Merck KGaA's Supercomputer for AI and Data Research
Merck KGaA has rolled out a new high-performance computing (HPC) platform to push AI and data-heavy modeling across life science, healthcare and electronics. Built by Lenovo and hosted in an AI-ready Equinix data centre in Munich, the system combines liquid cooling with a hybrid cloud setup to handle complex workloads at scale.
Laura Matz, Chief Science and Technology Officer at Merck KGaA, says: "At Merck KGaA, Darmstadt, Germany, we believe that data and technology are the cornerstones of scientific progress. With this high-performance computer, we are enhancing our computational capabilities and transforming how we conduct research and develop solutions that matter.
"This platform empowers our teams to leverage AI and advanced analytics more effectively, ultimately accelerating the pace of innovation across our sectors."
Why this HPC matters
AI-scale research needs infrastructure that can train, simulate and iterate without bottlenecks. Merck's system supports generative AI, molecular modeling and materials research for next-gen semiconductors on a unified backbone shared by multiple teams.
The hybrid model blends private and public cloud resources. That lets teams scale during peak demand while keeping sensitive research data under strict governance.
- Life science: Speed up drug discovery workflows
- Manufacturing: Shorten development cycles in life science production
- Electronics: Model novel materials for future semiconductor devices
Architecture at a glance
- Core hardware: Lenovo systems integrated with Neptune liquid cooling for sustained high utilization and lower energy use.
- Thermals at scale: As AI workloads raise rack heat and density, direct liquid cooling improves efficiency and stability.
- Location: Equinix International Business Exchange data centre in Munich with established liquid cooling and low-latency interconnection.
- Hybrid connectivity: Private systems interconnect with public cloud AI services, research partners and data sources for consistent performance.
- Global readiness: Equinix reports that 100+ sites are equipped for liquid cooling to host large GPU clusters and other high-performance systems.
Andreas Thomasch, Director & Chief Technology Officer (CTO), Germany and Austria at Lenovo, explains: "At Lenovo, we are committed to enabling the global adoption of innovative and sustainable technologies. By designing and configuring this solution together, we've demonstrated that organisations can achieve exceptional performance without compromising on sustainability. With Lenovo Neptune Liquid Cooling, even the most demanding workloads will run efficiently while helping Merck drive progress in both science and industry."
A digital foundation for AI convergence
Equinix contributes more than space and power; its facilities act as hubs for distributed AI infrastructure. Merck can place compute close to data, interconnect privately with clouds and partners, and keep performance steady while staying consistent with its ESG commitments.
Harmeen Mehta, Chief Digital and Innovation Officer at Equinix, says: "This partnership embodies what the future of innovation looks like - where digital infrastructure, compute, and science converge to solve humanity's biggest challenges." She noted that scaling technology responsibly can accelerate progress and redefine what's possible across industries.
Practical takeaways for IT, dev, and research teams
- Unify compute across labs and teams, with clear data boundaries and access controls.
- Use a hybrid approach: burst into public cloud for peak training or simulation, keep sensitive datasets on private infrastructure.
- Plan thermal design early. Liquid cooling is moving mainstream as GPUs push rack heat and density.
- Place compute near data. Leverage carrier-dense colocation to lower latency and egress to cloud services and partners.
- Track performance per watt and total cost of ownership. High sustained utilization beats chasing peak specs.
Learn more
Explore Lenovo Neptune Liquid Cooling for HPC and AI workloads: Lenovo Neptune.
If you're building skills in AI infrastructure, MLOps and HPC foundations, see our curated learning paths: Latest AI courses at Complete AI Training.
Your membership also unlocks:
