Contribution
I valued my hon. Friend’s earlier point that sovereignty has to be seen in the round. We cannot make everything here; we have to look at the entire bundle that we have to offer. In the context of LLMs, there is some uncertainty as to whether all the capability will ultimately accrue in closed proprietary models, or whether open-source, open-weight models might be part of it. To me, as things stand today, it is a pretty important part of the stack. The question then is whether we have enough of it to be able to make the most of it by adopting it for economic and national security usage here, or whether there are aspects in which, at least from a distillation or small-model point of view, we need to develop some capabilities here as well. I do not think there is a binary answer to the overarching question; the answer is much more nuanced. I am happy to discuss that further if it is of interest.
As I said, the third rung of the ladder is, ultimately, to build British and focus on areas in which we can develop our strengths. I have to point out that we made sure that Nscale, one of our neocloud hyperscale providers, was an important part of the supply chain for AI growth zones. I noticed that yesterday Nscale raised the largest ever series-C funding in Europe, in part as a result of the Government’s support and convening in that context. Arm, the leading chip design company globally, is still headquartered in Cambridge, and we have fantastic companies in the AI inference chip part of the stack, Fractile and Olix being two of them. It is an area that I spend a lot of my time on.
When it comes to models, we have huge strengths, not just because a number of the Gemini teams and researchers continue to sit in King’s Cross at DeepMind, but because companies developing foundation models in AI for science and autonomous vehicles, embodied AI, and aspects of world models and computer vision reside here in the UK. Wayve raised £1.5 billion just this year, the largest funding round in Europe to date for that stage. It is a fantastic company that looks in particular at embodied AI and vision. I am proud of those companies. It is right that the Government are supporting them through the lens of tech sovereignty, as that is what both Britain’s and the companies’ best interests dictate.
The sovereign AI unit will be crucial to that. I am glad to see the level of interest in that across the House. It will concentrate efforts on priority areas. There was interest in my specifying those areas. The four areas that are of interest at the outset are novel compute, in particular focusing on the inference chip part of the stack; novel model architecture; AI for science—I point hon. Members to the AI for science strategy published by the Department three or four months ago, which set out particular areas of focus and priority—and embodied AI.
To give a concrete example of early action that the sovereign AI unit has taken, we have already invested £8 million in the OpenBind consortium to accelerate AI-driven drug discovery, and £5 million in the Encode: AI for Science fellowship to support the next generation of world-class talent. The focus of the unit will be on both capital and compute, to incrementally anchor more and more British companies here, but I know that the unit will only be part of the solution. We have a role to look at innovation and market support much more broadly across the tech landscape.
In November, we also announced a significant advance market commitment—a deeply innovative procurement shift—which meant that up to £100 million in Government funding was available to buy products from promising UK chip companies once they reach a high-performance benchmark. That presents UK start-ups with an exciting opportunity to grow and compete right here, building for the world.
AI is of course just one area of Britain’s flourishing tech ecosystem. I point out to my hon. Friends the Members for Milton Keynes Central (Emily Darlington) and for Lichfield (Dave Robertson), who made important points about quantum, that the Government have doubled the rate of investment in quantum, with about £1 billion committed over the next four years. The points on helium made by my hon. Friend the Member for Lichfield have very much been taken into account. The Government are looking at the developing situation on helium supply in the middle east, which is of concern.
Through our national programme, we broadly want to anchor development and access to technological capabilities that are most important to economic growth and national security. That means, in the context of quantum, more companies starting, growing and staying here and, in the context of AI, not just developing capabilities in particular parts of the stack, but in part looking upstream for skills as well.
In that context, I agree totally with my hon. Friends the Members for Cambridge (Daniel Zeichner) and for Southend East and Rochford (Mr Alaba) that the quality and scale of our talent and skills in our universities and schools is the single biggest determinant of where we end up. I am happy to write to my hon. Friend the Member for Cambridge about the UKRI changes that we are making. In answer to my hon. Friend the Member for Southend East and Rochford, IP capitalisation is a deeply important part of what I focus on with the Intellectual Property Office, and I am happy to engage him on the question of Essex University in particular.