OpenAI is turning away new $200 Pro subscribers as Astra demand soars

Share
OpenAI is turning away new $200 Pro subscribers as Astra demand soars

OpenAI hit the one problem every other subscription business dreams of — too many people willing to pay — and responded by telling new customers to wait. Two days after the most hype-heavy launch window in the company's history, the $200 Pro tier is closed to new sign-ups, and Amazon just made its answer to workplace agents generally available on the desktop.

OpenAI paused new subscriptions to its $200-per-month ChatGPT Pro plan, saying demand for GPT-6 Astra has hit levels the team describes as "unprecedented." Tibo (@thsottiaux), an engineer on the Codex and ChatGPT team, announced the freeze on X: "To make sure our current users have an incredible experience and continued access to Astra, we are going to pause subscriptions to our $200 Pro plan. These put the most strain on our systems." The mechanics matter: existing Pro subscribers keep full access, every cheaper ChatGPT tier is unaffected, and the developer API remains open to new customers. Only the heaviest consumer tier — the one whose users run the most agentic and coding workloads — is closed, and OpenAI has not said when it reopens. A day earlier Tibo had flagged that Astra's growth spike was unlike anything OpenAI had seen, even against its biggest launches.

Read it as a supply-constraint story wearing a growth-story costume. OpenAI is a company rumored to be spending billions on compute, and it still cannot serve everyone who wants to hand it $200 a month — the clearest signal yet that Astra's inference cost per heavy user is extraordinary, and that capacity, not demand, is the binding constraint on consumer revenue right now. We covered the launch itself when OpenAI shipped GPT-6 Astra and said the AGI era had started; a week later, the AGI era comes with a waitlist. There is also a quieter read: capping the $200 tier while leaving API sales open suggests enterprise and developer dollars — higher margin, more predictable — get the GPUs first.


Amazon's agentic workplace platform Quick is now generally available on Windows and macOS desktops. The GA follows April's preview revamp, and it ships with a refreshed mobile activity feed for iOS and Android that consolidates email, calendar, messaging, and CRM into a single priority view — AI agents triage the routine items in the background so only the high-value ones reach a human. Quick runs on AWS infrastructure, which is the pitch: agents keep working when the laptop closes, data stays inside enterprise firewalls, and every agent conversation is auditable through CloudWatch and CloudTrail.

The interesting part is what Quick is really for. Amazon spent a year blocking third-party AI agents from its storefront, and its own ad business is now entangled with ChatGPT — meanwhile it is selling enterprises an agent supervisor that sits across their SaaS sprawl and quietly decides what a human needs to see. Amazon does not have a frontier consumer model to sell, so it is selling the layer around the models: triage, continuity across devices, audit trails. That is the same wager every cloud incumbent is making — that the money in agentic AI lands with whoever owns the security and workflow surface, not the model itself.

What to watch: whether the Pro pause lasts days or stretches into weeks — a long freeze would hand Anthropic and Google a rare window to poach power users with capacity they can actually deliver.

If OpenAI can't sell subscriptions to people waving $200 at it, what does that say about the compute behind the AGI talk? 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