After GPT-5 launches there will be another big spike in demand that OpenAI will not be able to serve.

Note: GPT-5 launched on 2025-08-07, before the claim date. Judge the claim over the weeks after the statement (through about end of September 2025). True: dated reporting or OpenAI/Altman statements show a big spike in demand after GPT-5's launch AND that OpenAI could not fully serve it, e.g. rate limits imposed or tightened, capacity shortages, degraded or unavailable service, delayed rollouts or waitlists caused by compute limits. Mostly true: both happened but the shortfall was brief or minor (e.g. short slowdowns or temporary rate-limit cuts only). Mixed: a clear demand spike that OpenAI largely served, with only scattered capacity problems, or capacity problems without evidence of a demand spike. Mostly false: demand rose only modestly and service problems were trivial. False: evidence shows no notable demand spike, or that OpenAI served demand without capacity-driven limits or outages.

Set before the research began.

1. E1 Sam Altman, over bread rolls, explores life after GPT-5 | TechCrunch TechCrunch · 15 August 2025

   “Altman says OpenAI’s API traffic doubled within 48 hours of GPT-5’s launch, and the company is effectively “out of GPUs” thanks to all the demand.”

2. E2 ChatGPT gets Auto, Fast, and Thinking options, but its limited to paid GPT-5 users The Indian Express · 13 August 2025

   “The AI startup also said that they might change the rate limit depending on usage.”

3. E3 Users encountering rate limit or model not found errors in GPT 5 - OpenAI Status OpenAI Status Own record

   “We have identified that users are experiencing elevated errors for the impacted services.”

4. E4 Introducing GPT-5 OpenAI · 7 August 2025 Own record

   “Once usage limits are reached, a mini version of each model handles remaining queries.”

TechCrunch (E1) reports Altman saying that OpenAI's API traffic doubled within 48 hours of GPT-5's launch and that the company was effectively "out of GPUs" because of the demand. The resolution criteria count OpenAI or Altman statements as evidence of both a spike and a capacity shortfall. Indian Express (E2) reports OpenAI saying it might change rate limits depending on usage, and an undated OpenAI status incident (E3) records elevated errors and rate-limit errors for GPT-5. The evidence does not show how severe or how long the shortfall was, and the main support is Altman's own account, so the verdict is mostly true rather than fully true.