The biggest AI company in the world just got breached. The way in wasn't a phishing email or a flaw in OpenAI's own systems.
It was a piece of open-source code that two of its engineers had installed. That should worry every tech company.
TanStack is a coding library that engineers use to build apps and websites. It's free, open-source, and trusted enough to live inside thousands of products.
On Monday, attackers pushed out 84 poisoned versions of TanStack packages in a six-minute window. A security researcher detected the attack within 20 minutes.
Even with that quick catch, the bad code landed on the laptops of two OpenAI engineers. The malware was built to grab login credentials and spread to other systems on the same network.
OpenAI confirmed Wednesday that those two devices were hit.
We break down what cyber attacks like this mean for investors every morning in Market Briefs - five minutes a day, with a free investing masterclass thrown in when you join.
OpenAI ran an investigation and said the damage was contained. No user data was accessed, production systems and intellectual property were not compromised, and its software was not altered.
The attackers did access internal source code repos those two employees worked in, which OpenAI called a "limited subset." From those repos, they grabbed some credentials.
One detail to flag: those repos held digital certificates OpenAI uses to sign its products. The company is now rotating those certificates as a safety step.
Mac users will see an app update prompt because of it.
This is the third major supply chain attack on a widely used open-source project in three months. In March, attackers tied to North Korea poisoned Axios, another widely used development tool.
Earlier this month, Chinese-linked hackers were accused of doing the same to Daemon Tools, a disc-imaging program that runs on a lot of Windows machines.
A hacking group called TeamPCP has been linked to past attacks of this type. The TanStack attacker has not been named yet.
For investors, the takeaway is that software supply chain risk is now a top-tier business risk for every tech company. If hackers can reach inside OpenAI through an open-source coding tool, they can reach inside almost anyone.
Cybersecurity stocks like CrowdStrike, Palo Alto Networks, and Cloudflare all sell products built to catch exactly this kind of attack. Demand for that work tends to spike after a name like OpenAI shows up on the victim list.
OpenAI is the biggest name on the list of recent victims. It will not be the last.
To get a daily read on stories like this and how they move markets, sign up for Market Briefs - you also get a 45-minute investing course as a bonus when you sign up.
