The headline says OpenAI filed a motion to dismiss. OpenAI hasn't. It wrote a blog post instead.

The headline running around this week claims OpenAI asked a judge to dismiss Apple's trade-secrets lawsuit. The thing is, OpenAI hasn't filed a motion to dismiss. It published a blog post.

That is an odd way to fight a lawsuit against a company worth roughly $5 trillion - one that just saw its stock drop 8% in five days to around $311 and may be throwing the full institutional weight of its litigation machinery at you. A blog post is what you'd expect from a consumer tech startup trying to rally a Reddit audience, not from a company facing allegations that it stole proprietary supply-chain information and metal-finishing techniques to build its own hardware.

But the blog post is exactly the right move if you think about what this lawsuit actually is, and what Apple is really trying to do with it.

What Apple's case is

Apple filed its complaint on July 10 in the Northern District of California. It accuses OpenAI - and more precisely, two former Apple employees now working at OpenAI, Chang Liu and Tang Yew Tan - of a coordinated effort to extract Apple's hardware trade secrets. Tan, Apple's former vice president of product design for iPhone and Apple Watch, allegedly directed job candidates who still worked at Apple to bring "actual parts" to OpenAI interviews for "show and tell" sessions. Liu, a former electrical engineer, allegedly failed to return a company laptop, exploited an authentication bug to access Apple's shared cloud storage after his departure, and downloaded dozens of Apple's confidential hardware-related files.

The complaint is detailed enough to make you think someone was having a very bad time at an OpenAI interview. One candidate allegedly told Tan he "didn't even know we could take those [parts] from the office."

Apple also claims Tan circulated Apple's internal departure-procedure document - marked "Need to Know" - to recruits before they gave notice, coaching them on how to avoid the "dreaded walkout" (being escorted out immediately rather than working their two-week notice, which would give them more time to download things).

This is the sort of complaint that makes the legal team at the hiring company have a meeting.

The shadow case

Here is the thing Apple's complaint was written in the shadow of. Six weeks before Apple filed, on June 15, 2026, Judge Rita Lin... permanently dismissed xAI's trade-secrets lawsuit against OpenAI - without leave to amend. xAI had alleged that OpenAI recruited a former xAI engineer and induced him to share reinforcement-learning techniques for Grok 4 during an interview presentation. Judge Lin found the theory deficient. Asking a candidate about their past work doesn't create trade-secret liability. Passive receipt of information doesn't count as "acquisition" under the Defend Trade Secrets Act. You need active conduct: showing that the hiring company knew, encouraged, or benefited from the disclosure.

xAI then got billed a million dollars in attorney's fees for bringing the case.

Apple's complaint reads as though its lawyers had Judge Lin's order open on the desk beside them. Where xAI alleged passive receipt, Apple alleges active acquisition - an authentication bug, unauthorized downloads, coaching on how to dodge security. Where xAI couldn't plead that OpenAI knew what it was getting, Apple pleads documented knowledge - Tan's circulation of Apple's own internal security procedures to new hires. Where xAI alleged a routine interview, Apple alleges an interview process designed to extract trade secrets.

It's a competent, well-structured complaint. The question is whether it survives the same standard that killed xAI's case.

What OpenAI actually did

On August 3, Apple filed a motion for a preliminary injunction - asking the court to immediately bar OpenAI, Liu, and Tan from accessing, using, or disclosing Apple's alleged confidential information. Apple also filed a motion for expedited discovery, wanting depositions of the key witnesses now, not months from now.

OpenAI's response on August 4 was the blog post, titled "Apple is Getting This Wrong." It calls the lawsuit "careless, aggressive, and oddly personal". It publishes iMessage exchanges showing Apple employees asking Liu for help locating files after his departure ("Apple employees reached out to him and asked for his help to locate this information"). It publishes an email chain showing Apple's outside counsel accidentally emailed the wrong OpenAI employee in February, then apologized, which OpenAI frames as evidence that Apple never properly contacted OpenAI's general counsel before filing suit.

This isn't a legal filing. It's a public relations counterstrike. OpenAI is trying to shape the factual record in a venue the court can't control.

That's unusual. It's also revealing. If OpenAI were sitting on a clean motion to dismiss - one that could say, in effect, "Apple's complaint doesn't meet the pleading standard Judge Lin just set for this exact court" - there would be a motion to dismiss in the docket. The docket shows Apple's motion for a preliminary injunction, but not a motion to dismiss from OpenAI. The available reporting suggests OpenAI's formal legal response hasn't been filed yet.

So why the blog post first?

The structure of the game

The simplest model is that OpenAI is buying time and shape. A blog post costs nothing, reaches everyone, and frames the narrative before the court files become public. It turns a trade-secrets allegation - where the default instinct is to stay quiet, stay clean, and let the lawyers work - into a public fight where OpenAI gets to play the scrappy innovator and Apple gets to look like the bully.

That framing works better in a browser than in a complaint. The iMessage screenshots, even if they don't directly refute the core allegation (Apple's injunction motion says the disputed downloads came from a third-party cloud provider, likely Box, not iCloud, and the messages address iCloud access), create a plausible confusion: "Apple's own employees were talking to this guy. Is this really theft?" That's the sort of murkiness that works in public but might not survive depositions, where the timeline, the authentication logs, and the actual downloaded files would become the record.

The email misdirection about the wrong address is even funnier in its own way. Apple's initial letter was sent to OpenAI general counsel Che Chang, who did not respond. The mistaken email was a follow-up sent by Apple's outside counsel to the wrong person, and they quickly apologized. But the headline works. "Apple sued us without talking to us" is a better press line than the actual sequence of events.

Who has the harder problem

Apple has the harder structural problem in court, even if its complaint is more detailed than xAI's. The DTSA requires misappropriation - improper acquisition, disclosure, or use. For the individual defendants, Liu's alleged conduct (auth bug, repeated downloads over three months from February through April 2026) is the strongest factual basis Apple has. If the forensic evidence holds up, that's active acquisition, not passive receipt.

But liability for OpenAI as a company requires showing the company knew or should have known, and that it benefited. Apple pleads that Tan - OpenAI's chief hardware officer - distributed Apple's own security documents to recruits. If that's true, it's a direct shot at the knowledge requirement. But if Tan's defense is "I was sharing what I knew about Apple's departure process to help new hires navigate a company I worked at for 24 years," and the jury buys that the document itself isn't a trade secret (it's internal procedure, not a confidential formula), the corporate-level claim weakens.

The interview-process allegations - asking candidates to bring parts for "show and tell" - are the messiest part for both sides. They sound predatory. They might also survive as aggressive hiring rather than trade-secret misappropriation, depending on whether Apple can show discrete confidential information was actually extracted and used, rather than general industry knowledge that former employees already had.

The actual headline

The competitor headline - "OpenAI asks US judge to dismiss Apple's trade secrets case" - runs ahead of the filing. OpenAI hasn't asked a judge to dismiss anything yet. It published a blog post responding to Apple's preliminary injunction motion. The formal legal response, if it includes a motion to dismiss, would come next.

Whether it does is the real open question. The xAI dismissal created a high bar for trade-secret claims in this court. Apple's complaint was clearly drafted to clear it. The question for OpenAI's lawyers is whether Apple succeeded, or whether the complaint is still too close to the xAI pattern to survive.

Either way, the blog post is a signal about what sort of company OpenAI thinks it is right now. Not the one that quietly lets its legal team work. The one that wants the public to read the receipts.

That's a fine strategy for a company preparing for a public offering. It's a less predictable strategy for a company in federal litigation against Apple.