World’s largest Nvidia AI supercomputer for pharma sector to be built by US firm

US pharmaceutical giant Bristol Myers Squibb (BMS) has teamed up with chip maker NVIDIA to build the sector’s largest AI supercomputer. The computing cluster is expected to be online by January 2027 and will be hosted by data center company Equinix under a co-location agreement. The wave of artificial intelligence that is sweeping companies across the globe isn’t leaving out the pharmaceutical sector either. Although it is not the large language models (LLMs) that are helping accelerate drug discovery, the high-performance compute that powers these models is helping pharmaceutical companies build their own AI. Top leadership at these companies is now convinced that computing power is as critical to the drug discovery process as the wet laboratory. However, unlike tech companies in the Silicon Valley, which went all in with rented compute and are now paying full price for their token usage, big pharma companies took a different route and decided to own their compute instead. Why pharma companies own compute At a time when NVIDIA’s processors are excessively in demand and seeing regular performance upgrades, the decision to buy the compute instead of renting may not seem wise. However, renting compute isn’t that straightforward either. When the BMS team did the math, they realized that even when reserving capacity for compute meant that they were getting access to chips that were three to four generations old. Moreover, the cost of renting was much higher than that of purchasing the chips from NVIDIA outright. This seems to be the line of thinking at other pharma giants like Eli Lilly and Roche as well, who bought their compute in October 2025 and March 2026, respectively. BMS has worked closely with NVIDIA for three years when it installed the DGX SuperPOD for its research and development activities. While Eli Lilly compared the buy decision to having access to a ‘near infinite’ number of tokens. BMS is much more conservative in its way of thinking and is just optimizing its resources and isn’t shy about renting compute in the future. The larger benefit of buying the compute is that the company’s proprietary scientific data and intellectual property remain on the compute is owns. BMS’ AI Supercomputer The AI Supercomputer that BMS is working to build consists of an NVIDIA DGX SuperPOD built with DGX Vera Rubin NVL72 systems making it the most powerful single-owned NVIDIA infrastructure available within the life science industry. The system delivers 10x the performance per megawatt of the infrastructure that it is replacing while also giving BMS researchers access to NVIDIA’s BioNeMo Agent Toolkit, which can run predictions, train AI models, and power workflows across the drug discovery pipeline. “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 in a press release. “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.” Recommended ArticlesGet the latest in engineering, tech, space & science - delivered daily to your inbox.Ameya is a science writer based in Hyderabad, India. A Molecular Biologist at heart, he traded the micropipette to write about science during the pandemic and does not want to go back. He likes to write about genetics, microbes, technology, and public policy.