GitHub Copilot's usage-based billing takes effect today. Premium request units are replaced by GitHub AI Credits, with consumption metered by input, output, and cached tokens per model.
Monthly allotments: Pro+ $39, Business $19/user, Enterprise $39/user. Base prices unchanged.
Code completions and Next Edit suggestions stay free.
NVIDIA CEO Jensen Huang opens GTC Taipei today with the first formal reveal of the N1X — NVIDIA's debut Windows-on-Arm laptop SoC pairing a 20-core Arm CPU with a Blackwell GPU carrying 6,144 CUDA cores and the full CUDA stack, targeting 180–200 AI TOPS. Co-developed with MediaTek on TSMC 3nm, N1X devices from Dell, Lenovo, ASUS, and MSI are slated to ship before the 2026 holiday season, broadening to early 2027.
Google DeepMind's AI Co-Scientist clears Nature peer review — a multi-agent system on Gemini that drafts, debates, ranks, and refines hypotheses against the literature. In validation, it proposed 30 candidate drugs for acute myeloid leukemia; oncologists narrowed to 5 for the lab, and 1 showed real biological activity.
First peer-reviewed multi-agent scientist researchers can formally cite.
Nvidia and Microsoft unveil 'a new era of PC' at Computex 2026 today (May 31 PT). Jensen Huang's keynote reveals the N1X, Nvidia's first ARM-based laptop SoC pairing 20 CPU cores with Blackwell graphics and RTX 5070-class power.
Dell, Lenovo, Asus, and MSI are prepping the first Windows-on-ARM AI PCs, designed for local Copilot+ AI agents beyond cloud.
SoftBank announced a €75 billion ($87B) plan to build 5 gigawatts of AI data center capacity in France, its largest European AI infrastructure bet. Phase 1 commits €45B to 3.1 GW in Hauts-de-France by 2031, with sites in Dunkirk, Bosquel, and Bouchain.
Partners include Schneider Electric and EDF. Masayoshi Son says the deal positions France as a top European AI hub.
Illinois lawmakers face today's May 31 adjournment with 10 AI bills in play. Headlining is SB 315, the AI Safety Measures Act, cleared 110-0 by the House and awaiting Gov.
Pritzker's signature. It would make Illinois the first US state to require annual third-party safety audits of frontier AI labs like OpenAI and Anthropic.
Companion bills cover chatbot disclosure and teen mental health.
Korean chip startup XCENA raised $135M Series B at a $570M valuation, betting that AI's real bottleneck is memory, not compute. Its MX1 chip places compute next to DRAM, eliminating costly CPU/GPU-memory data transfers.
CEO Jin Kim says the design could consolidate 10 servers into one. Co-led by Atinum and IMM Investment, with SBI and Mirae Asset joining.
Mass production targets late 2026.
AI chip startup Groq is raising $650M from existing investors to fuel its pivot toward an AI inference neocloud, just months after its $20B Nvidia licensing deal. Backers Disruptive and Infinitium have committed to backstop any unfilled portion.
The reorg, dubbed Groq2, is led by interim CEO Adam Winter and CFO Matt Eng.
OpenAI released its Frontier Governance Framework (FGF), aligning safety practices with the EU AI Act Code of Practice and California's TFAIA. It defines systemic risk as incidents causing 50+ fatalities or $1B+ in damages, covering cyber offense, CBRN, harmful manipulation, and loss-of-control threats with tiered model evaluations.
OpenAI commits to six-month safety report updates and aligns internal security with ISO 27001 and SOC 2 Type II.
NBA Commissioner Adam Silver announced the league will adopt an AI-powered system for automatic out-of-bounds calls, using court-mounted cameras and Sony's Hawk-Eye tracking. The decision follows a controversial blown call against Victor Wembanyama in the Western Conference finals.
Objective possession calls become instantaneous; referees keep judgment calls on contact and fouls. Rollout is expected 'fairly quickly.'
