Summary
Bristol Myers Squibb (BMS) is buying a new Nvidia DGX SuperPOD built on the latest Vera Rubin architecture. This will be the first time a life sciences company uses this advanced AI system. The new setup will help BMS speed up drug discovery and development by running larger AI models and making predictions about potential medicines. The purchase shows how drug companies are relying more on powerful computers to find new treatments faster.
Main Impact
BMS is getting a major boost in computing power to support its AI work in drug research. The company already has an older Nvidia SuperPOD that has been running for about three years. But that system is now working at full capacity. The new Vera Rubin system will give BMS up to 10 times more performance for each unit of electricity used. This means scientists can run more complex models and test more drug candidates without waiting for access.
Key Details
What Happened
BMS is buying eight DGX Vera Rubin NVL72 systems from Nvidia. Each system combines Nvidia Vera central processing units and Rubin graphics processing units. The company will use this infrastructure to train its own AI models and run predictions across all its research programs. The system will handle work involving compounds, proteins, and other scientific data. BMS did not say how much it paid for the system.
Important Numbers and Facts
The new system will replace an older SuperPOD that is two or three generations behind. BMS has been using its current system for about three years. The company says AI already helps design every small-molecule program and most large-molecule programs. AI has cut some manual research work by several weeks. BMS has also used AI to shorten the time needed to produce medicines for clinical trials by 20% to 30%, and hopes to reach 50% in the coming years.
Background and Context
Drug companies are turning to AI to make research faster and cheaper. Finding new medicines usually takes many years and costs billions of dollars. AI can help by predicting which molecules might work as drugs before scientists have to make them in a lab. This saves time and money. BMS has a method called "Predict First" that uses AI to rule out molecules that do not have the right properties. This way, researchers only test the most promising candidates. The company has also used AI to expand its library of CELMoD compounds, which are designed to destroy cancer-causing proteins.
Public or Industry Reaction
BMS executives are excited about the new system. Greg Meyers, the company's chief digital and technology officer, said computing needs have grown as BMS uses larger AI models. Erin Davis, vice president of research business insights and technology, said the current system is running at full capacity because of demand for large-molecule predictions and internal model development. Robert Plenge, chief research officer, said the new system will let scientists test more drug candidates. "Maybe before we could do 10 and now we can do dozens," he said. Payal Sheth, senior vice president of therapeutic discovery sciences, said the AI predictions help prioritize which molecules to make in the lab.
What This Means Going Forward
The new system will connect BMS research sites around the world. Data from one site, like Lawrenceville, New Jersey, can be used by teams in San Diego. This shared environment will help the company keep knowledge from experiments and programs. BMS also plans to use the system for digital-twin applications, though it did not give details. The company says the system will be more energy-efficient, which matters because electricity costs are rising. BMS did not say when the new system will be ready or where it will be located.
Final Take
BMS is making a big bet on AI to speed up drug discovery. The new Nvidia system will give its scientists more computing power and better access to AI tools. This could help the company find new medicines faster and at lower cost. But human researchers will still make the final decisions about which drugs to pursue. The move shows how AI is becoming a key part of modern drug development, not just a side project.
Frequently Asked Questions
What is the Nvidia DGX SuperPOD Vera Rubin system?
It is a powerful computer system made by Nvidia. It uses the latest Vera Rubin architecture, which includes new central processing units and graphics processing units. The system is designed to run large AI models and handle complex calculations. BMS is the first life sciences company to buy this system.
How will BMS use AI in drug discovery?
BMS uses AI to design small-molecule and large-molecule drugs. The technology helps identify targets for new drugs, optimize lead compounds, and predict which molecules will work best. AI also helps the company screen potential drug candidates before making them in a lab. This saves time and resources by focusing only on the most promising options.
What is the "Predict First" method at BMS?
"Predict First" is BMS's approach to using AI predictions before making molecules in a lab. The system uses computer models to check if a molecule has the right properties to become a drug. Molecules that do not meet the requirements are set aside. This way, researchers only synthesize and test the candidates with the highest chance of success.
