OpenAI has temporarily stopped accepting new sign-ups for its $200-per-month ChatGPT Pro plan, citing the strain that demand for its newest frontier model, GPT-6 Astra, is placing on the company's infrastructure. The pause, first reported by TechCrunch, was confirmed publicly by OpenAI product lead Thibault Sottiaux and remains in effect as of September 11, according to Fortune.

The move is one of the clearest signs yet that the rollout of Astra, which OpenAI launched on September 3 and has billed as the beginning of the "AGI era," is testing the limits of the compute capacity behind one of the world's most heavily used AI products. For readers tracking the latest AI developments, it is also a rare case of a major AI company throttling its most expensive subscription tier rather than degrading service for existing users.

What OpenAI Confirmed

Sottiaux, who leads core products including Codex and ChatGPT at OpenAI, announced the pause in a post on X. He explained that the Pro plan puts more load on the company's systems than any other tier, which is why new subscriptions for that plan specifically are being disabled while the company works through the backlog of demand.

"We wanted to take the smallest step that allows us to continue giving the broadest access possible," Sottiaux wrote.

The pause did not come as a complete surprise. A day earlier, Sottiaux had warned that the company might have to act. "Demand for Astra is really unprecedented," he wrote on Wednesday. "We're pulling all the levers possible to sustain the demand, but I've not seen anything like it until now and we went through very steep growth before. Priority will always be to keep excellent service for existing users, but we might have to pause new Pro subscriptions for a bit if this continues."

Why Astra Demand Is Different

GPT-6 Astra went live on September 3 and has been rolling out across OpenAI's plan tiers, including Pro, Plus, Enterprise, and Business accounts. OpenAI has positioned the model as a generational leap in reasoning, coding, and computer use — claims that have fueled intense interest among developers, researchers, and enterprise buyers.

The company's own marketing has amplified the demand curve. OpenAI publicly described Astra's release as the start of the "AGI era," language that drew enormous attention across the tech industry and, evidently, a flood of users hoping to test the model's limits for themselves.

OpenAI raised usage limits for Codex users as recently as last month, a decision that in hindsight suggests the company was already managing significant load — and that the current crunch is a recent and steep deterioration.

What the Pause Does — and Does Not — Affect

The restrictions apply only to new sign-ups for the $200 Pro tier. According to Sottiaux, the company's other access routes remain fully available:

  • The API, which developers use to build on OpenAI's models
  • The lower-cost Go and Plus consumer plans
  • Existing Pro subscriptions, which continue to be served

OpenAI has not said how long the pause will last, and it has not disclosed how many people are attempting to subscribe each day — figures that would help quantify just how unusual the current demand spike is. Existing subscribers, meanwhile, keep their access, which is consistent with Sottiaux's stated priority of "excellent service for existing users."

The Capacity Squeeze Behind the Scenes

Running a frontier model at this scale is expensive in a way that consumer subscriptions rarely capture. Pro users reportedly consume disproportionate amounts of compute relative to their subscription price, and a model as capable — and as heavily used — as Astra magnifies that imbalance. Pausing new Pro sign-ups caps the most compute-intensive category of demand while keeping revenue flowing from every other tier.

The episode also lands at a delicate moment for OpenAI, which is simultaneously contending with public debate over the pace of frontier development. CEO Sam Altman said this week that he is open to slowing frontier AI development as safety concerns mount among the company's own researchers — a separate but thematically related acknowledgment that the race to ship ever-more-capable models is straining more than just data centers.

What It Means for Would-Be Subscribers

For users considering an upgrade, the practical reality is a waiting game. OpenAI has committed to reopening Pro sign-ups but has given no date. Historically, capacity pauses at AI companies have lasted from days to weeks, depending on how quickly new compute comes online.

The broader lesson for the industry is uncomfortable for the "scale at all costs" narrative: even the best-resourced AI labs can now be outpaced by their own product launches. Astra's demand spike is, in one sense, a commercial triumph — but it is also a reminder that inference capacity, not just model quality, has become a competitive battleground in the AI race of 2026.