01
TechCrunch AI ↗
Anthropic has signed a $10 billion, six-year agreement with Volta, an AI cloud startup founded earlier in 2026 that Reuters reports is now valued at $2.4 billion. Volta will supply compute from a 133 megawatt data center in Norway, built with crypto mining firm Bitdeer and running on Nvidia's Vera Rubin chip architecture. The deal extends an aggressive compute buildout at Anthropic, which has struck recent capacity partnerships with SpaceX and Amazon as it competes with rival labs.
02
The Verge ↗
SpaceX booked more revenue from AI than from space in its first quarterly report since its June IPO, with AI revenue more than tripling year over year to $2.6 billion. Total second quarter revenue rose 92% to $7.8 billion, beating Wall Street's $6.81 billion estimate, though the AI segment still posted a $1.26 billion operating loss. Most of that revenue came from deals selling compute capacity to other AI companies, and the stock fell more than 5% after hours.
03
The Verge ↗
AMD's data center revenue more than doubled year over year to $6.7 billion in the second quarter, up 107% from $3.2 billion a year earlier and up from $5.8 billion in Q1, driven by demand for AI capacity. Total revenue rose 50% to $11.5 billion against estimates of $11.3 billion, but the company's third quarter forecast landed below some analyst expectations and the stock dropped more than 7% after hours. Gaming revenue fell 31% over the same period as the business tilts further toward AI accelerators.
04
TechMeme ↗
The Ninth Circuit Court of Appeals overturned a lower court order that had temporarily barred Perplexity from running its AI powered agentic shopping tools on Amazon's platform. The court found Amazon was unlikely to succeed on its claim that the agents violated the federal computer hacking statute, reasoning that it was Perplexity's users, acting through its agents, who accessed Amazon rather than Perplexity itself. It is the first federal appellate ruling on whether AI agents acting on behalf of users can legally access online platforms, a question with broad implications for agentic commerce. Amazon said it disagrees with the decision and is evaluating its next steps.
05
Ars Technica ↗
OpenAI publicly rebutted Apple's July 10 trade secrets lawsuit in a blog post, calling the allegations careless, aggressive and oddly personal, and stating that it does not have and does not want any of Apple's trade secrets. OpenAI published an email chain showing that Apple's outside counsel had contacted the wrong person in February after confusing two similar last names, along with an apology from that counsel. Apple, meanwhile, says an internal investigation identified 11 additional former employees who may have been involved, and it has moved to bar OpenAI and ex-employees Tang Tan and Chang Liu from using its confidential information while seeking expedited discovery.
06
Ars Technica ↗
Texas Governor Greg Abbott ordered state regulators to freeze all new data center connections to the power grid until a full audit of their energy and water use is complete. In a letter to the Public Utility Commission of Texas and grid operator ERCOT, Abbott directed the agencies to collect disclosures covering tax incentives, power consumption projections, on-site generation, cooling technology and community protections. ERCOT is currently reviewing 474 gigawatts of connection requests, roughly five times the state's peak demand, with data centers accounting for about 90% of the incoming load. Neither agency has said when approvals will resume.
07
Unite.AI ↗
Mistral AI released Shieldstral, a 3 billion parameter open weights safety classifier that screens both text and images against moderation policies written in plain language at inference time, rather than against fixed harm categories baked in during training. The model is available on Hugging Face under the Apache 2.0 license, covers 12 languages and runs on a single 16GB GPU. Mistral says it matches open guard models up to seven times its size on text safety benchmarks.
08
Bloomberg Technology ↗
Samsung Electronics unveiled its most advanced memory roadmap, previewing a design that vertically stacks high bandwidth memory directly on top of AI accelerators to improve performance and power efficiency. The company also detailed a new NAND product built on its V-NAND technology and unveiled what it calls the industry's first V10 BV-NAND architecture. The push is aimed at overtaking SK Hynix and Micron in the race for long term dominance in AI memory.
09
Unite.AI ↗
NVIDIA released Alpamayo 2 Super, a 34 billion parameter reasoning model for autonomous driving, for commercial use. The weights are posted on Hugging Face under OpenMDW-1.1, the Linux Foundation's permissive license for AI model distributions, which permits fine tuning, derivative models and commercial redistribution. The move clears a licensing barrier that had confined earlier versions of the Alpamayo family to research, opening robotaxi development to commercial teams.
10
TechCrunch AI ↗
Spotify said Merlin, which represents more than 30,000 independent labels and distributors, has joined Universal Music Group in backing its upcoming AI powered remix and covers product. The paid tool will let fans generate AI covers and remixes of participating artists' music, with artists opting in, receiving credit and being compensated. The partnership gives Spotify licensed coverage across a large share of the independent music catalog as it moves generative music features toward launch.