Internal memo: Meta plans to begin production of its in-house AI chip, codenamed Iris, in September, and aims to boost its computing power to 14GW in 2027

Techmeme

MiniMax is looking to raise as much as $1.9B by selling 35.6M shares at ~$34 each and offering ~$830M in convertible bonds due 2027 1X says its humanoid robot Neo's five-fingered hands have tendon-style actuators that give it 25 degrees of freedom of movement, nearly the same as a human hand (Boone Ashworth/Wired) How the internet, smartphones, social media, and now generative AI are accelerating a drop in the reading of longer works like books SAP will make it easier for customers to switch to rival service providers or end contracts, averting a possible EU antitrust fine after a September 2025 probe SK Hynix raises $26.5B in the largest ever US market debut by a foreign company, selling 177.9M ADRs for $149 each; the sale was more than 7x oversubscribed Fidji Simo says she has decided to leave her full-time role at OpenAI and transition to being a part-time adviser after her medical condition worsened Mercor acquires Deeptune, which builds reinforcement learning environments for AI agents, three months after CEO Brendan Foody backed Deeptune Sources: Netflix executives are increasingly worried about declining engagement and are exploring adding live TV and bundling streaming services like Peacock Coinbase Federal Reserve Chair Kevin Warsh names Marc Andreessen and Xbox CEO Asha Sharma to lead a task force on the economic impact of new technologies, including AI Google says it will automatically add a disclosure to ads that are made with its AI advertising tools, expanding the disclosure beyond election ads to all ads PitchBook: US venture funding hit $412.7B in H1 2026, up 30% on all of 2025, with AI startup funding accounting for 86%, or $355.9B; Q2 saw seven $1B+ rounds Anthropic, set to add usage-based billing for Fable 5 on July 12, aims to return the model to Claude The EU Parliament advances a bill letting tech companies scan for CSAM, reviving a proposal rejected in March, with an exemption for E2EE services like WhatsApp OpenAI is discontinuing ChatGPT Atlas, its standalone desktop browser, in favor of the new ChatGPT desktop app Sources: AI contractor marketplace Mercor is discussing raising new funds at a roughly $20B valuation, less than a year after raising money at a $10B valuation OpenAI merges Codex and ChatGPT desktop apps for Mac and Windows under a new ChatGPT desktop app, allowing users to switch between Codex, Chat, and Work (Zac Hall/9to5Mac) GPT-5.6 Sol costs $5 per 1M input tokens and $30 per 1M output tokens, GPT-5.6 Terra costs $2.50 and $15, and GPT-5.6 Luna costs $1 and $6 OpenAI broadly releases GPT-5.6, and launches ChatGPT Work, an AI agent that can gather context across apps and files to create documents, on Mac and Windows Microsoft President Brad Smith says the US now has AI “regulation without transparent or complete rules” and “businesses can Kraken Technology, which designs and builds autonomous maritime platforms, such as uncrewed subsurface vessels, raised a $175M Series B at a $1B valuation Bun creator says he rewrote Bun from Zig to Rust using a pre-release version of Claude Fable 5 in 11 days; the task would Character.AI launches three human-written, AI-generated microdramas, whose characters users can chat with, and aims to eventually let users make their own shows Muse Spark 1.1 costs $1.25 per 1M input tokens and $4.25 per 1M output tokens; Alexandr Wang says coding and agentic tasks were key focuses Meta launches a Meta Model API, whose pricing Zuckerberg says will be “aggressive and attractive” at ~25% of the cost of OpenAI and Anthropic models Meta releases Muse Spark 1.1, capable of more advanced coding and a “step-change” from the first generation, available to US developers via a public API preview European startups raised $24B in Q2, up 66% YoY and the strongest quarter in four years; UK startups raised $10.4B, above Germany Anthropic launches a “reflection” dashboard in beta for Free, Pro, and Max users to track Claude usage patterns over 1-, 3-, 6-, or 12-month intervals