Brief IA

OpenClaw: 100 AI Agents at $1.3 Million per Month

💼 Business & Startups·Tom Levy·

OpenClaw: 100 AI Agents at $1.3 Million per Month

OpenClaw: 100 AI Agents at $1.3 Million per Month
Key Takeaways
1Peter Steinberger uses 100 AI agents to develop OpenClaw, an open-source project.
2The OpenClaw team spent $1.3 million in 30 days on the OpenAI API.
3The AI agents review PRs, find bugs, and suggest automatic fixes.
💡Why it mattersThe massive use of AI by OpenClaw illustrates the potential and cost of advanced technologies in software development.
Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

📄
Full Analysis

Peter Steinberger and AI Agent Management

Peter Steinberger, the founder of the open-source project OpenClaw, has revealed the behind-the-scenes of his intensive use of artificial intelligence for software development. His team, although reduced to three people working at OpenAI, orchestrates around 100 instances of Codex in the cloud. These AI agents are versatile: they review pull requests (PR), identify security vulnerabilities in commits, deduplicate issues, and draft patches.

Some of these agents take the initiative to open PRs in line with the project's vision, while others monitor benchmarks and report regressions on Discord. They are even capable of listening to meetings and launching PRs for features discussed by the team. For bug analysis and security, the team also uses tools like Clawpatch.ai, Vercel's Deepsec, and Codex Security.

Cost of the OpenAI API

In just 30 days, the bill for the OpenAI API reached an impressive $1.3 million, corresponding to 603 billion tokens and 7.6 million requests. The most requested model was GPT-5.5, and OpenAI covers these costs. According to Steinberger, the perception of this cost as cheap or expensive depends on the perspective.

Steinberger justified this expenditure by explaining that it is an exploration of how software could be designed if token costs were not a constraint. He noted that disabling "Fast Mode" would reduce costs by 70%. Regarding return on investment (ROI), he emphasized that everything his team develops is open source and utilizes cutting-edge models as well as open models, stating that the ROI is "quite high."

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.