Research body launches tender for AI supercomputer and maintenance
A major AI supercomputer tender aims to equip an AI Factory with high-performance, energy-efficient hardware and long-term maintenance support.
A major AI supercomputer tender aims to equip an AI Factory with high-performance, energy-efficient hardware and long-term maintenance support.
Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.
On 20th July 2026, the European High-Performance Computing Joint Undertaking published a contract notice to acquire, deliver, install and maintain a high-performance AI supercomputer to support an AI Factory. Through its AI Supercomputer Acquisition and Maintenance tender, the joint undertaking is seeking a contractor with proven experience of similar environments and a solution that balances performance, programmability and energy efficiency. The move underlines how AI-specific supercomputing has become a core part of research and innovation infrastructure, and will attract interest from high-performance computing vendors, systems integrators and long-term support providers.
The notice sets out a clear intention to appoint a single contractor responsible for every major stage of the AI Factory supercomputer’s lifecycle. The successful supplier will be expected to oversee acquisition of the system, deliver it to site, complete installation and provide ongoing maintenance.
Rather than listing detailed technical specifications, the joint undertaking defines the system in functional terms. It must be a high-performance AI supercomputer dedicated to supporting an AI Factory, and it must place particular emphasis on:
This combination points towards a machine tuned to intensive AI workloads but also accessible to developers and researchers who need to build and run complex applications. The explicit focus on energy efficiency signals that power consumption and operating costs are key design constraints alongside raw compute capability.
The requirement for proven experience in similar environments is another strong signal. The buyer is looking for organisations that can show a track record of delivering and supporting comparable high-performance AI or supercomputing installations. For smaller or newer players in the market, that bar may be challenging; for established vendors and integrators, it reinforces the value of reference sites and long-standing support arrangements.
By bundling acquisition with delivery, installation and maintenance, the joint undertaking is avoiding a fragmented approach in which separate contracts cover hardware, integration and operations. Instead, it is asking bidders to stand behind an integrated solution over time. That places as much weight on project management, engineering and support capabilities as on the choice of processors, accelerators or storage technologies.
The AI Factory supercomputer is one of a growing number of AI-focused high-performance computing projects to come to market in 2026. Across research, education, government, healthcare and business, buyers are commissioning infrastructure that looks beyond general-purpose servers to systems explicitly designed for AI workloads.
In March 2026, the University of Galway issued a contract notice for CASPIr Supercomputer Hardware and Software. That project, like the AI Factory tender, covers acquisition, delivery, installation and maintenance of a supercomputing system to provide advanced computing services for researchers. It shows universities moving to secure long-term, integrated arrangements for supercomputers rather than treating them as short-term capital purchases.
Also in July 2026, Vysoká škola báňská – Technická univerzita Ostrava launched its HPCQC-AI Computational Cluster procurement. The university aims to modernise equipment by acquiring a supercomputer as a comprehensive system solution for high-performance computing and artificial intelligence, including implementation, integration, training and maintenance services. The expectation that a single supplier will handle deployment, integration and user training closely mirrors the lifecycle approach taken by the European High-Performance Computing Joint Undertaking.
National research and education networks are following a similar trajectory. In June 2026, HRVATSKA AKADEMSKA I ISTRAŽIVAČKA MREŽA CARNet published a notice for a High-Performance Computing Platform, seeking a local high-performance computing platform to support computationally intensive operations in scientific, educational and infrastructural contexts.
Specialist research centres are also scaling up. In June 2026, Forschungszentrum Jülich issued a contract notice for Servers and Network Infrastructure Procurement to enhance the computing capacity of its ICE-2 Institute for a range of high-performance computing tasks. While not described as an AI-specific system, it reflects the same pressure on research institutions to expand and modernise their computing backbones.
Beyond the research sector, demand for AI-capable infrastructure is becoming visible in security and healthcare. In June 2026, the United States Department of Justice, acting for the FBI, released an AI Computing Infrastructure RFI seeking information on sources of high-performance AI computing infrastructure, including hardware and software, with a particular focus on small business participation. And in May 2026, the health insurance body Česká průmyslová zdravotní pojišťovna went to market for the Delivery of AI Supercomputers, covering two AI supercomputers for large language model inference, together with a high-speed connection cable, warranty and documentation.
Local and regional bodies are looking at similar capabilities on a different scale. In July 2026, the city council in Salamanca published a notice for a High-Performance Computing Platform to supply, install, configure and commission a system for intensive computing tasks and AI-related workloads. In April 2026, the Chamber of Commerce of Ceuta launched the AI Computing Infrastructure for Ceuta procurement to implement a computing centre for artificial intelligence, spanning hardware, storage, networking, software and associated installation and support services. And on 20th July 2026, Hochschule Zittau/Görlitz issued its AI Cluster Installation and Support notice for delivery and installation of an NVIDIA DGX B200 cluster system with software and network components.
Taken together, these procurements show how AI-capable high-performance computing is moving from a niche concern to a mainstream requirement across sectors. The European High-Performance Computing Joint Undertaking’s AI Factory supercomputer sits firmly within this trend, but stands out for its clear emphasis on AI workloads and its insistence on a single contractor covering acquisition, installation and maintenance.
Although the AI Supercomputer Acquisition and Maintenance notice is concise, it contains several pointers for potential bidders and for the wider AI infrastructure market.
First, it confirms that buyers at this level now see AI-oriented supercomputers as strategic infrastructure. The machine is not described as a research pilot or prototype; it is framed as the compute backbone for an AI Factory, with long-term maintenance built in. That aligns with projects such as Galway’s CASPIr supercomputer and the HPCQC-AI cluster in Ostrava, where supercomputers are positioned as shared services for large user communities.
Second, the joint undertaking’s focus on performance, programmability and energy efficiency reflects a balancing act seen in other 2026 procurements. East Durham College’s Infrastructure Refresh Proposal, for example, asks suppliers to renew infrastructure and end-user computing equipment while meeting public sector standards and sustainability objectives. In the AI Factory tender, energy efficiency is elevated to one of three core design priorities, indicating that performance gains will need to be delivered within tight power and sustainability constraints.
Third, the AI Factory supercomputer is framed as shared infrastructure rather than a system for a single project. By contrast, some 2026 tenders focus on targeted installations. The University of Craiova’s Procurement of AI Server and UPS is tied to a deep learning project on fetal morphology, while Nexus Media’s AI Infrastructure Hardware Procurement is dedicated to the PIONEER AI autonomous business robot project. In those cases, the computing infrastructure is clearly scoped around a defined use case. The AI Factory tender, by contrast, describes a supercomputer intended to underpin a broader AI facility, with the supplier responsible for keeping it operational over time.
For prospective bidders, the notice crystallises several capabilities that are now close to mandatory in this part of the market:
The AI Supercomputer Acquisition and Maintenance notice is high-level. It does not set out the system’s planned size, specific technologies or budget, and it offers only headline priorities rather than detailed performance or efficiency targets. From the notice text alone, it is not yet possible to gauge the eventual scale of the AI Factory supercomputer.
Even so, the combination of a dedicated AI supercomputer, an AI Factory remit, and a requirement for proven experience marks this out as a significant opportunity for established high-performance computing suppliers. As more organisations — from universities and research centres to insurers, city councils and chambers of commerce — move to procure AI-capable infrastructure, the terms and eventual outcome of this tender will provide a useful reference point for how public-sector buyers specify integrated, energy-efficient AI supercomputers.
Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.