instead of treating AI as a specialized tool, companies are now embedding it across the entire pipeline, from discovery to production, something that will likely define the next phase of AI Drug Discovery.
This innovative shift reflects a broader belief across the industry: AI isn’t just an add-on anymore, it’s becoming central to how modern medicine is built. And as technology continues to advance, many expect it to unlock innovations that were once thought out of reach.
