NVIDIA is putting $250 million of computing power behind AI safety research in the UK, creating a five-year program designed to help researchers develop and test stronger safeguards for advanced AI systems. The commitment delivers $50 million in compute access every year, with infrastructure hosted in the UK through UK-based NVIDIA Cloud Partners.
Announced on October 14, 2026, the program reaches across the wider UK AI safety and security research community. It will support work with the British Open-Ended Learning and Discovery Lab, the Science of Fundamental AI Research Lab, and Imperial College London, connecting research into AI safety and security with work on agents, open models, robotics, and real-world threats.
A Five-Year Compute Engine for UK AI Research
The donation gives UK researchers access to the computing resources needed to develop, test, and reinforce safeguards for AI systems. Most of the $250 million will support the wider UK AI safety and security research community, while specific work will include secure agents, open-model security, and real-world threat mitigations.
Compute access will come through UK-based NVIDIA Cloud Partners, and the infrastructure will remain hosted in the UK. That structure gives the program a clear national footprint while linking research groups across different areas of AI development.
- BOLD: The British Open-Ended Learning and Discovery Lab is developing approaches for AI systems to learn, adapt, and work alongside people.
- SOFAIR: The Science of Fundamental AI Research Lab is exploring new open source AI architectures and ways to make AI development and deployment more accessible.
- Imperial College London: Its work will include safety engineering and robotics research.
The research will also include safeguards using NVIDIA OpenShell™ and NVIDIA Sentry. Together, these efforts aim to give researchers room to examine how AI systems behave, how they can be secured, and how their risks can be addressed in real-world settings.
Agents, Open Models, and Machines That Act
The program’s focus reaches beyond static software. Research will cover secure agents, systems that can learn and adapt, and methods for managing threats as AI moves into more capable forms of action.
Jakob Foerster described the goal this way: “This compute will help us explore transparent and human-centric approaches to AI that allow us to collectively rise to the challenge by building in governance and human-AI co-improvement from the ground up.”
That human-centered direction runs through the participating research areas. BOLD is working on AI that learns and works alongside people, while SOFAIR is exploring open source architectures that can make development and deployment more accessible. Imperial College London adds safety engineering and robotics to the mix.
Alessandra Russo pointed to the physical systems at the center of this next phase: “The next frontier of artificial intelligence is physical — robots, autonomous systems of agents that perceive, reason and act in the real world.”
That focus makes safety engineering a central part of the program. When systems perceive, reason, and act, researchers need ways to test their safeguards before those systems operate in real-world environments.
Government Support for AI Security
The UK government confirmed that its AI Security Institute, or AISI, will be able to examine cutting-edge technology. That capability sits alongside the new compute program, creating a direct connection between research capacity and AI security examination.
UK AI Minister Kanishka Narayan called the commitment a vote of confidence in the country’s AI research strength, saying: “This commitment is a vote of confidence in our world-leading strength and expertise in AI research — and, in particular, frontier AI safety that is becoming ever more important as AI’s capabilities develop.”
Business Secretary Jonathan Reynolds also connected safety with the wider promise of AI. “To realise the opportunities of AI, people need to feel the technology is safe and secure. The government is committed to ensuring AI can be securely harnessed to fuel new discoveries and drive growth,” he said.
Jensen Huang, NVIDIA’s founder and CEO, summed up the company’s position with a direct message: “AI capability and safety must advance together.”
The scale of the commitment stands out. NVIDIA, valued at $5.5 trillion, is dedicating $250 million in computing power across five years, giving UK researchers $50 million in access each year. The investment now sets a substantial foundation for work on AI safeguards, secure agents, open-model security, and robotics.
As AI systems gain new abilities, the UK program will give researchers more computing capacity to examine those abilities and build protections around them. The next five years will show how far that combination of compute, safety engineering, open research, and government examination can push the country’s AI security work forward.
Based on
- NVIDIA Pledges Five Years of Compute to UK Agentic Safety Research — unite.ai
- Nvidia, worth $5.5tn, donates $250m to UK for AI safety research | Nvidia | The Guardian — theguardian.com
- NVIDIA Commits $250 Million to Advance UK Agentic Safety and Security Research Over the Next Five Years | Markets Insider — markets.businessinsider.com
