Authors: Florian Berberich (Forschungszentrum Jülich), Kengo NAKAJIMA (The University of Tokyo, RIKEN Center for Computational Science (R-CCS)), Yutong Lu (National Supercomputing Center in Guangzhou & Shenzhen, Sun Yat-sen University), Happy Sithole (CSIR - NICIS), John Towns (University of Illinois Urbana-Champaign), Katerina Antypas (National Science Foundation (NSF)), Hal Finkel (U.S. Department of Energy)
Abstract: HPC infrastructures are being reshaped by rapid hardware innovation, pressure for greater energy efficiency, and the rise of AI and data-intensive workloads. Simultaneously, future directions are influenced not only by scientific communities and public research agendas, but also by industry roadmaps and emerging QC ambitions.
This BoF will examine how different countries plan future public HPC investments, how priorities are set, and who has influence in that process. The aim is to better understand whether public HPC remains community-driven or is increasingly shaped by industrial and market forces, and what this means for the future of research infrastructures.
Long Description: High-performance computing (HPC) is a highly dynamic domain, and the procurement of new public HPC systems has become a near-continuous process. Rapid advances in processor, accelerator, storage, and cooling technologies are pushing HPC service providers to renew systems more frequently, often in pursuit of better energy efficiency, higher performance, and improved operational sustainability. At the same time, new requirements from data-intensive workloads, Artifical Intelligence (AI) training and inference, and emerging quantum-HPC integration are placing unprecedented pressure on HPC data centers.
Public investment strategies for HPC are therefore evolving. In Europe, the EuroHPC Joint Undertaking (JU) coordinates major investments in supercomputing and develops its strategic direction through instruments such as the Multi-Annual Strategic Programme (MASP), with input from advisory bodies including Infrastructure Advisory Group (INFRAG), Research and Innovation Advisory Group (RIAG) and Quantum Technologies Advisory Group (QTAG).
Other regions appear to follow different models, including long-term national planning, mission-driven investment, industrial partnerships, and increasingly strong influence from hyperscalers and AI infrastructure providers. At the same time, AI is reshaping HPC priorities by increasing demand for large-scale compute, high-bandwidth memory, storage throughput, and resilient data infrastructure.
This BoF will bring together HPC practitioners, strategy experts, infrastructure providers, and community representatives from different countries to discuss how future public HPC investments are planned, who influences these decisions, and how community priorities can remain visible in an era increasingly shaped by AI, energy constraints, industrial roadmaps, and quantum ambitions.
Historically, public HPC strategies were strongly influenced by scientific communities, national research priorities, and long-term capability building. Today, the landscape is more complex. Procurement cycles are tightening. Energy efficiency is becoming a first-order constraint. AI and Quantum Computing (QC) is entering strategy discussions. Industry roadmaps, accelerator vendors, cloud providers, and hyperscalers may now have more influence than ever before.
This creates a strong need for open discussion across countries and stakeholder groups:
- How are future public HPC investments actually planned?
- What processes exist for involving the research and user community?
- How much influence do practitioners have compared with governments, industry, or hyperscalers?
- Are AI and QC expanding the strategic vision for HPC, or narrowing it around a few dominant technology paths?
- What can countries learn from each other when designing future funding and procurement models?
This BoF aims to:
- Compare how different countries and regions prepare and prioritize public HPC investments.
- Discuss the respective roles of governments, funding agencies, scientific communities, operators, vendors, hyperscalers, and AI actors.
- Examine how AI, data-intensive science, energy efficiency, and quantum computing are changing infrastructure strategy.
- Identify whether current planning processes adequately reflect user and community needs.
- Share examples of good practice in strategic planning, advisory structures, and community engagement.
- Explore whether public HPC is still primarily community-driven, or increasingly shaped by industrial supply and market power.