The UK’s plan to become an “AI maker, not an AI taker” is beginning to take shape through new sovereign compute infrastructure, startup support and enterprise AI deployments powered by NVIDIA technologies.
At London Tech Week, NVIDIA and its partners highlighted how the country is building the foundations for a domestic AI ecosystem. The effort focuses on keeping more AI development, data processing and computing capacity inside the UK while helping local startups and enterprises scale advanced AI systems.
Sovereign AI has become a major priority for governments worldwide. For the UK, the goal is to strengthen national AI capability by supporting local infrastructure, homegrown companies and sector-specific AI applications in areas such as healthcare, coding, robotics, life sciences and agentic AI.
NVIDIA AI Infrastructure Expands Across the UK
A major part of the UK sovereign AI strategy is access to domestic compute. Over the past year, the number of AI cloud providers planning to deploy AI infrastructure in the UK has doubled, according to NVIDIA.
Nebius is expanding its UK AI cloud operations with new deployments of NVIDIA AI infrastructure. These deployments are expected to reach 65 megawatts when fully ramped in 2027.
CoreWeave is also building AI infrastructure in the UK Government’s AI Growth Zones. Meanwhile, BT and Nscale have announced plans to create sovereign AI data centers across three existing BT sites in the UK. The plan combines NVIDIA AI infrastructure, Nscale’s full-stack platform and BT’s national connectivity network.
This expansion matters because AI development increasingly depends on high-performance computing. For startups, universities and enterprises, local access to GPU infrastructure can reduce dependence on overseas cloud capacity and support more secure AI development.
Isambard-AI Powers UK Research and Startups
At the center of the UK’s sovereign AI push is Isambard-AI, described by NVIDIA as the UK’s most powerful computer. The system is built on 5,400 NVIDIA GH200 Grace Hopper Superchips and runs entirely on zero-carbon electricity.
Isambard-AI is being used to support ambitious research and startup projects through the UK government’s Sovereign AI Fund. The fund is designed to help domestic companies access the compute needed to train and deploy advanced AI systems.
One of the early recipients is Ineffable Intelligence, which is collaborating with NVIDIA on reinforcement learning infrastructure. Several UK-based NVIDIA Inception startups are also using Isambard-AI to build AI systems across coding, agentic AI, inference and biology.
UK Startups Use Sovereign Compute for Advanced AI
Several UK startups are now using NVIDIA-powered sovereign compute to develop specialized AI platforms.
Cosine Builds a Sovereign AI Coding Platform
Cosine is developing an end-to-end sovereign AI coding platform for highly regulated sectors, including financial services, critical infrastructure and national security.
The company is using Isambard-AI to train a large-parameter, mixture-of-experts, multimodal agentic large language model. The system is designed to handle more than text and images, making it useful for complex software and enterprise workflows.
Cursive Develops Self-Improving AI Systems
Cursive is building AI systems that can learn continuously from real-world data and operate autonomously for longer periods. The company is working on memory-augmented AI architectures with larger context windows.
Cursive has also adopted NVIDIA Megatron-LM, a framework used for distributed training at scale.
Doubleword Focuses on Faster AI Inference
Doubleword, described as the UK’s first dedicated inference lab, is working to optimize AI inference across the stack. The company uses open models, including NVIDIA Nemotron 3 Super 120B, and builds on NVIDIA Dynamo.
On Isambard-AI, Doubleword reported major improvements in model loading and KV cache compression. These advances are important for agentic AI workloads, where models may need to run continuously and respond efficiently.
Prima Mente Applies AI to Alzheimer’s and Neurodegenerative Disease
Prima Mente is building biological foundation models to study diseases such as Alzheimer’s, Parkinson’s and ALS. Its work focuses on identifying biomarkers, disease subtypes and potential drug targets.
Using its Isambard-AI allocation, the company is developing Pleiades 2, a foundation model that combines five biological data modalities. Prima Mente is also using NVIDIA tools such as Parabricks and Transformer Engine to improve genomic processing and model optimization.
NVIDIA’s £2 Billion UK Startup Investment Strengthens AI Ecosystem
NVIDIA’s £2 billion investment in the UK startup ecosystem is another major part of the country’s AI growth strategy. The investment is being made in collaboration with venture capital firms and is expected to support AI hubs including London, Oxford, Cambridge and Manchester.
NVIDIA also reported that UK membership in its Inception startup program has grown by 50% over the past year. AI-native companies such as Doubleword, Synthesia and PolyAI are among the UK-rooted firms scaling globally.
This signals that the UK is not only investing in infrastructure but also building a pipeline of AI companies capable of competing internationally.
AI Skills and Developer Programs Expand
The UK’s sovereign AI strategy also includes talent development. NVIDIA previously announced a collaboration with the UK Department for Science, Innovation and Technology focused on 6G and AI skills.
That 6G collaboration has supported testbeds at four UK universities. NVIDIA’s Deep Learning Institute has also delivered new courses for participants from more than 30 UK universities.
In addition, NVIDIA says its Developer Program now includes more than 200,000 UK developers. This growing developer base could help the UK turn AI infrastructure into real products, services and research breakthroughs.
Enterprise AI Moves Beyond Pilots
NVIDIA also highlighted several UK enterprise AI projects moving from pilot programs into production.
Apian is building digital twins of two NHS hospitals using autonomous devices, ground robots, computer vision and robotic simulation. Deliverance AI is helping regulated companies run and manage AI agents inside their own environments. Glass Futures has installed an AI-driven digital twin of a glass furnace to test more efficient manufacturing methods.
Other examples include OneAdvanced fine-tuning NVIDIA Nemotron 2 Nano 9B for NHS primary care triage data, Orbital Industries working on AI factory infrastructure, and Reading Football Club partnering with Stelia to create an AI Centre of Excellence.
These projects show that sovereign AI is not limited to government policy or research labs. It is beginning to affect healthcare, manufacturing, sports, enterprise software and industrial operations.
Why UK Sovereign AI Matters
The UK sovereign AI movement reflects a broader global shift. Countries increasingly want more control over the infrastructure, data and talent behind artificial intelligence.
For the UK, this means building local compute capacity, backing domestic startups and applying AI to sectors where data security and national capability matter. NVIDIA’s role gives the country access to advanced GPU infrastructure, AI software frameworks and startup ecosystem support.
The biggest test will be whether these investments can produce long-term commercial and scientific results. If the UK can turn sovereign compute into globally competitive AI companies and real-world deployments, it could strengthen its position in the next phase of artificial intelligence.
For now, the message is clear: the UK wants to build AI at home, and NVIDIA is becoming one of the key technology partners behind that ambition.
