Full Breakdown
OpenAI Halts New $200 ChatGPT Pro Subscriptions Amid Astra Demand Surge
By Drooid · · How we work
Core Event: Pause of $200 ChatGPT Pro Tier
On September 10, 2026, OpenAI announced a temporary suspension of new sign-ups and upgrades for the $200-per-month ChatGPT Pro plan (referred to as “Pro 20X”). The company said the pause is necessary because demand for its newest model, Astra, is straining its compute infrastructure. Existing Pro subscribers are unaffected, and all other plans—including Plus, Go, the API, and enterprise tiers—remain available worldwide. OpenAI has not set a date for reopening the paused tier.
Background & Context
Astra, marketed as a GPT-6 model, launched on September 3, 2026. OpenAI promoted the model as a major leap in reasoning, code generation, and “computer-use” capability, allowing the AI to interact with a desktop—filling out forms and navigating web pages at “superhuman speed.” The model also consumes usage allowances faster than the previous flagship, GPT 5.6 Sol.
The pause follows a similar capacity-driven halt in November 2023, when OpenAI temporarily stopped new ChatGPT Plus subscriptions after a post-devday usage surge.
European customers face an additional limitation: Astra is deployed only in U.S. Azure regions, with no dedicated EU data zone, meaning processing can occur outside the EU data boundary.
To meet growing demand, OpenAI has expanded its compute capacity from 0.2 GW in 2023 to about 1.9 GW in 2025, a 9.5-fold increase. The company has diversified hardware sources beyond Microsoft Azure, adding CoreWeave and launching the Stargate $500 billion, four-year AI infrastructure program. New capacity slated for the second half of 2026 includes at least 10 GW of Nvidia systems, a 6 GW AMD GPU agreement, and a 10 GW Broadcom custom-accelerator partnership.
Data & Statistics
- Astra’s “computer-use” feature is the primary driver of higher compute consumption.
- The Pro 20X tier provides 20 times the usage limits of the Plus tier and offers 200 credits per week, compared with 50 credits per week for the Pro 5X tier.
- OpenAI’s compute capacity grew from 0.2 GW (2023) to 1.9 GW (2025), a 9.5-fold increase.
- Astra’s usage allowance depletes approximately 2.5 times faster than GPT 5.6 Sol, according to internal metrics cited by OpenAI.
Official Statements & Responses
OpenAI declined to comment further for this report.
Verbatim Quotes
- “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,” — Thibault Sottiaux
- “The surge in usage post devday has exceeded our capacity and we want to make sure everyone has a great experience,” — OpenAI CEO Sam Altman
Why It Matters
The suspension highlights the scaling challenges of delivering compute-intensive AI models to a rapidly expanding user base. Heavy-usage developers and enterprises that rely on the Pro 20X tier may need to shift to lower-cost plans or the API, potentially affecting productivity and project timelines. The lack of an EU-specific data zone also raises regulatory considerations for European organizations. OpenAI’s ongoing infrastructure investments suggest that future capacity expansions could alleviate the bottleneck, but the episode underscores how quickly demand can outpace supply in the frontier AI market.