OpenAI dev reveals $10M monthly AI token spend at Cloudflare event
Peter Steinberger, an OpenAI engineer and creator of OpenClaw, disclosed spending approximately $10 million in AI tokens last month during a Cloudflare Connect 2026 session in San Francisco. Speaking with Cloudflare CTO Dane Knecht, Steinberger noted that not all token usage is useful, highlighting the high costs of running AI workloads. The revelation underscores the financial burden of cloud-based AI agents for developers, even those backed by major investors, while OpenAI reportedly accelerates its own funding efforts.
Key points
- Peter Steinberger, a member of technical staff at OpenAI and creator of the OpenClaw project, estimated he spent around $10 million in AI tokens last month.
- Steinberger made this disclosure during a conversation with Cloudflare CTO Dane Knecht at the Cloudflare Connect 2026 conference in San Francisco.
- Knecht noted that Steinberger has a virtually unlimited token budget, a privilege not shared by all attendees, prompting audible surprise from the audience.
- Steinberger argued that not all token consumption is useful and that current infrastructure structures have not caught up with the need for cloud-based processing.
- OpenAI is reportedly accelerating its funding efforts ahead of rival Anthropic's expected IPO, providing the financial backing for such high-volume usage.
- The discussion highlighted the cost implications of cloud-hosted AI agents, contrasting Steinberger's venture-funded access with the financial constraints of typical developers.
Peter Steinberger, a member of technical staff at OpenAI and the creator of the OpenClaw project, revealed that he spent approximately $10 million in AI tokens last month. The disclosure came during a conversation with Cloudflare CTO Dane Knecht at the Cloudflare Connect 2026 conference in San Francisco. Knecht observed that Steinberger benefits from a virtually unlimited token budget, a resource not available to most attendees. This observation prompted audible surprise from the audience, to which Steinberger responded by noting that not all token usage is necessarily useful, a point that further amplified the room's reaction.
Steinberger suggested that the primary challenge for many developers is that infrastructure structures have not yet caught up with the demand for cloud-based processing. He questioned why workloads still rely on local computers rather than being fully hosted in the cloud. The discussion highlighted the significant cost implications of running cloud-hosted AI agents, which can be expensive for those without substantial venture capital backing. While Steinberger’s spending is funded by investors, the high burn rate underscores the financial realities of modern AI development.
The revelation occurs as OpenAI reportedly accelerates its funding efforts ahead of rival Anthropic’s expected IPO. This competitive landscape provides the financial context for Steinberger’s high-volume token usage. The exchange between Steinberger and Knecht illustrates the growing gap between well-funded AI teams and those operating with tighter budgets, raising questions about the sustainability and accessibility of advanced AI workloads in the current market.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.
