• Weekly GenAI
  • Posts
  • The GenAI Power Shift: Europe Blinks, Big Tech Consolidates

The GenAI Power Shift: Europe Blinks, Big Tech Consolidates

Let's get started.

In partnership with

This Changes Everything

It’s not an update. It’s a new era for creators.

On November 13, beehiiv is pulling back the curtain on what the future of creating and publishing content online looks like. From creators to publishers to entire media brands, this is the moment everything changes.

At our Winter Release Event, we’ll reveal our vision for building, growing, and earning all in one place, with total control over your audience and your business.

If you make things on the internet, this is the day you finally see what the future holds.

RSVP for our free virtual event now.

Want to appear here? Talk with us

EU AI Act Under Pressure: Brussels Considers Softening Landmark Regulation

Source: Fortune

The European Commission is weighing whether to delay or weaken parts of its groundbreaking AI Act following pressure from the Trump administration and U.S. tech giants. Sources report Brussels has been "engaging" with Washington on potential adjustments to simplify the legislative framework.

Why This Matters:

  • The EU is considering changes to its landmark AI Act amid pressure from Trump administration and tech companies arguing rules are too complex and stifle innovation

  • U.S. Vice President J.D. Vance publicly warned at Paris AI Summit that "excessive regulation" of AI in Europe could cripple the emerging industry

  • Commission launched work on November 5th on code of practice for marking and labeling AI-generated content

The Strategic Shift: Europe's "AI Act as global standard" narrative is fracturing. While the Commission publicly remains "fully behind the AI Act," internal discussions about delays signal the geopolitical AI race is forcing regulatory recalibration.

 

Microsoft Deploys GPT-5 Across Copilot: In-Country Data Processing for 15 Nations

Source: Microsoft

Microsoft announced GPT-5 will become the default model in Copilot Chat in November 2025, powered by GPT-5's real-time router that dynamically selects the right model for each prompt Microsoft Community Hub. Simultaneously, Microsoft announced in-country data processing for Microsoft 365 Copilot interactions in 15 countries by end of 2026.

The Numbers:

  • Australia, UK, India, and Japan get in-country processing by end of 2025

  • 11 more countries including Germany, Canada, Italy, and UAE follow in 2026

  • Copilot now includes prominent clickable citations and improved file summaries in Teams chats

Microsoft's Hedge: With in-house models (MAI-V1, MAI-Vision-1) in development, Microsoft is reducing OpenAI dependency while offering data sovereignty options that could accelerate European enterprise adoption despite AI Act uncertainty.

 

Apple's $1 Billion Bet: Google's 1.2 Trillion Parameter Gemini to Power New Siri

Gemini 3.0 is coming soon

Apple is planning to pay about $1 billion annually for access to Google's ultrapowerful 1.2 trillion parameter AI model to power its long-promised Siri overhaul Bloomberg. The deal marks Apple's most aggressive AI infrastructure move yet.

Why This Matters:

  • Apple evaluated multiple providers before selecting Google

  • Gemini 3.0 expected late Q4 2025 with enhanced multimodal capabilities and superior coding abilities Xole

  • Integration planned for iOS 26, iPadOS 26, and macOS Tahoe

The Irony: Apple, which built its brand on privacy and ecosystem control, is outsourcing its AI intelligence layer to its search rival—acknowledging that building frontier models requires capital and compute Apple hasn't deployed at scale.

 

Google Embeds Gemini Everywhere: Maps Gets Conversational Navigation

Source: The Hill

Google Maps received major Gemini AI upgrades on November 6th, introducing hands-free conversational driving experience with multistep task support. Users can now ask complex queries like "Is there a budget-friendly restaurant with vegan options along my route?"

New Capabilities:

  • Gemini processes review data, websites, and Google Maps pages to provide contextual recommendations The Hill

  • Lens in Maps feature uses visual AI to identify restaurants, shops, and landmarks via camera Tom's Guide

  • Rolling out on Android and iOS over next weeks

The Broader Play: Gemini 3 Pro preview spotted in VertexAI code with November 2025 label, featuring 1M token context window YouTubeBleeping Computer. Google is racing to embed Gemini across every surface before year-end.

 

OpenAI Releases GPT-5-Codex-Mini: 4x Usage at Fraction of Cost

Source: YouTube

OpenAI announced GPT-5-Codex-Mini on November 8th, offering developers 4x more usage than GPT-5-Codex with only minor capability tradeoff Neowin. The model scores 71.3% on SWE-bench Verified versus 74.5% for full GPT-5-Codex.

Developer Impact:

  • Recommended for easier software engineering tasks and when approaching rate limits Neowin

  • ChatGPT Plus, Business and Edu users get 50% higher rate limits thanks to GPU efficiency improvements

  • Available now in CLI and IDE extensions, API support coming soon

Strategic Context: As coding becomes AI's most monetizable use case, OpenAI is creating tiered pricing that mirrors cloud compute economics—optimize for cost or capability, but not both.

 

Five Principles for Navigating November's AI Shifts

  1. Sovereignty Wins Enterprise Deals: Microsoft's in-country processing directly addresses European compliance concerns. If you're selling AI to regulated industries, data residency is now table stakes.

  2. The Model Picker Dies: OpenAI and Microsoft are eliminating model selection UI. Users want "AI that just works"—intelligence routers that auto-select optimal models per task will define 2026 UX.

  3. Frontier Models Need Frontier Budgets: Apple's $1B/year Gemini deal and Anthropic's tens-of-billions TPU commitment prove: competitive AI requires hyperscale partnerships or hyperscale capital.

  4. Regulation Remains Fluid: EU's wavering on AI Act enforcement shows even "final" regulations bend under geopolitical pressure. Build for compliance, but prepare for policy shifts.

  5. Code Generation Consolidates: With GPT-5-Codex-Mini, Claude's SWE-bench dominance, and Gemini 3's coding focus, the message is clear—if your AI strategy doesn't include agentic coding, you're already behind.

Community Spotlight: Best of the Week

Thomas Kurian (Google Cloud CEO): "Anthropic's choice to significantly expand its usage of TPUs reflects the strong price performance and efficiency its teams have seen with TPUs for several years." Japan Today

Mustafa Suleyman (Microsoft AI CEO): "Microsoft is developing humanist superintelligence to deliver all the goodness of science and invention without the uncontrollable risks part."

 

Executive Summary

First Week of November 2025: The AI Power Broker Era

Three Key Developments:

Regulatory Retreat: Europe's AI Act faces its first major pressure test. Brussels engaging Trump administration on "simplifications" signals regulations will evolve based on geopolitical leverage, not just ethical frameworks.

Consolidation Accelerates: Apple-Google's $1B/year deal, Anthropic-Google's multi-billion TPU partnership, and Microsoft's multicloud sovereignty play prove 2026 will be defined by strategic alliances, not solo innovation.

The Coding War Intensifies: OpenAI's Codex-Mini, Claude's SWE-bench dominance (77% vs competitors' 63%), and every major model prioritizing development tools show code generation is AI's first genuine enterprise revenue stream.

Bottom Line: November 2025 marks the end of the "every company builds AI" era. Winners will either (1) own frontier compute, (2) broker access to it, or (3) build compelling applications on top. The middle ground—custom models without hyperscale backing—is disappearing.

Did you enjoy this edition?

Login or Subscribe to participate in polls.

Did you enjoy this newsletter?
Feel free to share it with a friend.

See you in the next one.

Want to sponsor the newsletter? Talk with us