EU Forces Google to Open Android, Oracle Axes 30,000 Jobs for Stargate, and Microsoft's AI Security Fix

Week of July 20, 2026

The European Commission just did what no competitor could: it pried open Google's distribution and data moats. In a binding decision under the Digital Markets Act, regulators ordered Google to open Android to rival AI assistants and share its search data with competing AI developers, hitting the company exactly as its flagship Gemini 3.5 Pro model reportedly missed its third delivery target. Meanwhile, Oracle is cutting up to 30,000 jobs to fund the massive Stargate infrastructure project, and Microsoft used AI to fix 570 vulnerabilities in a single month, hinting at a future where software heals itself.

The EU Breaks Open Google's Android and Search Moats

The European Commission adopted binding requirements ordering Google to do two things it has fiercely resisted: allow rival AI assistants to work natively on Android, and share anonymized ranking, query, click, and view data from its search engine. Eligible third-party assistants will gain voice activation and cross-app capabilities across 11 Android feature groups, subject to certification and user consent. Search data sharing begins in January 2027, with Android interoperability due by July 2027.

This is the most consequential regulatory move in AI this year because it attacks the two assets that make Google nearly unassailable: default placement on over two billion Android devices, and two decades of search behavior data that no competitor can replicate. A rival assistant that can activate by voice and operate across apps loses the structural handicap of not being preinstalled. Search signal handed to AI developers provides training and ranking data they could never otherwise acquire. Google's president of global affairs, Kent Walker, pushed back, warning that the decisions risk undermining privacy and security for millions of Europeans, but the binding nature of the ruling leaves little room for delay.

Google Stumbles as Gemini 3.5 Pro Misses a Third Target

The timing for Google is brutal. Gemini 3.5 Pro reportedly missed its July 17 target—the third slip for the flagship model—and the company is now exploring a stopgap Gemini 3.6 Flash release simply to put something in the market. Alphabet shares fell about 4 percent on the delay reports, and no official model card, pricing, or benchmarks have been published, leaving every circulating claim unconfirmed.

One delay is engineering discipline; three suggests something structural. The unanswered question is whether Google is grappling with safety alignment, data quality, or scaling bottlenecks. For every AI company that is not Google, the combination of regulatory access and model wobble is the best news of the month: a legal path onto Android and access to search signal arriving right as the incumbent looks vulnerable.

Oracle's $500 Billion Stargate Bet Comes with a Human Cost

Oracle is cutting up to 30,000 jobs to fund the $500 billion Stargate infrastructure buildout. The sheer scale of the workforce reduction signals just how much capital intensity the AI infrastructure race now demands. Stargate, designed to be a massive AI compute platform, is pushing Oracle to reallocate resources away from traditional enterprise software toward cloud and AI operations, even at the expense of large segments of its existing workforce.

This restructuring mirrors a broader industry pivot: the infrastructure required to train and run frontier models is becoming so expensive that even major players are forced to sacrifice their legacy cash cows. The question is whether Stargate can generate enough AI revenue to justify the human and financial cost, especially with the EU now forcing market access that could shift demand toward rival assistants and away from the very platforms Stargate is built to serve.

The Bottom Line

This week signals that AI's next chapter will be shaped as much by courtrooms and balance sheets as by model benchmarks. The EU's move to open Android and search data could reset the competitive landscape, especially when combined with Google's own execution struggles. Oracle's dramatic job cuts show that the infrastructure race is becoming a winner-take-most battle with real human consequences. And Microsoft's quiet AI-assisted patching binge hints at a near future where software vulnerabilities are closed before humans even notice them—a capability that, in an era of proliferating AI-powered attacks, might be the most important feature of all.