AI Daily · Archive
AI Daily archive, 7 issue
- 112026-11OpenAI Quarterly Revenue Hits $10B as Capital and On-Device AI Both AccelerateOpenAI's Q3 revenue grew nearly 50% to $10 billion, and the company is seeking a new funding round at a $1.25 trillion valuation. CoreWeave beat expectations and raised guidance, easing market worries about slowing AI demand. Meanwhile, SoftBank, Grab and Petros are advancing more than 5GW of AI infrastructure in Malaysia, and Apple is preparing a Mac app that makes running local large models easier. Capital, compute and on-device AI all moved forward on the same day.
- 102026-11Compute Scheduling and Embodied AI Lead Today's Two StorylinesAlibaba Cloud won a China Patent Gold Award for its compute scheduling patent, putting the AI-era compute tidal problem officially on the table; Ant Lingbo's embodied large-model challenge narrowed to 15 finalists, with a 48-hour real-machine hackathon about to begin. On the consumer side, Apple Intelligence completed filings for tablet and laptop scenarios, while Samsung Galaxy AI was reported to be testing a $4.99/month subscription. Two Japanese labs made moves the same day: Sakana AI was reported to be eyeing an IPO as early as fall 2027, and Shizuku AI unveiled a full-duplex Japanese dialogue model.
- 092026-11AI Moves Into the Lab and the Workflow: From Autonomous Experiments to Agent DeploymentThe day's most notable development is AI shifting from "talking" to "doing": Shanghai Jiao Tong University's AutoDNA appears in Nature, with an AI-native architecture spanning the design, execution and iteration of molecular biology experiments. Tencent, meanwhile, is steering the Agent race toward deployment gaps in data, security and engineering, with chief AI scientist Yao Shunyu saying the moat lies at the product layer. On the compute side, AMD says data center revenue nearly doubled, and China Mobile Cloud has allied with Zhipu on domestic compute and Token efficiency.
- 082026-11Anthropic Races Toward IPO as the Trust Cost of AI Comes Into FocusAnthropic plans to launch its IPO roadshow next week and aims to list before Thanksgiving, but investors are balking at a $1.5 trillion valuation. On the same day, Google was hit with a lawsuit over Gemini's role in a minor's suicide, and an Accenture report was found to contain 15 fake citations. Model capabilities keep advancing — Meta is pushing proactive memory and proactive focus down to Muse and Meta AI, and BioMap teamed up with 11 partners to expand AI-driven biocomputing — but capital and compliance are repricing how fast AI can actually be deployed.
- 072026-11Microsoft Pulls Back from Oracle AI Compute as OpenAI Valuation Hits $1.25 TrillionNew signals are emerging on both the compute and capital fronts: Bloomberg reports that Microsoft is phasing out its use of Oracle's AI computing resources as it reassesses their multibillion-dollar cloud partnership, while OpenAI has raised $11 billion in a new funding round at a $1.25 trillion valuation. Domestically, sources say the CSRC has given window guidance to investment banks to slow IPOs by AI and chip companies. On the research side, arXiv has seen a cluster of papers on multi-agent information design, LLM reviewer drift, and agent engineering.
- 062026-11AI Stocks Hit by Selloff as Compute and Evaluation Draw AttentionOn Nov 5, the U.S. midterm election results triggered a sharp correction in tech stocks, with the Nasdaq 100 falling 6.4% in a single day. AI names bore the brunt of the selling, bringing the market's disagreement over AI asset valuations into the open. On the industry side, Moonshot AI reportedly held talks to rent Nvidia GPU compute in Europe, a sign that securing compute remains a key constraint for model companies. On the research side, a systematic evaluation of LLMs' ability to write literature reviews appeared alongside a resource-orchestration method for Agentic AI, both pointing to reliability problems in real-world deployment.
- 052026-11US Midterms Deal a Blow to AI as Compute Expansion Hits a WallIn the US midterms, voters overwhelmingly backed restrictions on AI and data centers, redrawing the political map of Texas, a key hub for AI data center construction, and the tech giants' $600 million in political spending failed to deliver. On the industry side, OpenAI and Anthropic are diverging in their funding paths, Apple has launched a developer trial of Apple Intelligence in China, and Bill Gates is warning about the risk of AI bioweapons.