The Arena

AWS orders 2 million more Nvidia GPUs for 2027–2028

Editorial

Editorial

— 2 min read
Share
AWS orders 2 million more Nvidia GPUs for 2027–2028

Amazon and Nvidia have expanded their decade-long partnership with a deal that adds 2 million more Nvidia GPUs to Amazon's data centers, deploying through 2027 and 2028. It is a fresh, concrete signal that top-tier AI infrastructure demand still has a long runway.

AWS is betting 2 million more Nvidia GPUs that the AI buildout hasn't peaked. The expanded partnership brings Nvidia's Blackwell Ultra, Rubin, and Rubin Ultra chips into Amazon Web Services' data centers across 2027 and 2028, according to the two companies. Neither disclosed financial terms, but with GPU unit costs where they are, TechCrunch put the order in the tens of billions of dollars. It lands hours after Nvidia posted a $96.2 billion second quarter — we covered that record quarter this morning — and it makes the "surging demand" claim behind those numbers concrete: a hyperscaler committing now to next-gen silicon two to three years out.

The relationship reaches well beyond buying chips. Nvidia's networking hardware that links tens of thousands of GPUs into single systems, its Vera CPUs, its robotics platform, and its Nemotron family of open models — which will now be offered on Amazon Bedrock and SageMaker — are all being wired deeper into AWS. The deal also extends to Amazon's warehouse robots. What makes it remarkable is that this deepening comes even as Amazon keeps pouring money into its own competing Trainium and Graviton silicon: for a customer actively building in-house alternatives, Nvidia still commands the order of the day. Jensen Huang cast it as the shift from pilots to production for agentic and physical AI, and the size of the commitment suggests the buildout is far from done.

What to watch: whether the "profitability of tokens" Huang keeps citing actually materializes as this compute comes online — and whether in-house chips like Trainium start eating into the next Nvidia order's scale.

When is a massive compute commitment like this a bet on future AI demand, and when is it a hedge that the current one is real? Tell us in the comments.

Read more

Akhetonics says its all-optical CPU reaches a customer in 2026

Akhetonics says its all-optical CPU reaches a customer in 2026

Light-based computing keeps promising more than it delivers — but one Munich startup has just put a date on its bet, and the interview laying it out is doing the rounds on Hacker News this week. Akhetonics says it will deploy its first commercial machine with a major customer by the end of 2026, with several more planned for 2027. The company, founded by Michael Kissner and Leonardo Del Bino, is building a computer where data enters as light, is switched as light, and circulates through memory

The Week in AI — October 5–11, 2026

The Week in AI — October 5–11, 2026

Every big claim this week turned out to rest on fine print more interesting than the headline: revenue only the company reporting it can define, safety tests sandboxed while the product keeps the web, and a Pentagon phase-out nobody would confirm until reporters kept asking. The week's top 5 1. OpenAI's revenue was $20 billion below the numbers everyone quoted — and the gap was definitional. The Financial Times reported Thursday that OpenAI's annualized revenue runs roughly $20 billion unde

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Russia's largest tech company is learning what the AI era's infrastructure war looks like from the receiving end — three data centers in four days, and with them much of the cloud layer Russian businesses run on. A Ukrainian drone strike knocked out Yandex's data center in Vladimir early Sunday morning, the third of the company's facilities hit since October 8. The site — reported at roughly 50 MW and designed for about 2,880 server racks — stopped operating completely after the attack, Yandex

Agent teams cost up to 5x more, barely score higher

Agent teams cost up to 5x more, barely score higher

The multi-agent hype train hit a benchmark this weekend — and the grid and the trucking regulators had quiet weeks of their own. Vals AI put agent teams head-to-head with single agents on its Vibe Code Bench, and the teams cost between 1.8 and 5.1 times more for almost no extra quality. The evals company ran GPT-6 Sol and Claude Opus 5.5 solo and in teams across 50 apps at two reasoning efforts; out of four comparisons, only one was statistically significant — Sol at medium effort, where the t