State of AI: November 2025 newsletter

Dear readers, Welcome to the latest issue of the State of AI, an editorialized newsletter formerly known as Guide to AI that covers the key developments in AI policy, research, industry, and start-ups over the last month. First up, a few reminders:State of AI Report 2025: At over 300 slides, you...

Dear readers, Welcome to the latest issue of the State of AI, an editorialized newsletter formerly known as Guide to AI that covers the key developments in AI policy, research, industry, and start-ups over the last month. First up, a few reminders:State of AI Report 2025: At over 300 slides, you can watch my 25 min summary of the report to digest our key findings. Podcasts! I discussed the report further on the MAD podcast (YouTube/Apple/Spotify), on TechBio Talks (YouTube/Apple/Spotify) and on Hidden Forces (YouTube/Apple/Spotify).AI meetups + RAAIS 2026: Join our upcoming AI meetups in London (2nd Dec ‘25), Munich (17 Feb ‘26) and Zurich (19 Feb ‘26) as well as our 11th Research and Applied AI Summit on 12 June 2026.Air Street Press featured Poolside’s path into AI’s power infrastructure, PARIMA’s cultivated meat approval in Singapore and acquisition of Vital Meat, and Air Street’s partnership with NVIDIA to supercharge the UK’s AI ecosystem with £2B.I love hearing what you’re up to, so just hit reply or forward to your friends :-)AI as national infrastructureThe media is drumming up the AI bubble narrative (again). Here, I’d point you to two essays: last October, we wrote how “AI isn’t the dotcom bubble”, and this past week, Stratechery ran an essay on “The benefits of bubbles”. In short, research is delivering real, repeatable breakthroughs, the adoption of AI in the enterprise is significant, while hyperscaler and AI lab revenues are already huge as we show in the State of AI Report 2025. Even a potential “overbuild” mostly drives costs down and catalyzes durable assets like fabs and power, rather than signaling collapse.Governments around the world continued to treat AI as critical infrastructure, though with differing extents of seriousness. In the United States, a public–private partnership between the Department of Energy (DOE) and AMD to build new “AI Factory” supercomputers at Oak Ridge National Laboratory reflected an entrenched belief that compute capacity is a matter of sovereignty. The Lux and Discovery systems, expected in 2026 and 2028 respectively, will expand federal AI capabilities under a roughly $1B budget shared between public and private funding. The same logic of scale drove NVIDIA’s $1B investment in Nokia, which made the U.S. chipmaker the telecoms company’s second-largest shareholder. The partnership aims to integrate Nokia’s networking technology with Nvidia’s data-centre hardware, a sign of how industrial and commercial agendas are converging.On the other side of the Atlantic, the European Commission unveiled a €1B “Apply AI” strategy intended to reduce dependence on U.S. and Chinese technology, alongside a broader €1.1B package to ramp up AI in key industries. The plan channels funds from Horizon Europe and Digital Europe toward healthcare, manufacturing, energy and defence. However, at this scale, the initiative is negligible compared to the hundreds of billions being deployed by the U.S. and Asia, if not the $1.5T announced by JPMorgan for Ameican security and resilience. Indeed, Europe’s AI market is dominated by U.S. cloud providers and high energy costs, slow buyers and few national champions impede an “AI-first” transition. Without vastly greater capital and faster regulatory reform, Apply AI risks being symbolic rather than transformative. Compounding the uncertainty, the EU was also reported to be weighing a pause on parts of its landmark AI Act amid pressure from U.S. Big Tech. Still, Commission President Ursula von der Leyen emphasised the need for a sovereign industrial base, hinting that Europe will have to coordinate energy, compute and policy to compete. Sounds good, show us the goods!Over in Saudi Arabia, which hosted their latest “Davos in the Desert”, it is clear that the kingdom is bidding to become a global exporter of AI compute. The NYT reported that they’re planning a $5B Red Sea data‑center complex (with another multibillion‑dollar site on the east coast), which aims to handle ~6% of global AI workload (from

Source: Air Street Capital — Published November 9, 2025 at 17:39 UTC — Category: Business