Nvidia Pledges $250M to UK AI Safety Research and Announces $2B More in European Investment as Stock Slides on Inflation Fears

Nvidia committed $250 million over five years to advance UK agentic AI safety and security research, structured as $50 million in annual compute access donations. CEO Jensen Huang separately announced that the company will disclose an additional $2 billion in AI investments across Europe within the coming week, layering on top of the $8 billion Nvidia has already deployed in the region over the past year. Huang also said he would cut off chip supply to any customer found to be building unsafe AI, framing safety concerns as normal industry growing pains.

The announcements came alongside a flurry of new partnerships. Nvidia teamed with ChipAgents to launch Build-A-Chip-Bench, an evaluation program for AI-powered chip design; ChipAgents had previously raised $134 million in Series A funding. Nvidia also unveiled cuPotential, a new CUDA-X library it says speeds GPU workflows 2 to 3 times in early access, and released a broader suite of CUDA-X AI tools targeting semiconductor design and manufacturing, with partners including Synopsys, Cadence, Applied Materials, Lam Research and KLA reporting speed improvements of up to 50 times in selected workflows. Honeywell launched its Forge Cognition agentic AI and simulation platform built with Nvidia and Dell Technologies to manage operations across buildings and industrial facilities. London Stock Exchange Group announced a partnership with Nvidia to accelerate tick history analysis, while Danaher said it would collaborate with Nvidia to advance autonomous science. In France, Huang announced investments including equity stakes in Doctolib and Genesis AI.

Nvidia shares fell roughly 2.9 percent on the day as hotter-than-expected inflation data rattled AI-related stocks broadly. A key index tracking major chipmakers including Nvidia and Broadcom dropped 20 percent from its June high, with investors citing rising interest rates, stretched valuations and mounting concern over the cost of AI infrastructure. CoreWeave bonds slid to distressed levels, and the company's chief described market sentiment as irrational. Nvidia's share of China's data-center AI chip market fell to 34 percent in the first half of 2026, down from prior levels, while Huawei climbed to 28 percent and smaller Chinese rivals including Alibaba's T-Head, Cambricon, Baidu's Kunlunxin and Hygon took additional share. Huang has previously warned that US export controls are effectively ceding China's estimated $50 billion AI chip market to competitors.

Analyst Dan Ives pushed back against the selloff, calling it way overdone and citing an estimated 13-to-1 chip demand-to-supply imbalance. Nvidia trades at roughly 14 times fiscal 2028 consensus earnings per share, a valuation bulls call inexpensive given the company's 83.4 percent revenue growth rate and its annualized return of 63.8 percent since the end of 2020, compared with 27.2 percent for Alphabet, 17.7 percent for Apple and 16.6 percent for Microsoft over the same period.