compute

Waymo for the first time disclosed details about the high-performance compute 'brain' in its robotaxis, while commentators highlight that AI driving systems like Waymo's and Tesla's chips are driving a rapidly growing demand for memory and on-board compute.

The author argues that SpaceX's core challenge is balancing token production between its own Grok Bot and serving other models like Claude at spot pricing, to fund massive data center and compute investment, while keeping pricing accessible and driving model cost reductions.

Anthropic's S-1 reveals explosive 30× revenue growth over five quarters to $16.3B in H1 2026, its first profitable quarter with $974M operating profit, and a massive $637B bet on compute contracts.

A tweet joking about the confusing possibility that Anthropic rents compute from SpaceX, highlighting the tangled web of AI infrastructure deals.

Claim of spending $10M+ of compute on Navier-Stokes research and $100M+ on other scientific discoveries, with much of it unpublished yet. The post links to an accompanying YouTube video.

The Verge analyzes Anthropic's S-1 filing, which reveals 14x revenue growth to $16.3B and $1.62B adjusted income, while arguing the company's business plan essentially requires massive compute spending and dramatic recursive growth to stay ahead of OpenAI, Google, and open-source rivals.

A discussion post exploring the emerging trend of giving AI agents their own wallets to pay for APIs, data, and compute, weighing the convenience against risks like costly mistakes and asking readers about spending limits and safety requirements.

Stripe is collecting and publishing empirical data on European startups and computing power, with Patrick Collison saying he hopes this will provide valuable insights for Europe's entrepreneurship and industry.

Anthropic's S-1 IPO prospectus omits key financial figures, including third-quarter revenue and compute lease terms, raising questions about disclosure practices and AI investment uncertainty ahead of its public debut.

Anthropic's S-1 IPO filing reveals unusual details, including executive pay and security costs for the Amodei siblings, existential risk factors around catastrophic AI harm and sentience, massive compute commitments with SpaceX, Google, Microsoft, and Amazon, and a special corporate structure giving cofounders lasting control.

The neocloud boom is essentially driven by Anthropic: of the $157 billion in orders at publicly listed prices signed this year, Anthropic accounts for $90 billion (57%), mitigating the risk of compute shortages through a multi-chip strategy spanning Trainium, TPU, NVIDIA, and AMD.

Google CEO Sundar Pichai discusses the company's commitment to frontier AI with its Gemini 4 Argon model, explaining how Google forgoes short-term cloud revenues to allocate compute toward cutting-edge research, and touches on competition with OpenAI, partnerships with Anthropic and Apple, and future AI products.

Elon Musk said SpaceX is being pursued by every hyperscaler and dozens of smaller firms to rent its compute, while the company deploys 220K Nvidia GB300 GPUs per month with acceleration expected next year.

Elon Musk responds that "lack of electricity" oversimplifies the challenge, arguing that deploying massive AI compute at high speed is difficult across the board, echoing a former Google employee's analysis that data center capacity and power, not GPU production, are the real bottlenecks.

Quote tweet highlighting a new Institute for Progress (IFP) report examining US-China AI competition, analyzing how Chinese AI companies' access to compute and distillation of American models affect the current capabilities gap.

Neil Chowdhury argues that if AI labs can spend $10M+ of compute on solving math problems, they should allocate similar resources to closed-ended mechanistic interpretability problems, noting that negative results would also be informative. The post quotes Jacob Hilton's quip about interpreting his 432-parameter model.

Suhail predicts that agent swarms are coming and that a tremendous amount of real-world problems will be reframed as compute problems.

Anjney Midha corrects AMPPublic's assets under management to $440B and thanks Zurich Insurance as an anchor investor supporting the 2GW National Compute Grid Trust program aimed at scaling frontier AI for public benefit.

Tibo Sottiaux 分享了为显著减少 "Selected model is at capacity" 提示所做的改动，并通过增加算力让服务在容量不足时能够自动继续而不是阻塞。

NVIDIA joins the European Commission's AI in Science Enablers Initiative at the AI in Science Summit in Dublin, pledging open models, datasets, compute, training resources, and collaborations to support Europe's scientific research community.