OpenAI resumes training by...?

The strongest signal is that OpenAI itself has already framed the pause as an open-ended safety and security hold, not a short interruption with a scheduled restart. The company said it would resume only after validating that the problem is resolved and after additional testing, which means the deciding factor is not time elapsed but completion of internal checks. Since the market question requires an official announcement that training has actually resumed, the burden of evidence is high and currently there is no reported qualifying statement.

The public reporting pattern also points toward delay rather than imminent restart. Coverage from the last couple of days indicates that OpenAI still has no exact timeline, and the language around the pause suggests the company is still in the process of diagnosing safeguards and verifying that the gap is closed. In markets like this, the absence of a firm restart date matters a lot because a resumption announcement usually follows either a technical remediation update or a clear operational decision, neither of which appears to have surfaced yet.

The main argument for Yes is that OpenAI may be able to complete internal validation very quickly and announce a fresh training run with little notice. Large labs can sometimes move from pause to restart abruptly once security approvals are done, especially if they are eager to minimize disruption. But the current market price near 5 percent seems directionally sensible because the required sequence is specific: the pause must be lifted for the covered training, and the company must make that fact clear publicly before the deadline.

Overall, the balance of evidence favors No. The company appears to be in a cautious, remediation-first posture, and there is no credible indication yet that the paused training has been officially resumed. Unless OpenAI issues an explicit announcement in the next couple of days, the most likely outcome is that the market resolves No.