In a striking demonstration of the financial demands of AI development, Peter Steinberger, the Austrian creator of OpenClaw, reported a staggering $1,305,088.81 spent on OpenAI's API over a single month. This bill resulted from processing 603 billion tokens across 7.6 million requests, primarily through about 100 Codex instances managed by a small team of three.
Steinberger, who joined OpenAI in February, shared a screenshot of his API usage dashboard, showing daily expenses of roughly $19,985.84 along with 206,000 requests on the day he posted. The project has attracted attention due to substantial financial backing from OpenAI, which covers the costs incurred by Steinberger's team.
The Implications of High Spending
This unprecedented expenditure raises questions about the economic viability of AI-assisted software development. Steinberger's use of Codex's "Fast Mode" pricing places his costs at the upper end of the spectrum. Disabling this mode could potentially lower his monthly API bill to around $300,000, still a significant amount reflecting the high costs of using advanced AI tools.
According to Steinberger, a typical Codex Pro subscription, priced at $200 per month, offers about $5,000 to $6,000 worth of API usage. By this estimate, his non-fast-mode API consumption would be equivalent to approximately 60 Codex Pro subscriptions, highlighting the gap between developer expenses and the actual computational costs incurred by providers.
The Competitive Landscape
The scrutiny surrounding AI coding tools is intensifying as competitors like Claude Code and Cursor seek to engage developers by offering subsidized inference costs. OpenAI’s recent shift to a token-based billing model has increased awareness of these expenses and introduced variability for high-volume users.
https://x.com/cantworkitout/status/2055346265869721905
Steinberger's findings reflect a broader trend within the industry, where developers are becoming more conscious of the computational costs tied to AI technologies. This awareness is especially relevant as demand for AI tools continues to grow, prompting developers to question the sustainability of existing pricing models.
A Research Perspective
Despite the eye-watering expenses, Steinberger remains undeterred, viewing the financial outlay as a form of research into the future of software development when budget constraints are lifted. He emphasized that all output from his team remains open source, indicating a commitment to community-driven innovation.
As the landscape of AI coding tools evolves, the implications of Steinberger's project will likely resonate throughout the industry. The balance between cost and capability will remain a key discussion point as developers and companies seek effective ways to integrate AI into their workflows without incurring overwhelming expenses.
With Steinberger's OpenClaw project serving as a case study, the tech community is left to consider how these costs will influence the future of AI development and the infrastructure that supports it.
Quick answers
What is OpenClaw?
OpenClaw is an open-source project developed by Peter Steinberger, utilizing AI coding tools to enhance software development.
How much did Peter Steinberger spend on OpenAI’s API?
Peter Steinberger reported spending $1,305,088.81 on OpenAI's API in a single month.
What does the spending reveal about AI development costs?
The spending highlights the significant expenses associated with AI coding tools and raises questions about the economic viability for developers.
How does Steinberger’s spending compare to typical usage?
His spending is at the extreme end of the cost spectrum, with typical costs for Codex ranging from $100 to $200 per developer per month.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.
