Trade Secrets After Apple v. OpenAI: Answers to the Questions Founders Are Actually Asking

By Babak Akhlaghi on July 25, 2026.  Apple’s lawsuit against OpenAI put trade secrets on the front page. If you run a startup, hire from competitors, or hold confidential know-how of your own, this case is a playbook worth studying.

This FAQ answers the questions we hear most, using the Apple v. OpenAI complaint as a concrete reference point throughout.

What exactly is a trade secret, and how is it different from a patent?

A patent protects an invention through public disclosure. You tell the world what you invented in exchange for a time-limited exclusive right.

A trade secret protects confidential information through secrecy. No filing, no disclosure, no expiration, as long as you keep it confidential and take reasonable steps to protect it. The tradeoff is simple: once the secret is out, the protection is gone.

Three elements must all be present:

• The information is confidential and not publicly known
• It derives commercial value from that confidentiality
• You have taken reasonable measures to protect it

Customer lists, internal workflows, supplier relationships, manufacturing specifications, pricing models. None of these qualify for a patent. All of them can be trade secrets. That makes trade secrets an essential part of your IP roadmap, alongside any patents you file.

What is the Apple v. OpenAI lawsuit actually about?

Apple filed suit on July 10, 2026, in the U.S. District Court for the Northern District of California. The complaint alleges that OpenAI orchestrated a coordinated effort to misappropriate Apple’s trade secrets, specifically hardware designs, manufacturing processes, and supplier relationships, to accelerate its push into consumer hardware.

Two former Apple employees are named. Chang Liu, a senior electrical engineer, allegedly retained a company laptop and used an authentication bug to download dozens of confidential files from Apple’s network. Tang Yew Tan, a former VP of product design with 24 years at Apple, allegedly emailed himself supplier information before leaving.

Apple also alleges that OpenAI encouraged job candidates to bring confidential Apple materials to interviews for “show and tell” sessions, and that OpenAI employees contacted Apple suppliers seeking to use Apple’s proprietary manufacturing processes. Apple wrote to OpenAI in February 2026 raising these concerns and received no response.

More than 400 former Apple employees now work at OpenAI.

What makes this case different from other AI trade secret disputes?

Most prior AI trade secret cases involved software: algorithms, training data, model architectures. This case centers on hardware. Physical designs, manufacturing tolerances, supplier know-how, supply chain relationships.

Hardware trade secrets are in some ways more durable and more valuable than software secrets. They represent years of physical iteration and supplier relationship-building that competitors cannot replicate quickly.

The case also involves allegations of directed misappropriation. Apple claims OpenAI actively encouraged and orchestrated the acquisition of confidential information, going well beyond passively benefiting from what employees carried in their heads. Courts treat those two scenarios very differently, and the legal exposure for directed conduct is far greater.

Where is the line between general knowledge and misappropriation?

This is one of the most important distinctions in trade secret law.

You are free to take your skills, experience, and general knowledge to a new employer. A hardware engineer who worked at Apple can go do hardware engineering at OpenAI. That mobility is legal and protected, particularly in California, where state law is specifically designed to allow employees to move to competitors.

Here is the test: general knowledge lives in the mind. It does not exist in retained laptops, downloaded files, or confidential documents that were never returned.

When someone keeps a company laptop, downloads confidential files, or emails themselves supplier details on the way out, that is misappropriation. Even after 24 years at a company, the line holds. If a document is labeled confidential, the information in it belongs to the company.

Can misappropriating a trade secret be a criminal offense?

Yes. Trade secret theft can trigger criminal prosecution alongside any civil case. Under federal law, individual offenders face up to 10 years in prison and substantial fines, and organizations can face fines of up to $5 million upon conviction.

💡 This point is commonly overlooked. Taking files from behind a computer screen feels nothing like walking out of a store with merchandise. The legal consequences are just as serious, and the value of what is taken is often far greater.

Reminding departing employees of this fact during exit interviews is one of the most effective deterrents available. It makes people think twice before doing something they cannot undo.

What should a startup do right now to protect its own trade secrets?

Six steps, in order of priority:

1. Conduct a trade secret audit. Identify the confidential information you hold that competitors would value.
2. Label it. Mark confidential information as confidential, consistently and visibly. Every time an employee opens that file, the label reminds them the information belongs to the company.
3. Limit access. Share sensitive information only with people who need it, even inside the company, even under NDA.
4. Train employees. At onboarding and continuously. One training session is never enough. Policy without training accomplishes nothing.
5. Run rigorous exit interviews. Collect equipment. Revoke access. Remind departing employees of their confidentiality obligations and the criminal exposure for violating them. Apple’s own exit process failed here, and a senior VP allegedly walked out with confidential files.
6. Document your development process. Contemporaneous records showing independent creation are your strongest defense if you ever face a misappropriation claim.

If a court asks what measures you took to protect your information, you want a real answer ready.

How do we avoid being accused of taking someone else’s trade secrets?

The risk runs in both directions. If you are hiring aggressively from competitors, which describes most AI startups today, you need deliberate protocols on the inbound side.

• Separate onboarding for competitor hires. Put it in writing: no confidential information from the former employer enters your workspace. Have them sign it.
• Train your managers. The most dangerous evidence in a trade secret case often comes from a manager who casually asked a new hire about secrets from their old job in a Slack message. Every internal communication may surface in discovery. Write every message as if a jury will read it.
• Build a speak-up culture. Inappropriate questions about a former employer’s confidential information go to HR immediately.
• Use a single authorized channel for suppliers. One contact, one source of instruction. When a key employee departs, update your suppliers the same day so requests from that person stop being processed.

⚠️ Warning: Apple alleges that OpenAI circulated internal Apple documents coaching departing employees on evading Apple’s exit security. If proven, that kind of coordination transforms an employee problem into company-wide liability.

What should we watch as the case proceeds?

Four things.

First, whether Apple can establish that OpenAI had actual knowledge that confidential information was coming in. Knowledge and direction determine the depth of OpenAI’s exposure.

Second, how the court values hardware trade secrets. Putting a dollar figure on secrets embedded in physical manufacturing processes and supplier relationships is legally and technically complex.

Third, whether the case produces injunctive relief that affects OpenAI’s hardware development timeline.

Fourth, what discovery surfaces. The emails, Slack messages, and meeting notes from both sides will tell the real story of what was known and when.

Is this case a one-off, or a sign of things to come?

It is a sign of things to come.

The AI sector’s appetite for talent from a small number of dominant players, combined with the billions flowing into hardware development, creates the structural conditions for exactly this kind of dispute. As AI companies move into physical devices, the value of manufacturing know-how, supply chain relationships, and proprietary processes will keep rising.

Here is the bigger shift. AI has made products dramatically cheaper and faster to build. The old Silicon Valley belief that your team and your head start protect you no longer holds. Your IP remains, and some of your most valuable IP is your trade secrets.

Treat them that way. Put them in the lockbox, limit who sees them, train your people continuously, and build your protocols before you need them. Startups that build serious trade secret programs now will be better positioned on both sides of the next dispute. If you are still figuring out how patents and trade secrets fit together for your company, working with a startup patent attorney early is the right move.