Artificial intelligence is often used to generate questionable images, summarize emails, or help companies cut labor costs. Drug discovery is where I really want to see the technology succeed.
Bristol Myers Squibb is making a huge bet on that possibility. The pharmaceutical company plans to build what it calls the most powerful privately owned NVIDIA AI infrastructure in the life sciences industry.
The system will use an NVIDIA DGX SuperPOD equipped with DGX Vera Rubin NVL72 hardware. BMS says the additional computing power will help its researchers train proprietary AI models, analyze scientific data, automate repetitive work, and potentially shorten the time required to identify promising new medicines.
That is an ambitious claim, of course. Pharmaceutical research is expensive, slow, and filled with failures. Even a compound that looks promising early in development can collapse during testing or produce unexpected safety problems.
Still, I love seeing AI used in an effort to help humanity rather than simply sell more advertisements or replace another customer service worker. If powerful computers can help scientists discover treatments faster, eliminate weak candidates earlier, or better understand serious diseases, that is a use of AI worth rooting for.
BMS says NVIDIA’s Vera Rubin architecture can deliver up to 10 times more performance per megawatt than the previous generation. That efficiency could become especially important as AI systems grow larger and require increasing amounts of electricity and cooling.
The company is not starting from scratch. Bristol Myers Squibb has worked with NVIDIA for nearly three years and already uses DGX SuperPOD infrastructure across its research and development operations.
The new system will expand that foundation and support work in oncology, hematology, cardiovascular disease, immunology, and neuroscience.
“BMS has made a deliberate bet on AI, and we are beginning to see it pay off in our pipeline and operations,” said Greg Meyers, chief digital and technology officer at Bristol Myers Squibb. “We’re committed to translating AI into real outcomes for patients which requires infrastructure built to match that ambition. Expanding our compute capabilities with NVIDIA gives our researchers and teams across the business the scale they need to keep BMS at the leading edge of what AI can do for drug discovery and development.”
BMS says its scientists already use AI agents for tasks such as identifying and validating potential drug targets. According to the company, some of that work previously required weeks of manual effort.
That does not mean scientists are being removed from the process. Ideally, it means researchers can spend less time digging through data and more time evaluating whether an idea makes biological sense.
The company also uses an approach it calls Predict First. AI-generated predictions are considered before researchers begin experiments in the laboratory. BMS says the process now influences every small-molecule program and most of its large-molecule programs.
In theory, this could help researchers design smarter experiments and avoid spending time and money on candidates that have little chance of success.
“Drug discovery is a sequence of decisions made under uncertainty, and better decisions come from better evidence, faster,” said Robert Plenge, executive vice president and chief research officer at Bristol Myers Squibb. “This infrastructure lets us learn from every experiment and every clinical readout to sharpen the next hypothesis, allowing BMS scientists to spend less time on manual work and more time on the questions that require human judgment.”
BMS calls this combination of people and software hybrid intelligence. AI systems handle complex and data-intensive tasks, while human researchers decide which questions to ask, how to interpret the results, and whether a potential treatment deserves further development.
That sounds far more realistic than the usual suggestion that AI will somehow replace highly trained scientists. Computers can process enormous datasets and identify patterns, but they do not remove the need for laboratory testing, clinical trials, regulatory review, or human judgment.
The NVIDIA cluster will also support foundation models trained on decades of proprietary Bristol Myers Squibb data. BMS may also use NVIDIA BioNeMo, a collection of AI tools aimed specifically at biology and drug development.
“BMS has built decades of extraordinary scientific knowledge across some of the most complex areas of human disease,” said Rory Kelleher, senior director of business development for healthcare and life sciences at NVIDIA. “With NVIDIA Vera Rubin and BioNeMo Agent Toolkit, BMS has the ability to transform that enterprise scientific data into proprietary intelligence and give agents domain-specific tooling that help scientists explore biology, design molecules, and evaluate hypotheses at unprecedented scale.”
NVIDIA clearly benefits from convincing pharmaceutical companies that bigger AI systems will improve research. BMS also benefits from presenting itself as a technology leader. Both companies have reasons to promote the project aggressively.
The real test will not be the size of the cluster, the number of GPUs, or performance-per-watt claims. It will be whether the technology helps Bristol Myers Squibb develop safer and more effective medicines.
That answer may take years. Drug discovery cannot be measured on the same schedule as a software release or an AI benchmark.
Even so, this is the type of AI investment I want to see. The industry has spent plenty of time showing us chatbots, synthetic videos, and automated marketing tools. Using AI to help scientists fight cancer, neurological diseases, and other serious medical conditions feels far more meaningful.
Support independent tech journalism
NERDS.xyz is independently owned and operated. If you enjoy my coverage of Linux, AI, hardware, cybersecurity, and tech culture, consider supporting the site on Ko-fi.
Support NERDS.xyz