August 18, 2026

Apple v. OpenAI: How the World’s Most Valuable Company Is Using Trade Secret Law to Defend Its Competitive Edge

Subscribe to Our Newsletter

Newsletter


Patrick Dempsey

|

August 18, 2026

On July 10, 2026, the most valuable company in the world accused the most talked-about company in the world of theft. Apple sued OpenAI in the U.S. District Court for the Northern District of California, alleging that OpenAI built its hardware ambitions on a foundation of Apple’s misappropriated trade secrets.¹ Few disputes touch as much of Certum’s Trade Secret Litigation Playbook at once: reasonable measures to guard a secret, identifying with particularity what was taken, and the human-centered points — recruiting and employee departures — where secrets actually walk out the door. Nearly every core theme in Certum Group’s Trade Secret Guide is in this case. And the lesson beneath it is worth sitting with: for the companies with the most to protect, trade secret litigation is not a last resort. It’s a front-line instrument of competitive strategy.


Background


The dispute sits at the intersection of two of the most closely watched storylines in technology. In 2025, OpenAI acquired io, the hardware venture founded by former Apple design chief Jony Ive and a group of other Apple alumni, for a reported $6.5 billion, and set out to build its first consumer hardware device, widely expected to compete directly with the iPhone.² To staff that effort, OpenAI hired aggressively from Apple. According to the complaint, more than 400 former Apple employees now work at OpenAI.³


Two of those hires anchor Apple’s allegations. Tang Yew Tan spent roughly 24 years at Apple, where he served as a vice president of product design responsible for the iPhone and Apple Watch, before becoming OpenAI’s chief hardware officer. Chang Liu spent about eight years at Apple as a senior systems electrical engineer before departing for OpenAI in 2026.⁴ Apple’s theory is not that a single rogue employee walked out the door with a file. It is that the movement of talent was accompanied by a coordinated effort, one Apple describes as operating “at every level," to extract and exploit the confidential information those employees carried in their heads and on their devices.⁵


The Allegations


The complaint reads less like a garden-variety departure dispute and more like a catalog of the exact conduct the Trade Secret Guide warns companies to watch for. Among Apple’s central allegations:


Apple claims OpenAI’s hardware leadership directed recruiters to use Apple’s confidential project code names during the hiring process, and instructed job candidates to bring “actual parts” and “CAD/design artifacts” to their interviews.⁶ It alleges that OpenAI circulated internal Apple documents marked “Need to Know” that coached departing employees on how to evade Apple’s exit-security procedures, including the “dreaded walkout,” and to alert OpenAI before signing their exit agreements.⁷


The specifics attributed to individual employees are what give the complaint its texture. Apple alleges that Chang Liu exploited an authentication bug to reach internal network storage after his access should have been cut off, messaging a colleague, “LOL, I found out I can access the [network storage], so funny,” and noting within hours of his departure that he “still ha[d] another computer.”⁸ And Apple alleges that io “exploited and used Apple’s secret, proprietary industrial design techniques,” misleading one of Apple’s own manufacturing partners about whether it was authorized to use a confidential metal-finishing technique.⁹


The trade secrets Apple says are at risk span the full arc of its product-development process: technical specifications for unreleased technologies, engineering presentations and prototype data, component and vendor selection processes, and the proprietary manufacturing techniques that turn a design into a shippable product.¹⁰ Notably, Apple’s opening ask is not a damages windfall. It is protection. Apple seeks to bar OpenAI from using or disclosing the information at issue, to compel the return of its confidential materials, and to preserve the evidence.¹¹ In other words, Apple is using the courthouse to do what its NDAs and exit interviews were supposed to do: keep its edge inside the building.


OpenAI’s Response


OpenAI has pushed back hard, and its answer is a preview of the fault lines any trade secret plaintiff should expect to fight over. On August 6, 2026, OpenAI moved to dismiss, characterizing the alleged conduct as “benign, lawful conduct” that Apple has mischaracterized, and arguing that its hardware executives simply followed standard industry recruiting practices.¹² As to Chang Liu, OpenAI contends he was “trying to help Apple” by assisting former colleagues who asked him to locate work information, not stealing anything.¹³


More pointed, and more instructive, is OpenAI’s argument that Apple’s own conduct undermines its case. OpenAI asserts that Apple allowed employees to use personal iCloud accounts for work and failed to properly revoke access when they left — noting that an Apple manager remained logged into Chang Liu’s personal iCloud account after his departure in order to transfer files.¹⁴ From that, OpenAI argues that Apple’s offboarding lapses created “confusion and unwanted access issues that Apple now characterizes as theft.” OpenAI also contends that Apple has not identified its trade secrets with adequate specificity, pointing instead to “generic categories of the product-development process.”¹⁵ OpenAI must file its full response by August 17, 2026, with oral argument on the motion set for October 1, 2026.¹⁶


Whatever the merits, OpenAI’s playbook is worth studying precisely because it is so conventional. Reasonable secrecy measures and identification of the trade secret with particularity are two of the elements every misappropriation claim rises or falls on, and they are exactly where a well-resourced defendant will apply pressure first.


What This Means


It is easy to read a case like this as celebrity litigation between two of the most valuable enterprises on earth. The more useful reading is that trade secret law has become core infrastructure for how modern companies protect competitive advantage. Apple did not respond to a $6.5 billion competitive threat with a press release or a patent portfolio. It responded with a trade secret complaint, because in a business where the crown jewels are unpatented know-how — manufacturing techniques, vendor relationships, unreleased designs — the Defend Trade Secrets Act and its state-law counterparts are the sharpest tools available.


