The UK's AI strategy is missing one crucial piece: compute
Reported by TechRadar1 source ↓
TechRadar argues that UK AI companies need more reliable access to computing power, including pooled domestic capacity, alongside government investment and international partnerships.
Key points
- The UK-US AI partnership announced by Prime Minister Andy Burnham is intended to protect critical infrastructure
- The Sovereign AI Fund received more than 200 applications requesting over 67 million GPU hours in three months
- The government has committed over £2 billion to public compute, including a twentyfold expansion of the AI Research Resource by 2030
- TechRadar proposes pooling existing UK computing capacity to make it available to companies
Details
TechRadar reports that Burnham announced a UK-US AI partnership at the UN General Assembly, focused on protecting critical infrastructure.
The article says an Oxford University study found that 32 countries host specialized AI data centers, and that US and Chinese firms operate more than 90% of them. It also cites a 23,000-cluster analysis that found average GPU utilization of 5%, suggesting that some capacity is idle even as other companies struggle to obtain it.
TechRadar says applications to the government's Sovereign AI Fund exceeded 200 in three months, seeking more than 67 million GPU hours. The government has committed over £2 billion to public compute, including plans to expand the AI Research Resource twentyfold by 2030; the article argues that pooling capacity across UK data centers, universities and businesses could give companies another source of compute under UK jurisdiction.
Written from coverage as of
Images
- TechRadar