Google I/O kicks off today with what could be its most consequential keynote in years, Anthropic is quietly negotiating a deal that would make it more valuable than OpenAI, and Microsoft's AI chief just put an 18-month countdown on your office job. Meanwhile, the US government has quietly locked every major AI lab into pre-release testing, and a DeepMind spinout just raised $2.1B to prove AI can design drugs, not just chatbots.
Google I/O 2026 Keynote Kicks Off Today With Expected Gemini 4.0 Reveal and Android XR Glasses
Google's annual developer conference opens today at Shoreline Amphitheatre with its most AI-heavy keynote yet. Expected announcements include a major Gemini model upgrade (likely Gemini 4.0), a preview of Android XR smart glasses with Samsung, Warby Parker, and XREAL, and a leaked 'Gemini Omni' model capable of generating text, images, and video in a single pipeline. With Anthropic and OpenAI trading benchmark leads, Google needs to prove that distributing Gemini across billions of devices matters more than winning any single benchmark crown.
Anthropic in Talks to Raise $30B at $900B+ Valuation, Would Surpass OpenAI
Bloomberg reports Anthropic is negotiating at least $30 billion in new financing at a valuation above $900 billion, which would nearly triple its $380 billion February valuation and leapfrog OpenAI's $852 billion. The round, expected to close by month's end, is being led by Greenoaks, Sequoia, Altimeter, and Dragoneer. Anthropic's annualized revenue has hit $14 billion with Claude Code alone generating $2.5 billion, and the company now has over 1,000 customers spending $1M+ annually. The capital is earmarked for compute infrastructure as Anthropic races to lock in capacity before Google I/O reshapes the competitive landscape.
Microsoft AI Chief Mustafa Suleyman: All White-Collar Work Will Be Automated Within 18 Months
In a Fortune feature published May 16, Microsoft AI CEO Mustafa Suleyman predicted that most tasks involving 'sitting down at a computer' will be fully automated by AI within 12 to 18 months, naming accounting, legal, marketing, and project management as vulnerable. However, the evidence so far is mixed: a METR study found AI actually made software developers 20% slower, Apollo Global Management data shows profit margin gains are confined almost entirely to Big Tech, and a Gartner study found AI-driven layoffs are failing to generate returns. The prediction matters less for its accuracy than for what it reveals about Microsoft's strategic direction toward building 'superintelligence' independently from OpenAI.
All Five Frontier AI Labs Now Under US Government Pre-Release Model Testing
NIST's Center for AI Standards and Innovation has finalized pre-deployment evaluation agreements with Google DeepMind, Microsoft, and xAI, joining existing deals with OpenAI and Anthropic. Every major frontier model in the US must now go through government security evaluation before public launch. CAISI has already completed over 40 model evaluations, including unreleased models, and developers often provide versions with reduced safeguards for thorough testing. The move was catalyzed by Anthropic's Mythos model, which pushed cybersecurity concerns past a tipping point.
DeepMind Spinout Isomorphic Labs Raises $2.1B Series B to Scale AI Drug Discovery Engine
Isomorphic Labs, founded by DeepMind co-founder and Nobel laureate Demis Hassabis, closed a $2.1 billion Series B led by Thrive Capital with participation from MGX, Temasek, CapitalG, the UK Sovereign AI Fund, and existing backers Alphabet and GV. The capital will scale its AI drug design engine (IsoDDE) and push therapeutic programs toward clinical trials. With strategic partnerships already in place with Novartis, Lilly, and Johnson & Johnson, the raise signals institutional investors now view AI drug discovery as a near-term commercial category rather than a research curiosity.
Test Your Understanding
Quiz
1 / 8
Which hardware partners are expected to showcase Android XR glasses at Google I/O 2026?
A federal judge just told the Pentagon it cannot punish an AI lab for refusing to hand over its models, which is the first time a court has drawn a hard line around what safety policies cost you in Washington. Elsewhere the theme is AI learning by watching rather than being told: a robot foundation model that copies a ten-minute task from one video, and Claude grinding through algebra that stumped a lab for eighteen months. Plus Meta patching the dumbest privacy hole in its glasses, and Pew's number on how many Americans now ask a chatbot about their symptoms.
Read full summary & take a quiz →
Top Stories
A Judge Just Ruled the Pentagon Illegally Punished Anthropic for Saying No to Mass Surveillance
Skild's Robot Model Watches One Video of You Making Pancakes and Then Makes Pancakes
Claude Did Eighteen Months of Algebra in Five Weeks, and Got Some of It Wrong Along the Way
Meta Is Finally Killing the Trick Where You Start Recording Then Cover the Light
A Third of Americans Now Ask a Chatbot About Their Symptoms, and Most Won't Share Their Health Data With It
