{"slug": "the-supercomputing-dna-of-the-ai-factory", "title": "The supercomputing DNA of the AI factory", "summary": "HPE and NVIDIA are positioning Cray supercomputing architecture as the foundation of their joint AI factory portfolio, with the Leibniz Supercomputing Centre's upcoming Blue Lion system built on the HPE Cray Supercomputing GX5000 architecture and NVIDIA Vera Rubin expected to run grand-challenge problems 30 times faster than LRZ's previous flagship systems. The University of Utah has partnered with HPE and NVIDIA on an on-premises sovereign AI factory that has more than tripled the university's computational capacity for genomic and protein-folding research. The US Mint is also releasing a US $1 Cray-1 coin honoring the supercomputer's system-architecture legacy.", "body_md": "In the world of high-performance computing (HPC), few symbols carry as much weight as the C shape of the Cray-1. The [US $1 Cray-1](https://www.usmint.gov/news/press-releases/2026-american-innovation-dollar-coin-rolls-and-bags-featuring-wisconsin-on-sale-april-7) coin is more than a tribute to computing history. It is a reminder that breakthrough performance has always depended on system architecture: how compute is arranged, connected, cooled, and operated.\n\nAI factories may be relatively new, but the challenge is familiar: orchestrating massive amounts of compute, data, networking, power, and cooling into a single system that consistently delivers outcomes. That’s exactly the problem Cray supercomputing was built to solve, and why its engineering DNA is foundational to today’s AI factories.\n\n**Co-engineering the HPE AI Factory**\n\nThe collaboration between HPE and NVIDIA is built on a shared philosophy: to handle the massive datasets required for agentic AI and deep scientific simulation, the hardware cannot be a collection of disparate parts. It must be a unified system.\n\nBy combining HPE’s world-leading supercomputing infrastructure with NVIDIA GPUs and NVIDIA Spectrum-X Ethernet networking, the companies have created a portfolio of AI factory solutions. These dedicated, on-premises engines are designed to turn raw data into actionable intelligence with [unprecedented scale and security](https://www.hpe.com/us/en/newsroom/press-release/2026/03/hpe-accelerates-secure-scalable-production-ready-ai-through-new-innovations-with-nvidia.html).\n\n**Innovation in action**\n\nThe true measure of this collaboration is found in joint customers. These organizations are not just buying servers; they are deploying massive-scale systems to redefine their industries.\n\n**1. Leibniz Supercomputing Centre (LRZ): Pushing the frontiers of physics**\n\nAt the Leibniz Supercomputing Centre in Germany, the mission is to bridge the gap between traditional physics-based simulations and data-driven AI. Their upcoming system, called Blue Lion, is built on the [HPE Cray Supercomputing GX5000](https://m.youtube.com/watch?v=f8pZv1CsvTo) architecture.\n\nBy leveraging NVIDIA’s Vera Rubin within this HPE framework, LRZ researchers are preparing to tackle grand-challenge problems—ranging from astrophysics to climate modelling—at speeds [30 times faster](https://www.hpe.com/us/en/newsroom/press-release/2024/12/hewlett-packard-enterprise-builds-next-generation-100-direct-liquid-cooled-supercomputer-at-the-leibniz-supercomputing-center.html) than their previous flagship systems. This on-premises powerhouse allows scientists to run complex, high-resolution models that were previously computationally too expensive, ensuring that European research stays at the global forefront of discovery.\n\n**2. The University of Utah: Democratizing sovereign AI**\n\nOrganizations are increasingly seeking sovereign AI—the ability to own and control their own AI development infrastructure. The [University of Utah](https://www.hpe.com/us/en/newsroom/press-release/2025/10/university-of-utah-and-the-state-of-utah-to-accelerate-medical-research-and-expand-regional-economic-development-with-a-sovereign-ai-factory-powered-by-hpe-and-nvidia.html) has partnered with HPE and NVIDIA to build a massive on-premises AI factory.\n\nThis deployment has [more than tripled](https://www.hpe.com/us/en/newsroom/press-release/2025/10/university-of-utah-and-the-state-of-utah-to-accelerate-medical-research-and-expand-regional-economic-development-with-a-sovereign-ai-factory-powered-by-hpe-and-nvidia.html) the University’s computational capacity. For the students and faculty, this translates to real-world impact in medical research, where researchers can now process genomic data and protein-folding simulations in a fraction of the time. By housing this power locally, the university ensures data privacy and provides a stable environment for regional economic development, proving that supercomputing is an essential utility for today’s university.