Daily Brief

Ex-Qwen chief Lin Junyang launches $2B agent startup

The Arena

Ex-Qwen chief Lin Junyang launches $2B agent startup

The model wars are quietly becoming founder wars: the architect of Alibaba's Qwen is out with a reported $2 billion startup, the Qwen consumer app just started charging for its best features, and Bernie Sanders wants the frontier labs to stop building. Lin Junyang, the former technical lead of Alibaba's Qwen, has officially unveiled Pragmatik Labs (语用科技) — a Shanghai startup betting on "next-generation agents" spanning both the digital and physical worlds. The August 12 announcement ends five m

China's biggest edge-AI unicorn ModelBest files for A-share IPO

The Arena

China's biggest edge-AI unicorn ModelBest files for A-share IPO

The AI story this cycle is usually written in data centers — but two items overnight point elsewhere: China's largest on-device model maker is heading for the public market, and a new paper hands diffusion language models scaling rules of their own. Beijing ModelBest, China's biggest edge-AI unicorn, has officially started the A-share IPO process, signing a tutoring agreement with CITIC Securities that was posted to the securities regulator's website on August 12. The Tsinghua University NLP la

Supermicro's record year shows AI server demand is still booming

The Arena

Supermicro's record year shows AI server demand is still booming

Two readings of the AI buildout landed overnight: the machine that assembles the servers posted a record year and a stock surge, and a new benchmark suggested AI models still can't do original research — only optimize what humans already built. Supermicro capped fiscal 2026 with record revenue of $39.1 billion, up 78 percent year over year, and fourth-quarter revenue of $11.1 billion, up 93 percent, sending the stock up about 9 percent in extended trading. The headline number for investors was

Alibaba Cloud's CUBE 5.0 builds AI data centers in 100 days

The Arena

Alibaba Cloud's CUBE 5.0 builds AI data centers in 100 days

Two infrastructure stories frame the day: Alibaba Cloud is industrializing the AI data center — compressing builds that once took a year into 100 days — and Ryanair is handing its operations to Gemini agents. Alibaba Cloud says its CUBE 5.0 modular architecture can deliver a large-scale AI data center in about 100 days, a build that traditionally runs six to 12 months in China and 12 to 18 months in the US. The design, announced Monday, raises prefabrication of the five core systems — power, co

Suno hits 100M users as UMG, Sony seek $9B

The Arena

Suno hits 100M users as UMG, Sony seek $9B

Suno became the defining product of AI music on scale alone. A new Bloomberg profile counts the users, the subscribers, and the legal bills piling up behind them. Bloomberg's profile of Suno — the AI music generator more than 100 million people have used since its 2023 launch — lands as the company's legal exposure keeps growing. Suno has crossed 2 million paid subscribers, worth roughly $300 million in annual recurring revenue by CEO Mikey Shulman's own disclosure, and it raised more than $400

Palantir's federal government AI contracts: the $244M no-bid Pentagon deal, explained

The Arena

Palantir's federal government AI contracts: the $244M no-bid Pentagon deal, explained

Two stories this evening trace where AI's center of gravity is moving: Washington is preparing to hand Palantir up to $244 million without a competitive bidding process, and the researcher who built Alibaba's Qwen is leaving to build agents that cross into the physical world. A draft memo from Deputy Secretary of Defense Stephen Feinberg proposes giving Palantir up to $244 million through 2028 for AI-assisted analysis of military production — and it reportedly skips competitive bidding. The int

Phia's founders knew about cookie stuffing for months

The Guardrails

Phia's founders knew about cookie stuffing for months

Phoebe Gates's AI shopping startup just lost its defense. Bloomberg reports that Gates and co-founder Sophia Kianni knew Phia's extension was claiming affiliate credit for sales it never drove — for months, and that they pushed for the features behind the practice. Bloomberg reported Monday that leaked Slack messages and people familiar with the matter show Phia co-founders Phoebe Gates and Sophia Kianni knew their AI shopping extension was "cookie stuffing" — automatically injecting affiliate

Shanghai's 15th Five-Year plan — ¥4 trillion software push led by AI

The Guardrails

Shanghai's 15th Five-Year plan — ¥4 trillion software push led by AI

China's commercial capital is writing its AI ambitions into a five-year industrial blueprint: on Tuesday afternoon Shanghai's economic regulator published its plan for the software and information services sector through 2030, and artificial intelligence drives nearly every section of it. Shanghai's Economic and Information Technology Commission on Tuesday issued its "15th Five-Year Plan" for the software and information services industry, targeting ¥4 trillion (about $560 billion) in industry

Kevin Weil raising $150M for AI science startup

The Arena

Kevin Weil raising $150M for AI science startup

The AI talent exodus from OpenAI keeps producing new companies: former chief product officer Kevin Weil is raising a nine-figure round for an AI-for-science startup, and CoreWeave's earnings showed the infrastructure buildout underneath it all is still accelerating. Kevin Weil, OpenAI's former chief product officer, is seeking to raise $150 million for a new AI science startup at a valuation of at least $750 million, according to three people with knowledge of the discussions. Business Insider

CME to launch GPU rental futures as AI compute goes tradable

The Arena

CME to launch GPU rental futures as AI compute goes tradable

The AI buildout took two steps toward becoming a real financial market today: CME Group is turning GPU rental into a traded commodity, and Super Micro's earnings showed the demand underneath it is still accelerating. Plus, Uber and Serve Robotics — the sidewalk-delivery robot company it helped spin out — are officially done with each other. CME Group will start trading GPU rental futures on October 5, turning the cost of AI compute into a standardized, tradable commodity for the first time. The

OpenAI ships ChatGPT and Codex desktop app for Linux

The Arena

OpenAI ships ChatGPT and Codex desktop app for Linux

OpenAI closed its last desktop gap Tuesday: a preview ChatGPT app for Linux that brings ChatGPT, ChatGPT Work, and the Codex coding agent to the operating system where much of the world's AI development already happens. OpenAI released a ChatGPT desktop app for Linux in preview, ending the platform's years-long status as the odd one out among major desktop operating systems. The app went out worldwide Tuesday, bundling ChatGPT, ChatGPT Work, and Codex, with OpenAI saying the launch "extends Cha

Mojo 1.0 ships — Modular's AI-native language goes stable

The Stack

Mojo 1.0 ships — Modular's AI-native language goes stable

Modular today released Mojo 1.0, the first stable version of Chris Lattner's AI-native programming language, three years after its debut. The release caps a roadmap laid out last December and turns a language long viewed as promising-but-moving into one the company says developers can build on for years. Mojo 1.0 is here: Modular shipped the stable version of its AI-native language on August 11 as part of the Modular 26.5 release. Mojo was launched in 2023 by Lattner — creator of LLVM and Swift