Databento, which sells financial data feeds, raised a $97M Series B led by NEA, taking its total funding to $127M, and says it is “profitable every month” Micron raises its US capex commitment to $250B through 2035, adding $50B for facilities in NY, ID, VA, and elsewhere, and invests $500M in GlobalWafers Ollama, which helps developers run open-weight AI models locally, raised a $65M Series B led by Theory Venture, following a $15M Series A led by Benchmark PrismML says it ran a 27B-parameter Qwen 3.6 model on an iPhone 17 Pro, bigger than any prior on-device model; sources: Apple held talks with PrismML about it India scraps import duties on some parts used to make phones and other devices until March 2029, removing 7.5% and 5% levies, in a move that might help Apple Apple supplier Luxshare fell in its Hong Kong trading debut, closing down 1.6%, after raising ~$3.1B in the city Enterprise AI agent startup Lyzr is raising a $100M Series B at a ~$500M valuation, up from $250M in a Series A in 2026, and says agents wrote investment memos Internal US DOJ memo: attorneys working on digital currency cases should expect less cooperation from Binance, which is adding requirements for asset freezes US cybersecurity and data resilience company Rubrik plans to invest $500M+ in the UK over the next five years and establish its European headquarters in London Interpol and law enforcement agencies arrest 5,811 suspects and seize $293M in an operation combating social engineering scams and fraud across 97 countries (Sergiu Gatlan/BleepingComputer) Leaked renders via Android Headlines show the Samsung Galaxy Z Flip 8, Z Fold 8, Watch 9, and Watch Ultra 2; the Galaxy Z Fold 8 is likely a new wide foldable (Lawrence Bonk/Engadget) Sources: Sonos' layoffs include some top design and product executives, like VP of Design Dana Krieger; CEO Tom Conrad says the move reduces management layers (Chris Welch/Bloomberg) The AI data center boom has led to surging demand for power transformers, with average lead times for orders, once measured in months, now stretching into years (Financial Times) A Brown University professor suspected his class used AI to cheat after a take-home midterm averaged 96%, prompting an in-person final, which averaged 48.6% (Emma Whitford/Inside Higher Ed) Companies are mobilizing internal groups of "AI champions" to drive adoption; BCG says 74% of front-line employees now use AI regularly, up from 51% in 2025 (Isabelle Bousquette/Wall Street Journal) Shanghai-based GPU maker Iluvatar CoreX raised ~$902M in a Hong Kong share sale; the company's stock has soared 257% since its January IPO, which raised $473M (Bloomberg) Filings: Grubhub parent Wonder is raising hundreds of millions of dollars at a $9B valuation; a source says founder Marc Lore indicated he will contribute $200M (The Information) Cognition releases SWE-1.7, trained from Kimi K2.7 and available at 1,000 tokens/second, claiming it nears GPT-5.5 and Opus 4.8 on benchmarks at a lower cost (Cognition) Block agrees to pay $45M and offer live customer support for Cash App to settle claims by 46 US states that the company failed to protect users from fraud (Daniel Wiessner/Reuters) Sources: Reno-based AI chip startup Positron is in talks to raise ~$750M in two phases, at valuations of $3.5B in the first tranche and ~$5B in the second (Bloomberg) The FTC settles with John Deere over a 2025 right-to-repair lawsuit, requiring the company to provide farmers and shops with equipment and software for 10 years (Boone Ashworth/Wired) OpenAI says it found widespread task issues in SWE-Bench Pro, estimates ~30% of tasks are broken, and retracts its earlier recommendation to adopt the benchmark (OpenAI) Meta announces its first data center in Canada, a 1GW facility in Alberta that will cost the company about $9B and take two to three years to construct (Jonathan Vanian/CNBC) Grok 4.5 is available in Grok Build, in Cursor, and from the SpaceXAI console, but not in the EU; it costs $2 per 1M input tokens and $6 per 1M output tokens (Axios)

2026-07-09 · via Techmeme

此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。原文来自 — 版权归原作者所有。

推荐订阅源

Techmeme