Draghi tells Europe to build AI data centres — it holds 5% of compute

Share
Draghi tells Europe to build AI data centres — it holds 5% of compute

Two stories out of this week's Inclusion Conference and the European policy pages: Mario Draghi is back with an industrial plan, and Ant is quietly proving that the agent business is sold deal by deal.

Mario Draghi is calling for a large-scale buildout of AI data centres inside Europe, arguing that sovereignty over data is now a growth question rather than a regulatory one. The former Italian prime minister and ex-ECB president published a long essay this week urging the bloc to stop depending on American cloud for inference, and the numbers behind the plea are worse than the rhetoric: a Bruegel policy brief estimates Europe's AI compute capacity at roughly 2 gigawatts in 2026, about 5% of the global total, against roughly 35 GW in the United States (78%) and 5 GW in China (11%). Even if global AI compute grows more than eightfold by 2031, Bruegel's authors project Europe's share creeping up only to about 5.6%. The uncomfortable detail is that the money is not the binding constraint — the European data-centre industry's own 2026 survey puts the blockers on grid readiness and permitting, not private capital, and the five AI gigafactories the Commission is funding add up to around 750 megawatts, roughly 4% of the 21 GW Bruegel models Europe will need in 2031. The brief's most actionable claim is also the least glamorous: shave a single year off infrastructure consent processes and Europe's 2031 share rises to about 6.7%. That lands a week after the bloc's first concrete site — Nokia gives Europe's AI gigafactory a street address — so Draghi is arguing into a debate that has already moved from white papers to land purchases. The read: Europe's compute gap is a construction-permit problem dressed as an industrial-policy problem, and no amount of InvestAI money changes the arithmetic until approval clocks do.


Ant Digital Technologies closed 31 partnerships at the Inclusion Conference in Shanghai, spanning green energy, smart terminals, hotels and tourism, and digital content. The individual deals are unglamorous — AI power-trading and storage assets on-chain in energy, robot-staffed retail kiosks and industrial AI glasses on the device side, an AI hotel concierge and ticketing agent in tourism, and a joint copyright model with Fujian-based Vobile covering text-to-video and short-drama work. What adds up is the shape of the strategy: Ant is not selling model access, it is embedding agents into workflows that already have invoices attached, alongside its Agentar enterprise-agent platform and the Lingbo robotics "brain" it demoed sorting medicine in an 80-centimetre pharmacy aisle. Ant Digital's chief executive Zhao Wenbiao told the conference that the unit's overall business grew close to 50% a year over the past two years, with AI enterprise revenue growing at triple-digit rates. The honest caveat comes from Ant's own leadership: asked whether the agentic-commerce "ChatGPT moment" he predicted was visible in the company's backend data, Ant's Han Xinyi answered that it was not.

What to watch: whether the Commission's Cloud and AI Development Act actually commits to the 18-month permitting target, and which consortiums make the gigafactory shortlist.

Is Europe's AI compute gap fixable by permits and power lines, or is the train already gone? Tell us in the comments.

Read more

Brain organoids classified Braille at 83% — in a trio, not alone

Brain organoids classified Braille at 83% — in a trio, not alone

Two stories today: human brain tissue on a chip decoded a tactile signal, and the data-centre company behind some of the largest AI contracts in Europe turns out to be six months behind on its flagship site. Human forebrain organoids grown on a microelectrode array classified Braille letters with 61 percent accuracy on their own and 83 percent when the responses of three organoids were combined, in a peer-reviewed study from the University of Bristol. The team took event-based tactile readings

Suleyman says Microsoft's base model beat OpenAI's on a quarter compute

Suleyman says Microsoft's base model beat OpenAI's on a quarter compute

Microsoft finished pre-training its biggest in-house model yet — then used it to make a point about the company it bankrolls. Microsoft AI completed pre-training of MAI-Base-2, a mixture-of-experts foundation model with 80 billion active and 2.8 trillion total parameters, and says the recipe is 4.2–12.3x more pre-training FLOPs-efficient than DeepSeek-V4-Pro-Base. The run itself is the number worth keeping: 16,000 Nvidia GB200 GPUs for 38 days across August and September, at 21% average model

Amazon's new Bee devices listen all day — for $11.99 a month

Amazon's new Bee devices listen all day — for $11.99 a month

Amazon turned its always-listening Bee line into a two-device family on Thursday, and NVIDIA opened a 2.8-trillion-token Indic corpus recipe. The connecting thread: both are bets that the next AI money is in memory — yours, or your model's. Amazon put two new always-listening devices on sale, and both now lead to a subscription. Bee Frontier, the next-generation wristband that also clips to clothing, ships now at $74.99 in Onyx, Lavender and Sage — up from the $50 predecessor — with seven-day

Perplexity open-sources its agent sandbox after 118 failed escapes

Perplexity open-sources its agent sandbox after 118 failed escapes

Perplexity handed its agent containment layer to the open-source world, and Washington's AI task force stopped reading briefings and started working. Perplexity open-sourced SPACE on Thursday — the sandbox it built to keep its own AI agents from getting out — and says nine frontier models failed to break out of it across 118 attempts. Each agent runs inside a small virtual machine with its own operating system, with credentials and network access managed outside the sandbox so an agent can re