OpenAI Uses $1.2M in API Calls to Find Kernel Flaws

OpenAI says an automated pipeline generated working exploits for 948 Linux kernel vulnerabilities at an API cost of roughly $1.2 million.

Updated on Nov. 18, 2026 in Artificial Intelligence

OpenAI Uses $1.2M in API Calls to Find Kernel Flaws

Live Poll

Do you trust AI companies to handle the technology responsibly as they grow more powerful?

OpenAI spent approximately $1.2 million in API costs to generate working exploits for 948 Linux kernel vulnerabilities, according to a technical report it published on November 4. Human security researchers reviewed the findings before they were sent to Linux kernel maintainers.

Why it matters

The report shows how automated AI systems can search for, reproduce and exploit software bugs at modest cost, a capability that bears on how kernel maintainers and security teams prioritize patching. Sam Altman said at a New York Times DealBook Summit that OpenAI is set up to be vertically integrated across the AI stack.

The automated pipeline searched for vulnerabilities, reproduced bugs, tested exploitability and wrote patches, with about $1.2 million in API spending as an upper-bound cost estimate. Figures from OpenAI's own report have not been independently verified, and the optimized cost of the approach remains unknown.

The players

OpenAI

OpenAI is an artificial intelligence research and deployment company known for developing the GPT family of large language models and the ChatGPT product.

Sam Altman

Sam Altman is the chief executive of OpenAI and a prominent public voice on artificial intelligence policy and infrastructure.

The details

The pipeline reproduced 1,724 vulnerabilities and produced working exploits for more than half of them, with 948 identified as Linux kernel vulnerabilities. OpenAI published the work in a technical report titled "A thousand fewer paths to root." Human security researchers reviewed the findings before they were sent to Linux kernel maintainers.

Timeline

  1. - November 4, 2026: OpenAI publishes its technical report on the Linux vulnerability pipeline.

  2. - Wednesday, November 18, 2026: Sam Altman speaks at The New York Times' DealBook Summit alongside OpenAI President Greg Brockman.

The Tech Race

OpenAI is developing a custom Jalapeño inference chip, which fits its stated plan to be vertically integrated across the AI stack. The automated vulnerability work and the long-running agents it projects both depend on compute, which Altman has called the biggest constraint on what AI systems can accomplish.

Users of Linux-based systems may see security patches arrive as maintainers process the reported kernel vulnerabilities. The findings also suggest that automated exploit generation is becoming cheap enough that patch timelines may matter more to everyday users.

The takeaway

Keeping systems updated matters more when automated tools can find and exploit bugs at low cost. Install kernel and distribution security updates promptly once they are released.

Further reading

For more on the tools reshaping software security, explore Artificial Intelligence coverage on Wisevoter.

Source note: This article includes information reported by The New Stack.

Live Poll

Do you trust AI companies to handle the technology responsibly as they grow more powerful?