Apple vs OpenAI: Judge Signals He'll Block Trade Secret Use

A federal judge in San Jose just told OpenAI, in open court, that he's "highly likely" to order it to stop using anything it learned from Apple's confidential files. That's about as close to losing as a company can get before a judge actually signs anything.

What the fight is actually about

Apple sued OpenAI back in July, along with OpenAI's hardware chief and a former Apple engineer, accusing them of pulling confidential details out of ex-Apple employees — things like unreleased product designs, manufacturing processes, component choices, and how Apple tests its hardware before launch. The case, filed in the US District Court for the Northern District of California, is really about one thing: both companies are racing to build the same kind of product, a screenless, always-listening AI hardware device meant to sit somewhere between a phone and a digital assistant, and Apple thinks OpenAI got a head start by poaching knowledge along with the people.

At a hearing this week, Judge Edward Davila said he leans toward granting Apple's request for a preliminary injunction — a court order that kicks in before the full case is even decided, usually because one side can show it will suffer real harm if it has to wait months for a trial. In plain terms: Apple doesn't have to win the lawsuit yet to get OpenAI legally boxed in right now.

OpenAI's counterpunch didn't quite land

Just before that hearing, OpenAI tried to flip the script. It claimed a former OpenAI employee who later rejoined Apple had sent himself more than a hundred confidential files about an unreleased OpenAI device, using a personal account. Apple called the filing a distraction and said it found nothing to back the claim. Notably, OpenAI's filing didn't actually turn into a counter-lawsuit — it was framed more as "look over there" than a real legal move, and the judge didn't seem to buy it as a reason to go easy on OpenAI.

The judge said he was "highly likely" to side with Apple on the injunction, though he pushed back on one part of Apple's ask — calling a proposed full forensic inspection of OpenAI's devices and internal systems a "huge ask."

Where things stand now:

• OpenAI has agreed not to use any Apple trade secrets it may have picked up, while the case is still ongoing.
• OpenAI will hand over forensic images from the devices of 14 employees who may hold Apple's confidential material.
• The court partially granted Apple's request for faster-than-usual evidence sharing, giving OpenAI roughly two weeks to produce records.
• OpenAI's separate request to throw the whole case out hasn't been decided yet.

Why this should matter to Indian tech teams too

It's tempting to read this as a Silicon Valley spat that stays in Silicon Valley, but it lands differently if you work at — or hire for — one of the dozens of global capability centres (GCCs) that Apple, Google, Microsoft, and other multinationals now run out of Bengaluru, Hyderabad, and Pune. These centres handle real product engineering, not just support work, which means Indian employees routinely sit on the same kind of confidential design and manufacturing data at the centre of this case.

Here's the uncomfortable part for Indian employers: the legal tools the US is using against OpenAI barely exist here. India has no standalone trade secrets law — protection depends on what's written into an employment contract and general principles of confidentiality, and non-compete clauses that try to stop someone from joining a rival after they leave are largely unenforceable under Section 27 of the Indian Contract Act, 1872, once the person is no longer employed. The Department for Promotion of Industry and Internal Trade (DPIIT) has floated a dedicated trade secrets bill for years, but nothing has become law yet. For a GCC employee moving between competing companies' India offices — a fairly normal career move in this market — the guardrails that Apple is leaning on in a US courtroom simply aren't there in the same form back home.

What happens next

No injunction has actually been signed yet, and OpenAI's motion to dismiss the case entirely is still pending, so this isn't over. But the direction is clear enough: courts are starting to treat the AI hardware race the same way they'd treat any other high-stakes industrial espionage dispute, and companies on both sides are going to get a lot more aggressive about locking down who gets to know what before they leave the building. If you manage engineering teams anywhere in the world right now, this is as good a nudge as any to check whether your own confidentiality agreements would actually hold up if tested — because increasingly, they are being tested.