Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.
Z.ai's GLM-5.1, MiniMax M2.7, Moonshot's Kimi K2.6, and DeepSeek V4 all landed at roughly the same capability ceiling on agentic engineering benchmarks—at one-third or less the inference cost of Western frontier models. DeepSeek V4 Pro leads Chinese models on NIST's aggregate benchmark but still trails the leading US frontier by approximately eight months.
The two companies formally restructured their partnership, removing Microsoft's exclusive right to host OpenAI models. OpenAI announced GPT-5.4 availability on AWS Bedrock the same day, with GPT-5.5 coming within weeks. Microsoft retains a non-exclusive IP license through 2032 and OpenAI continues paying a 20% revenue share through 2030.
The Defense Department finalized agreements with OpenAI, Google, Microsoft, Nvidia, Amazon, Oracle, SpaceX, and Reflection AI for classified network access. Anthropic remains blacklisted over its insistence on safety guardrails for autonomous weapons and mass surveillance use cases. Anthropic is now contesting the supply-chain risk designation in court.
Google committed $10B cash at a $350B valuation with another $30B contingent on performance targets, plus 5 gigawatts of TPU compute capacity over five years. The deal deepens a partner-rival relationship and follows Anthropic's restricted release of its Mythos model, which identified thousands of zero-day vulnerabilities across major software.
Google agreed to let Gemini models run inside classified military networks for "any lawful purpose," including mission planning and weapons targeting. Over 580 employees—including DeepMind researchers—signed an open letter warning of "irreparable damage," but leadership has dug in, pointing to identical terms accepted by OpenAI and xAI.
VCs have funneled $18.8B into AI startups founded since early 2025. David Silver (ex-DeepMind) raised a record $1.1B seed round for Ineffable Intelligence, Tim Rocktäschel (ex-DeepMind) is raising up to $1B for Recursive Superintelligence, and Yann LeCun left Meta to launch AMI Labs with a $1B raise.
Google, Amazon, Microsoft, and Meta collectively plan $725B in capital expenditures for 2026, up 77% from last year's record $410B. Meanwhile, 80,000+ tech workers lost jobs in Q1 alone. Meta raised AI spend guidance to $125–145B while cutting 8,000 roles; Amazon cut 16,000 corporate positions while AWS posted its fastest growth in 13 quarters.
Profluent will use AI to design custom site-specific recombinases for large-gene insertion therapeutics—a capability beyond conventional CRISPR systems. The deal includes upfront payment, R&D funding, and up to $2.25B in milestones plus royalties. It's Lilly's second recombinase deal this year.
IBM's Institute for Business Value surveyed 2,000 CEOs across 33 countries and found CAIO adoption tripled from 26% in 2025 to 76% in 2026. CEOs expect 48% of operational decisions to be made by AI without human intervention by 2030, and 29% of employees will need reskilling for different roles by 2028.
Anthropic's internal "Project Deal" experiment—where 69 employee-backed agents negotiated 500+ listings—found Opus 4.5 agents systematically out-negotiated Haiku 4.5 counterparts on price and selection. The finding suggests agentic markets inherently reward superior models with hidden premiums, compounding compute advantages.