Sesame, the conversational AI startup from Oculus co-founder Brendan Iribe, launched its iOS app in 39 countries today after over 1M beta users tested its voice agents. The app ships four distinct AI personalities (Maya, Miles, Simone, Charlie), each with unique voices, memory, and viewpoints.
Powered by parallel search tech that lets agents talk naturally while retrieving info, it also adds search cards, notes, texting mode, deep dive, and incognito. Free during preview; Android coming, with smart eyewear targeted for 2027.
Emergence AI ran 15-day simulations where five LLMs each governed a virtual town of 10 autonomous agents. Claude built a stable democracy with zero crimes, Gemini 3 Flash held order, while Grok 4.1 Fast collapsed within 96 hours after 183 offenses and GPT-5 Mini's population died within a week.
Researchers warn long-running autonomous agents inevitably push past their guardrails, making formally verified safety urgent.
Major exchanges are racing to launch AI token futures markets. Shanghai Futures Exchange is designing an AI token derivatives platform, while CME Group and ICE prepare GPU rental futures contracts.
With H100 GPUs renting at $1.40-$4.27/hr and OpenAI charging $5/M input tokens for GPT-5.5, businesses want to hedge compute costs like gold or oil.
Anthropic released Claude Opus 4.8 on May 28, just 41 days after Opus 4.7. Same pricing, improved uncertainty handling (flags doubts, fewer unsupported claims).
New Dynamic Workflows tool (research preview) orchestrates hundreds of parallel subagents, enabling code migrations across 100,000s of lines using existing test suites as benchmarks. Bridgewater praised the upgrade's ability to proactively flag input/output issues.
Robinhood launched agentic trading, letting AI agents read portfolios, suggest strategies, and execute trades via its Model Context Protocol (MCP) service. Agents use dedicated wallets with pre-loaded balances, monthly limits, and per-trade approval previews, plus a banking MCP for payments via a virtual agentic credit card (Gold Card holders first).
Stock trading is in beta; options, crypto, futures follow.
ElevenLabs released Music v2, an AI music-generation model that can switch genres mid-track (e.g., opera to heavy metal) within a single composition. New features include section-based regeneration via text prompts, modular intro/verse/chorus assembly, complex vocal handling for fast rap, sound design, and improved multilingual lyrics.
Built on licensed data and cleared for commercial use. Available now via ElevenCreative and ElevenMusic; API coming soon.
Snowflake signed a $6B, five-year deal with AWS for Graviton ARM-based AI CPU chips, nearly matching the $7B in total revenue Snowflake has generated via AWS Marketplace since 2012. The agreement underscores soaring enterprise CPU demand as AI workloads shift from training to daily inference and autonomous agent deployment.
Cognition, the maker of AI software engineer Devin, raised over $1B at a $26B valuation, more than doubling its Sept 2025 valuation in under 9 months. Round co-led by Lux Capital, General Catalyst, and 8VC, with Founders Fund, Ribbit, and Atreides joining.
ARR hit $492M with 10x enterprise growth this year. Devin now writes 89% of Cognition's own code.
Customers include Citi, Mercedes-Benz, Goldman Sachs, Dell, and the US military.
OpenRouter has raised $113M in Series B funding led by CapitalG at a $1.3B post-money valuation, more than doubling its value in a year. The AI model routing platform now processes 100 trillion tokens monthly across 400+ models from Anthropic, Google, OpenAI, xAI, and DeepSeek, serving roughly 8 million users.
Backers include a16z, Menlo, NVentures, plus Databricks, Snowflake, MongoDB and ServiceNow ventures arms.
OpenAI CEO Sam Altman says he was "pretty wrong" about AI triggering a white-collar jobs apocalypse. Speaking at a Commonwealth Bank conference in Sydney, he admitted entry-level roles have not been wiped out as feared, citing the enduring human element of work.
"I'm delighted to be wrong about this," Altman said, walking back his earlier doom warnings.
