AI Daily Briefing

Google & Microsoft's $475B Capex Lock-In: Who Controls AI's Physical Layer

AI Daily Briefing

Show Notes

(00:00:00) Google & Microsoft's $475B Capex Lock-In: Who Controls AI's Physical Layer

(00:00:58) Azure and Google Cloud Earnings

(00:01:34) Google Gemini 4 Carbon Tops Rankings

(00:02:04) General Compute Alternative Inference Cloud

(00:02:43) AI Agents Enter Mission-Critical Operations

(00:03:45) Domestic Manufacturing Capital Wave

Google and Microsoft have drawn a hard line around the physical foundation of AI. With Alphabet committing over $300 billion and Microsoft $175 billion annually to data centres, compute, and networking, the infrastructure layer of artificial intelligence is consolidating around just two companies — and today's earnings numbers confirm it's accelerating.

Azure grew 48% last quarter. Google Cloud doubled to $30.7 billion in revenue. The Gemini app crossed one billion users, and Google Cloud is now processing 32 billion tokens per minute — four times the rate from a year ago. Against that backdrop, Google released Gemini 4 Carbon, which now sits first on the Artificial Analysis Intelligence Index with an 88.2% score on cybersecurity vulnerability detection.

One layer up from infrastructure, AI agents are leaving pilot programs behind. Hyphenate reported 80x ARR growth automating corporate accounting. Ultra's warehouse robots have processed over 500,000 orders across four states. And Causal Security's autonomous agent Cipher discovered four critical vulnerabilities — including a CVSS 9.8 severity flaw — in systems running over 90 particle accelerators and fusion reactors, in under four hours.

Meanwhile, General Compute raised $100 million to build an alternative inference cloud on Cerebras and SambaNova processors, and Re:Build secured $2.1 billion from General Catalyst and JPMorganChase for U.S. industrial manufacturing capacity.

The capex has been committed. The agents are live. The open question now is whether audit, compliance, and liability frameworks can catch up before something goes wrong in a high-stakes environment.

This episode includes AI-generated content.

More from AI Daily Briefing

Oct 28, 2026

(00:00:00) Anthropic's $80B Loss IPO, Meta's $95B Raise & Genie 4 Robots

Oct 27, 2026

(00:00:00) Anthropic S-1 Deep Dive: $80B Losses, FTC Probe & Claude's Rogue Move

Oct 26, 2026

(00:00:00) AWS Sold Out to 2028, Project Meridian & Agent Containment Failures

Oct 25, 2026

(00:00:00) Black October: PHP Crisis, OpenAI's Patch Gap & Meta's AI App

Oct 24, 2026

(00:00:00) Mythos 5, GPT-6 Monitoring Evasion & the Pentagon's $700M AI Bet

See all episodes →