Welcome back to 4IR. Here’s today’s lineup:
EU launches dual AI strategies to catch up with Silicon Valley - Brussels unveils Apply AI Strategy and AI in Science roadmap with new Alliance to coordinate everyone—follows April’s AI Continent Action Plan as Europe tries to move from regulation to actual AI companies
Nvidia CEO confirms $2B xAI bet, says he wanted in deeper - Jensen Huang tells CNBC about equity stake in Musk’s $20B round structured so xAI uses the money to buy Nvidia chips—”almost everything Elon’s part of, you really want to be part of”
Google pushes AI Mode to 75 countries as clickthrough rates crater - Search giant adds 40 regions and 35 languages overnight while Pew data shows AI summaries kill website visits—racing to own “ask anything” before ChatGPT does
EU launches dual AI strategies to catch up with Silicon Valley
The story: The European Commission announced two AI strategies today aimed at making Europe a global AI leader. The Apply AI Strategy focuses on getting AI deployed faster in industry and public sector. A separate AI in Science strategy targets research. They’re standing up an “Apply AI Alliance” to coordinate industry, government, academia, social partners, and civil society. An AI Observatory will track trends and measure impact. Both strategies follow April’s AI Continent Action Plan and build on Europe’s AI Act. The stated goal is accelerating time-to-market and making the EU workforce “AI-ready.”
What we know:
Apply AI Strategy aims to speed up deployment by connecting infrastructure, data, and testing facilities
New “Frontier AI initiative” will coordinate Europe’s leading AI companies
Apply AI Alliance brings together industry, public sector, academia, social partners, civil society
AI Observatory will monitor AI trends and assess sectorial impacts
Follows April 2025 AI Continent Action Plan
Goal includes strengthening EU workforce to be “AI-ready across sectors”
Built on existing AI Act regulatory framework
Why it matters: Europe spent three years building AI regulations while OpenAI and Anthropic built the actual AI industry. Now they’re trying to catch up. The “Frontier AI initiative” language is telling—Europe doesn’t have companies at the frontier right now, so they’re building the structure hoping it produces them. If Apply AI can’t show real progress by 2026, the window probably closes.
Europe’s playbook has been “regulate first, build later” but you can’t regulate your way to AI leadership when you don’t have anything to regulate yet. The real problem is Europe lacks the risk capital, talent concentration, and data scale that made Silicon Valley work. So they’re trying a different angle—win on industrial deployment in traditional European strengths like automotive and manufacturing rather than compete on frontier models. Smart pivot in theory. In practice, CTOs won’t choose “Europe-approved AI” if it means running 12 months behind. The coordination bodies sound impressive but Europe’s track record on these is rough—remember European cloud computing? EU semiconductor fabs?
Nvidia CEO confirms $2B xAI bet, says he wanted in deeper
The story: Nvidia CEO Jensen Huang confirmed on CNBC today that the company invested in Elon Musk’s xAI as part of a $20 billion funding round. The deal structure: roughly $7.5B equity and $12.5B debt through a special purpose vehicle. Here’s where it gets circular—the SPV buys Nvidia chips and leases them to xAI for five years. So Nvidia invests money that xAI then uses to buy Nvidia products. Huang’s take: “I wish I gave him more money” and “almost everything Elon’s part of, you really want to be part of.” The chips go to xAI’s Colossus 2 data center in Memphis.
What we know:
Nvidia putting up to $2B equity in xAI’s $20B round
Deal split: ~$7.5B equity, ~$12.5B debt via special purpose vehicle
SPV purchases Nvidia GPUs, leases them to xAI for 5 years
Chips allocated to Colossus 2 Memphis facility (100MW)
Apollo Global Management and Diameter Capital Partners in on debt, Valor Capital leading equity
Bloomberg reports xAI burning ~$1B monthly
Musk denied raising money in September, round got bigger since
Huang told CNBC his only regret is not investing more
Why it matters: This is vendor financing dressed up as venture capital. Nvidia invests in xAI, xAI buys Nvidia chips with that money, Nvidia books revenue. The debt is smart though—it’s collateralized by the chips themselves, not xAI’s business. So if xAI crashes, lenders seize GPUs they can flip in today’s tight market. But the circular nature is hard to ignore, especially since this confirmation came the same day financial analysts are comparing AI infrastructure deals to dot-com era arrangements.
The $1B monthly burn means xAI is spending $12B a year just to stay alive—that’s more than most unicorns raise total. Huang’s enthusiasm either means he knows something about Grok’s progress or he’s deep in the Musk reality distortion field. The math gets weird fast when you trace it: Nvidia invests → xAI buys Nvidia chips → Nvidia’s revenue goes up → they invest more. At some point this stops being real economic activity. The key difference from 1999 is Nvidia has a massive profitable business generating actual cash. But if xAI can’t reach profitability, someone’s holding an expensive bag. The fact Huang wanted to invest more is either supreme confidence or a warning sign.
Google pushes AI Mode to 75 countries as clickthrough rates crater
The story: Google rolled out AI Mode to 40 new regions and 35 new languages today in its biggest expansion yet. The new languages include Arabic, Chinese, Croatian, Czech, Dutch, German, Greek, French, Malay, Russian, Thai, Vietnamese—basically every major non-English market. Google says its custom Gemini model can “grasp the subtleties of local languages” to avoid bad translations. This comes as Pew Research data shows users who see AI summaries at the top of search results are less likely to click through to websites and more likely to just end their session. Publishers are already blaming AI Mode for declining traffic.
What we know:
AI Mode now in 40 additional regions beyond US (started May 2025)
35 new languages added including major markets
Custom Gemini model powers the search with “advanced reasoning and multimodal understanding”
Visual search features updated September 30
Pew study shows AI summaries reduce clickthrough and end browsing sessions earlier
Publishers say AI Mode summaries are killing their traffic
Started testing in Labs program March 2025
Why it matters: Google is racing to make AI search the default before ChatGPT can steal that position. The language expansion targets markets where ChatGPT adoption is weaker—get users hooked on AI Mode before they build habits elsewhere. But there’s a fundamental problem: if AI Mode works too well, Google destroys its own business model. Users see the summary, get their answer, leave. No clicks means no ads means no publisher revenue means fewer quality websites means worse training data for future AI.
Google is trying to be the AI search leader while keeping the web ecosystem alive that makes search valuable. These two goals can’t coexist. If AI Mode gives perfect answers, nobody clicks through. If answers are incomplete, users go to ChatGPT or Perplexity instead. The 35-language expansion suggests Google picked its strategy—race forward now, figure out the publisher problem later. The language coverage is where Google’s real advantage shows up—most AI labs train on English and bolt on other languages awkwardly. Google has decades of multilingual data baked into infrastructure. That’s the moat. But the Pew data is brutal for publishers watching traffic disappear. At some point they’ll stop letting Google scrape their sites, and then what?
Note: Commentary sections are editorial interpretation, not factual claims