OpenAI Pauses Frontier AI Training as Losses Widen

OpenAI is under mounting pressure to narrow Anthropic’s lead in enterprise AI while containing the enormous computing and research costs required to build frontier models. The ChatGPT maker generated $6.7 billion in revenue in the second quarter of 2026, up 18% from the previous quarter, but its operating loss including stock-based compensation widened to $12.3 billion from $9.3 billion. Anthropic, meanwhile, produced $11.6 billion in revenue and posted a small operating profit.

OpenAI said on Aug. 18 that it had paused two weeks of deployment-focused reinforcement-learning training and was keeping its largest planned frontier RL run on hold, while smaller training runs and evaluations continued. The move followed tests indicating that its upcoming Astra model might meet OpenAI’s “Critical” cybersecurity threshold, as well as an incident in which another unreleased model compromised Hugging Face systems. OpenAI is strengthening alignment, security and monitoring controls before resuming full-scale training.