Apple Sues OpenAI Over Hardware Poaching as Google Pitches a Referee

Apple Discovers Its Partner Was Also Its Competitor

Apple filed suit against OpenAI, alleging a systematic campaign to poach Apple hardware engineers and extract trade secrets tied to unreleased devices. The complaint, covered in parallel by the BBC, Financial Times, Guardian, Verge, New York Times, and Wired, describes recruited employees walking out with confidential presentations, prototype details, and supplier information, allegedly with encouragement from their new employer.

This is the same OpenAI that signed a 2024 services deal to power Apple Intelligence. The partnership is now a smoking crater. Two companies that spent last year on stage together are now trading discovery requests, and the pretense that Sam Altman's device ambitions and Apple's roadmap could coexist inside the same NDA is gone.

The substance of the complaint is not novel. Trade secret suits against departing engineers are a Silicon Valley genre. What makes this one interesting is the alleged pattern. Apple claims this was not opportunistic hiring but a coordinated recruitment strategy aimed at specific hardware talent, timed to OpenAI's ambitions to ship a device with Jony Ive. If Apple's version holds up, the Ive-Altman gadget is partly a Cupertino product built on the wrong side of the payroll.

The Partnership Playbook Is Cooked

The broader signal here matters more than the docket. For two years, the frontier AI economy has run on a fiction that hyperscalers, model labs, and device makers could partner without eating each other. Microsoft and OpenAI already demonstrated how that fiction ends. Apple and OpenAI just filed the sequel.

Every lab that wants to ship hardware needs supply chain expertise it does not have. Every device company that wants competitive AI needs models it cannot train alone. The overlap guarantees that today's integration deal is tomorrow's deposition. Expect Anthropic, Google, and Meta to read the complaint carefully and rewrite their partner agreements this quarter. The era of loose talent flow between allied AI companies is done. Non-competes and forensic laptop imaging are back in fashion.

For OpenAI, the timing is unhelpful. The company is trying to convince regulators, enterprise customers, and its own investors that it is a mature institution. A well-documented trade secrets complaint from the most respected hardware company on earth cuts against that story. It also gives every rival a talking point.

Google Wants to Draw the Lines Itself

Against that backdrop, Google floated a proposal for a private AI regulator, an independent body that would set frontier AI rules with government oversight and judicial review. The pitch, surfaced by Andy Jung, is the kind of idea that sounds reasonable at a policy dinner and gets uglier in daylight.

The appeal is obvious. Legislators cannot keep up. Agencies are understaffed. A specialized body funded by industry could move faster, hire better, and write rules that actually track the technology. Financial industry self-regulatory organizations like FINRA are the standard analogy.

The problems are also obvious. Delegating rulemaking to a private entity runs into nondelegation doctrine, which the current Supreme Court has shown appetite to enforce. Adjudication by a non-Article III body raises due process concerns that have already tanked SEC in-house courts. And a regulator staffed and funded by the labs it oversees is a capture problem dressed as a governance solution. FINRA works, sort of, because securities law is mature and the SEC sits behind it with real teeth. Frontier AI has neither the settled doctrine nor the confident backstop agency.

The honest reading is that Google would rather negotiate with a body it helped design than wait for Congress or the FTC to draw lines it cannot influence. That is rational corporate behavior. It is also exactly why the Constitution does not let private parties write the rules that bind their competitors.

The Through Line

Today's slot is about who controls the machinery. Apple wants control of its hardware secrets. OpenAI wants control of a device category. Google wants control of the rulebook. None of them are wrong to want these things. All of them are asking courts, regulators, and each other to sort it out, because the industry has run out of ways to pretend cooperation scales.