\n\n**3. Mercedes-AMG PETRONAS Formula One team: Engineering at the limit**\n\nIn the world of Formula One, milliseconds matter. The [Mercedes-AMG PETRONAS team](https://community.hpe.com/t5/ai-unlocked/deep-learning-in-the-fast-lane-the-compute-behind-the-car-to-win/ba-p/7263873) relies on the computational density provided by HPE and NVIDIA to power their digital wind tunnels. Because F1 regulations strictly limit physical testing time, the team’s success depends on the accuracy of their Computational Fluid Dynamics (CFD) simulations.\n\nUsing on-premises [high-performance computing infrastructure](https://www.hpe.com/us/en/newsroom/blog-post/2025/10/beyond-the-grid-how-true-partnership-drives-success-for-mercedes-amg-petronas-formula-one-team-and-hpe.html), the team can run thousands of iterative design simulations every week. This allows them to debut aerodynamic upgrades at a pace that keeps them competitive at the highest level of motorsport.\n\n**4. COMLINE: Automating service with local intelligence**\n\nFor German IT service provider [COMLINE,](https://www.hpe.com/us/en/newsroom/press-release/2025/02/comline-to-launch-genai-as-a-service-based-on-hpe-private-cloud-ai.html) the move to GenAI required a solution that met strict data sovereignty and performance standards. By deploying Private Cloud AI, a turnkey, on-premises solution co-developed with [NVIDIA,] they were able to launch their AI services in record time.\n\nThis system integrates NVIDIA AI Enterprise software directly with HPE’s modular hardware, allowing COMLINE to automate complex service delivery tasks for their clients. It’s a prime example of how the HPE AI Factory model can be scaled and packaged for enterprise use, bringing supercomputing-grade innovation to the corporate data center.\n\n**5. Los Alamos National Laboratory: Where breakthroughs begin**\n\nFrom advancing scientific discovery to accelerating national security missions, Los Alamos National Laboratory exemplifies the growing convergence of supercomputing and AI. Leveraging innovations that span HPE’s Cray supercomputing heritage and next-generation NVIDIA accelerated computing, the HPE Cray Supercomputing GX5000 platform represents a new class of AI factory infrastructure, bringing together the compute, networking, cooling, and software needed to power intelligence at unprecedented scale. The result is a foundation capable of supporting both the world’s most complex simulations and the AI workloads that increasingly drive innovation and competitive advantage.\n\n**A legacy worth more than a dollar**\n\nThe lesson of Cray is that architecture wins. The world’s most complex challenges have never been solved by technology in isolation, but by systems engineered to work as one. That same philosophy is now shaping the AI era, where integrated compute, networking, cooling, software, and services are the foundation of AI factories.\n\nHPE and NVIDIA bring decades of supercomputing expertise to this new frontier, helping organizations transform data into insight at unprecedented scale. The HPE Cray coin represents more than computing history. It represents a vision that continues today, powering breakthroughs that expand scientific understanding, strengthen national capabilities, and help solve some of society’s most important challenges.\n\nWhether it is the [Blue Lion at LRZ](https://www.lrz.de/en/news/detail/blue-lion-based-on-vera-rubin-chip-architecture) or the aerodynamic simulations at Mercedes-AMG PETRONAS, the results are clear: when world-class system architecture meets world-class acceleration, the impossible becomes achievable. We don’t just look at the coin to remember where we started; we look at it to remind ourselves of the speed at which we are moving toward the future. \n\nLearn more about [HPE Cray](https://www.hpe.com/cray) and explore [hpe.com/ai](https://www.hpe.com/ai).\n\n******************\n\n*As AI becomes increasingly central to economic competitiveness, scientific advancement, and national priorities, organizations require infrastructure that balances performance with security and sovereign control. Together, HPE and NVIDIA co-engineer rack-scale AI systems that integrate AI computing, high-performance networking, and supercomputing expertise to support large-scale AI workloads. This provides enterprises, governments, and research institutions with a trusted foundation for sovereign AI initiatives while maintaining control over critical data, models, and operations.*", "url": "https://wpnews.pro/news/the-supercomputing-dna-of-the-ai-factory", "canonical_source": "https://www.cio.com/article/4224921/the-supercomputing-dna-of-the-ai-factory.html", "published_at": "2026-09-22 13:10:38+00:00", "updated_at": "2026-09-22 13:22:53.631888+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-chips", "ai-research"], "entities": ["HPE", "NVIDIA", "Cray-1", "Leibniz Supercomputing Centre", "Blue Lion", "HPE Cray Supercomputing GX5000", "NVIDIA Vera Rubin", "University of Utah"], "alternates": {"html": "https://wpnews.pro/news/the-supercomputing-dna-of-the-ai-factory", "markdown": "https://wpnews.pro/news/the-supercomputing-dna-of-the-ai-factory.md", "text": "https://wpnews.pro/news/the-supercomputing-dna-of-the-ai-factory.txt", "jsonld": "https://wpnews.pro/news/the-supercomputing-dna-of-the-ai-factory.jsonld"}}