AI infrastructure

Uncountable announced that Merck KGaA, Darmstadt, Germany selected its platform to digitalize its central analytical laboratories; Uncountable says the solution will improve data access to enable ML- and LLM-powered use cases.

Partners: Uncountable

BMS will expand its NVIDIA DGX SuperPOD with DGX Vera Rubin NVL72 systems to train proprietary foundation models, building on nearly three years of NVIDIA collaboration.

Partners: NVIDIA

Sanofi's internal platform for building AI, used by its data and AI developers.

Sanofi is moving AI workflows directly onto its Snowflake data using Cortex AI and the Elementum workflow platform, pitched as AI replacing some applications.

Partners: Snowflake, Elementum

An enterprise platform on Databricks that lets over 10,000 employees build, deploy and run AI agents under central governance, including a control tower, policy-enforcement agent and agent marketplace.

Partners: Databricks

Innovent's 2025 ESG report says protective measures implemented in 2024 for the WeMol computing platform and AI intelligence repository were followed by a security baseline for the electronic lab notebook system covering the R&D computing environment.

Roche added 2,176 NVIDIA Blackwell GPUs on premises in the US and Europe, bringing its combined on-premise and cloud footprint to more than 3,500 GPUs, as the next phase of an NVIDIA collaboration begun in 2023.

Partners: NVIDIA

Lilly's NVIDIA-built supercomputer for AI-driven research came online in early 2026 to support discovery and other scientific work.

Partners: NVIDIA

A self-service platform that lets employees build their own chatbots on Amazon Bedrock foundation models. AWS reports about 2,500 chatbots and 25,000 employees using it.

Partners: AWS, Cloud2 Oy

Vendor case study: Weights & Biases says GSK and Microsoft collaborated on AIGA, an internal framework for generative AI and agentic solutions.

Partners: Microsoft, Weights & Biases

Takeda is the founding partner in TetraScience's program to re-platform scientific data for AI use in R&D and CMC.

Partners: TetraScience

At its AI Day, BioNTech highlighted InstaDeep's new near-exascale supercomputer, Kyber, with the aim of enabling BioNTech to compute at scale. The release describes this as the introduction of a new near-exascale supercomputer.

Amgen installed an NVIDIA supercomputer called Freyja at its deCODE genetics subsidiary to train AI models on its very large store of de-identified human data for research.

Partners: NVIDIA

At the 2024 J.P. Morgan Healthcare Conference, Celltrion said its CEO shared plans to expand into digital healthcare with a company-owned data bank. That bank would use artificial intelligence to synthesize Celltrion's clinical and genomic data.

AbbVie's ARCH platform centralizes R&D data and hosts chemists' predictive models, which AbbVie says are rerun on 30-day cycles; AbbVie also says its generative AI platform is being embedded into ARCH.

Bayer and Google Cloud announced a collaboration that applies Google Cloud TPUs, described as accelerators designed to run machine-learning models and computationally intensive workloads, to accelerate and scale Bayer's quantum chemistry calculations.

Partners: Google Cloud