Research body launches tender for high-capacity AI platform hardware
Procurement covers servers, storage and networking to support simultaneous workloads on at least 96 GPUs, signalling rising demand for AI-ready infrastructure.
Procurement covers servers, storage and networking to support simultaneous workloads on at least 96 GPUs, signalling rising demand for AI-ready infrastructure.
Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.
Naukowa i Akademicka Sieć Komputerowa - Państwowy Instytut Badawczy has issued an AI Research Platform Expansion contract notice to grow its artificial intelligence research capacity. The institute is looking for servers, storage and network infrastructure able to support simultaneous computing on at least 96 GPU units, underlining how advanced AI workloads are shaping public-sector technology buying.
The contract centres on expanding an existing research platform for artificial intelligence. The successful supplier will be expected not only to provide the hardware, but also to bring it into service and keep it operational as a single environment.
The notice specifies four main elements:
The requirement to support simultaneous computing on at least 96 GPU units points to a high-density, parallel processing environment suited to training and running complex AI models. Rather than a simple equipment refresh, the brief describes an integrated platform in which compute, storage and networking need to be configured together to handle demanding research workloads.
The contract notice was published on 8th September 2026. It arrives in a year marked by a steady flow of AI- and high-performance-computing procurements from universities, research institutes and frontline public services.
Across the research landscape, recent tenders show institutions building dedicated environments for AI and data-intensive science. In March 2026, Instytut Biologii Doświadczalnej im. M.Nenckiego PAN launched a Personalized Analytical Platform procurement for an IT system to visualise and analyse biological and biomedical images, aimed at enhancing bioimaging data analysis. That same month, Wojskowa Akademia Techniczna im. Jarosława Dąbrowskiego issued a Computer Equipment Delivery notice for GPU/AI computing servers and network disk arrays.
By May and June 2026, universities were following suit. Politechnika Wrocławska went to market for a Server for AI Model Training, while Uniwersytet im. Adama Mickiewicza w Poznaniu published an Expansion of Computing Cluster tender for its Artificial Intelligence Center, divided into three parts with detailed delivery and documentation requirements. In July 2026, Uniwersytet Radomski im. Kazimierza Pułaskiego sought a complete HPC/AI class GPU server set through its Computer Equipment Supply for University contract, and ΤЕΧНИЧЕСКИ УНИВЕРСИТЕТ issued a Supply of AI Training Terminals and Storage to equip specialised AI training terminals and a large data storage platform.
Several research-focused bodies are also procuring full-scale high-performance infrastructures. In July 2026, the Medical University – Varna published a High-Performance Computer Equipment Delivery notice for computer equipment dedicated to artificial intelligence, and Ίδρυμα Τεχνολογίας & Έρευνας (ΙΤΕ) launched an HPC and AI Research Infrastructure project combining high-performance servers, application development servers, user access servers and a high-speed computational network, all of which must be compatible and guaranteed by the supplier. Taken together with the 96-GPU ambition of the new research platform expansion, these projects suggest that advanced AI is now treated as a core workload for institutional computing, not an experimental sideline.
Frontline services are also reshaping their infrastructure and systems around artificial intelligence. In March 2026, Pabianickie Centrum Medyczne Sp. z o.o. issued an IT Equipment and Cybersecurity Solutions notice covering IT hardware and software licences alongside the implementation and expansion of a hospital information system, including electronic medical documentation and imaging solutions, with training and support built in. Szpital Dziecięcy Polanki im. Macieja Płażyńskiego w Gdańsku sp. z o.o. followed with an AI Systems for Hospital Optimization contract for systems using ML, LLM and AI algorithms to streamline hospital processes and staff workflows; the hospital has chosen not to divide the order into parts, in order to preserve compatibility and coherence. Further health-sector tenders in April and May 2026, from Szpital w Puszczykowie im. prof. S.T. Dąbrowskiego S.A. (IT Infrastructure Management System), Akademickie Centrum Zdrowia w Mikołowie Sp. z o.o. (Expansion of Medical System and AI Solutions) and Regionalny Szpital Specjalistyczny im. dr Władysława Biegańskiego (AI Solutions for Hospital Processes), seek AI-enhanced systems to support diagnostic, therapeutic and managerial decision-making, often coupled with user training and ongoing technical support.
Other public bodies are upgrading core infrastructure with AI in mind. In April 2026, the Sąd Apelacyjny w Krakowie published a Data Center Expansion in Krakow notice to expand its Data Processing Center with dedicated servers for AI applications, blade servers and additional disk arrays, while Szpital Praski P.W. Przemienienia Pańskiego Spółka z Ograniczoną Odpowiedzialnością launched an IT Infrastructure Modernization tender that stresses compatibility with existing systems and compliance with environmental and safety regulations. In May 2026, the Urząd Komisji Nadzoru Finansowego sought Network Infrastructure Expansion, and Statutární město Brno went to market for an AI Platform Development and Management project to deliver a universal multi-agent AI platform for the city, combining chatbots, internal AI agents, automated workflows, hardware delivery, staff training and high service availability.
Several organisations are procuring dedicated computing capacity for AI workloads in their own domains. Krajowy Ośrodek Wsparcia Rolnictwa’s April 2026 AI Model Training Server contract focuses on a server designed for training and running artificial intelligence models. In July 2026, Zakład Informatyki Lasów Państwowych im. Stanisława Kostki Wisińskiego announced a Server Purchase for AI/ML Environment with strict compatibility and legal requirements and a manufacturer’s warranty, and Centrum Usług Informatycznych we Wrocławiu followed in August 2026 with a Data Processing Infrastructure Expansion tender for rack servers and accelerator cards for an AI system, subject to confidentiality rules around the detailed specifications. In August 2026, Lotnicze Pogotowie Ratunkowe’s Computing Infrastructure Expansion notice combined servers, equipment and software licences with a 36-month warranty service, and by early September 2026 the Państwowy Fundusz Rehabilitacji Osób Niepełnosprawnych was procuring Server and Network Infrastructure with virtualisation software and support. A cultural institution, EC1 Łódź – Miasto Kultury w Łodzi, has meanwhile published a Supply of AI Computing Server notice for a GPU-accelerated server to run and fine-tune generative AI models under tightly defined operational and warranty conditions.
The AI Research Platform Expansion from Naukowa i Akademicka Sieć Komputerowa - Państwowy Instytut Badawczy is notable for the scale of GPU capacity it brings into a single research environment, but it clearly forms part of a broader pattern. Across research institutions, hospitals, courts, regulators and municipal bodies, 2026 contract notices show AI moving from pilot projects into the core of operations, backed by significant investment in servers, storage, networking and long-term support. How suppliers respond to demands for integrated, compatible and maintainable platforms – and how quickly these projects translate into live services – will do much to determine the shape of public-sector AI in the years ahead.
Tenderlake Monitor reads every public sector notice published across the UK, the EU and the EEA, in the language it was published in, and tells you which ones match what your company sells and why. It works from a plain description of your business, not from keywords or CPV codes.
Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.