Merck KGaA, Darmstadt, Germany is turning up the compute dial—hard. The science and technology giant has deployed a new high-performance computing (HPC) platform built on Lenovo ThinkSystem servers with advanced liquid cooling, hosted inside an Equinix AI-ready data center in Germany. For a company with deep stakes in life science, healthcare, and semiconductor materials, compute is no longer background infrastructure—it’s the fuel powering its next decade of scientific discovery.
The system represents a major modernization step for Merck, which is now leaning on AI, simulation, and data-dense modeling as essential tools for R&D. And by selecting Lenovo hardware, Equinix infrastructure, and a hybrid cloud design, the company is making a clear statement: scientific innovation now depends on flexible, sustainable, high-performance compute at industrial scale.
What is effectively being built here is a blueprint for how large, research-driven organizations should rethink compute strategy in an AI-dominated era.
Why Merck Needs a New Class of Compute
“Data and technology are the cornerstones of scientific progress,” said Laura Matz, Chief Science and Technology Officer at Merck KGaA, Darmstadt, Germany. And it’s not hyperbole. In life science alone, R&D increasingly depends on massive molecular simulations, structure prediction models, and multimodal datasets requiring HPC-class hardware just to produce viable outputs.
In healthcare, ML-assisted drug discovery and virtual screening methods have become compute-intensive enough that traditional on-prem systems can’t keep up. Meanwhile, Merck’s electronics business—particularly in semiconductor materials—demands precise modeling and computational chemistry workloads that spike and dip depending on project cycles.
The newly deployed HPC platform is designed to handle all of the above—with enough headroom to scale aggressively as AI models get larger and simulation workloads get heavier.
A Lenovo-Equinix Hybrid: Built for AI, Efficiency, and Scale
The new platform is powered by Lenovo ThinkSystem servers equipped with the company’s Neptune liquid-cooling technology. Liquid cooling remains uncommon outside hyperscale environments, but it’s becoming essential as processor thermal design power (TDP) balloons in step with accelerated AI workloads.
Lenovo’s take on liquid cooling allows the system to run more densely, more quietly, and more efficiently—key for Merck, which is balancing increased compute requirements with strict ESG goals.
“Even the most demanding workloads will run efficiently,” said Andreas Thomasch, CTO for Germany and Austria at Lenovo. The message: you can have HPC performance without blowing past sustainability thresholds.
Equinix’s role, though, is arguably the long-term differentiator. The system sits inside an Equinix International Business Exchange (IBX) facility, offering:
- AI-ready data center design
- Distributed infrastructure for AI/ML pipelines
- Direct, secure interconnection to clouds and partners
- Liquid cooling support across dozens of metros globally
- Data sovereignty and compliance inside German borders
Equinix has become a go-to provider for AI workloads because of its neutral, interconnected architecture—something mid-sized and global enterprises increasingly need as they adopt multi-cloud, HPC-plus-AI hybrid stacks.
Harmeen Mehta, Equinix’s Chief Digital and Innovation Officer, framed the partnership as a preview of the future: “Where digital infrastructure, compute and science converge to solve humanity’s biggest challenges.”
How Merck Plans to Use Its New Compute Muscle
The HPC system underpins innovation across all three Merck sectors. That’s unusual—most companies segment compute by business unit. Merck is opting for a unified digital backbone.
Life Science
Merck is using the platform to accelerate product development, from lab reagents to bioprocessing solutions. HPC enables more accurate predictive models, protein interaction simulations, and cloud-enhanced digital lab workflows.
Healthcare
The system will support drug screening, computational chemistry, genomics, and advanced analytics. As generative AI begins shifting into regulatory-safe environments for drug R&D, compute requirements will only grow.
Electronics
Merck is a key supplier to the semiconductor industry. Here, HPC feeds into material discovery, chemical modeling, process optimization, and next-generation materials simulation—areas where even small computational improvements can unlock major revenue opportunities.
By using one infrastructure layer across all three sectors, Merck reduces fragmentation and encourages cross-disciplinary collaboration—an approach many R&D organizations have yet to adopt.
The Bigger Picture: AI Is Forcing Enterprises Into HPC Whether They Planned For It or Not
What Merck is doing mirrors a broader shift happening across global industry:
- AI models require HPC-level compute even for inference at scale.
- Hybrid cloud architectures are replacing monolithic on-prem HPC clusters.
- Liquid cooling is moving from “experimental” to “necessary.”
- Data sovereignty and ESG are driving where compute can physically live.
Equinix and Lenovo are trying to position themselves as the default partners for enterprises facing these transitions. With more than 100 global data centers now liquid cooling-ready, Equinix is betting on a future where power density—not square footage—is the defining characteristic of the modern data center.
Meanwhile, Lenovo is aggressively expanding its Neptune liquid-cooled HPC and AI portfolio beyond academic supercomputing and into enterprise R&D, pharma, and industrial engineering.
Merck becomes one of the clearer case studies of what a modern HPC environment looks like: hybrid, interconnected, sustainable, and designed for AI from day one.
A Strategic Win for All Three Players
For Merck, the payoff is simple: more computational capability means faster experimentation and faster innovation cycles.
For Lenovo, it’s a marquee deployment of Neptune liquid cooling in a high-visibility, high-stakes environment.
For Equinix, it’s validation of its “AI-ready global interconnect” strategy—one designed for enterprises that need sovereign, high-performance, hybrid compute without building their own hyperscale-class facilities.
This partnership is less about a single HPC deployment and more about a template for the next phase of digital science. As AI and simulation workloads surge, expect more enterprises to follow suit.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