The case also throws the Trade Secret Guide’s central lessons into relief. The value of a trade secret program is only as good as the “reasonable measures” behind it; OpenAI’s opening move is to argue that Apple’s own iCloud and offboarding practices were not reasonable at all. The ability to describe what was taken, with specificity, is not a formality. It is frequently the whole ballgame at the pleading stage. And the human element — recruiting, exit procedures, the “dreaded walkout” — is where secrets actually leak, long before anyone reaches a courtroom. Companies that treat these as compliance checkboxes learn the hard way, in a complaint, that they were the strategy all along.


For those of us who evaluate disputes for a living, Apple v. OpenAI is also a reminder of why high-stakes trade secret matters are among the most compelling on the plaintiff’s side. The conduct is often concrete and documentable, the competitive stakes are enormous, and, as the Federal Circuit’s recent decision in Versata Software v. Ford underscored, the damages framework can reach the full value of what the misappropriation delivered to the wrongdoer, not merely a discounted license fee. That combination is exactly what makes these cases worth pursuing, and worth backing.


Apple’s complaint will be tested, as it should be, and the allegations remain just that — allegations. But the strategic signal is already unmistakable. When the most valuable company in the world wants to defend its future, it reaches for trade secret law. Certum Group’s Trade Secret Guide is built to help plaintiffs and their counsel do the same, whatever their size, and this case is a live illustration of why that playbook matters now more than ever.


Certum Group can help. If you are evaluating a trade secret dispute or want to talk through options for funding or de-risking one, get in touch.


Footnotes


¹ Complaint, Apple Inc. v. OpenAI, Inc., No. 5:26-cv-07078 (N.D. Cal. filed July 10, 2026); see Apple sues OpenAI over alleged trade secret theft, TechCrunch (July 10, 2026).

² The wildest allegations in Apple's trade secrets lawsuit against OpenAI, TechCrunch (July 13, 2026).

³ Id.

⁴ Apple sues OpenAI over alleged trade secret theft, TechCrunch (July 10, 2026).

⁵ Apple sues OpenAI alleging trade secret theft, says scheme was "at every level," CNBC (July 10, 2026).

⁶ The wildest allegations in Apple's trade secrets lawsuit against OpenAI, TechCrunch (July 13, 2026).

⁷ Id.

⁸ Id.

⁹ Id.

¹⁰ Apple sues OpenAI over alleged trade secret theft, TechCrunch (July 10, 2026).

¹¹ Id.

¹² OpenAI Asks Judge to Toss Apple's Trade Secrets Lawsuit, Claims Journal (Aug. 7, 2026).

¹³ Id.

¹⁴ OpenAI says Apple's own security practices undermine its trade secrets case, TechCrunch (Aug. 6, 2026).

¹⁵ Id.

¹⁶ OpenAI Asks Judge to Toss Apple's Trade Secrets Lawsuit, Claims Journal (Aug. 7, 2026).


Certum Group Can Help

Get in touch to start discussing options.

Recent Content

By Patrick Dempsey • September 29, 2026
This is the third post in Certum Group's seven-part series bringing our Trade Secret Litigation Playbook to the blog. It draws on Part III of the Playbook, Before You File. Read or download the full Playbook here . The call usually comes from a sales leader, a head of engineering, or a chief of staff, and it usually starts the same way: "We think someone took X, and we think they took it to Y." What happens in the next three days is disproportionately important. The most expensive mistake in trade secret practice is waiting — every week that passes lets the defendant commercialize your advantage, and worse, lets the evidence quietly disappear. Slack messages, badge records, and git histories often sit behind auto-delete settings as short as ninety days. So before you do anything else, move to preserve. Here is what good early triage looks like. Issue a litigation hold — in writing, within 24 hours A written instruction to preserve any document, chat, email, calendar, or file relating to the possible misappropriation, sent to the full legal, HR, IT, and leadership teams. It should be specific enough to be useful, broad enough to catch the unexpected, and documented enough that you can produce it later. This is the single most important thing you can do in the first day. Lock down forensic images — and don't let internal IT do it Laptops, phones, and company devices belonging to any suspected party should be forensically imaged by an outside forensics firm, not wiped or "checked" by internal IT. Chain of custody matters, and internal teams generally cannot testify to it at trial. The right instinct is to image broadly and review narrowly: you can always decline to look at a device you preserved, but you can never go back for one that was reissued, wiped, or sold. And don't forget the non-obvious sources — personal cloud drives, USB connection histories, printer spooler logs, and screen-capture utilities are often where the decisive evidence actually lives. Extend retention on cloud and platform logs Have IT extend retention on the relevant accounts and export the log data — access logs, download histories, egress traffic — before anything rolls off. The forensic story of who did what, when, and with which file is usually more persuasive to a judge than any human witness. A download at 11:47 p.m. the night before a resignation is powerful evidence because of what the timing means. Do not tip off the adversary Resist the urge to confront the suspected employee, their new employer, or their counsel. Confrontation at this stage tends to accelerate deletion and pull communications behind privilege claims. A carefully timed cease-and-desist letter, sent after preservation is secure, is a very different move — and a far more effective one. Investigate under privilege Any investigation you run should be conducted under attorney direction, so that the work product is privileged and your witness interviews do not become admissions used against you later. This is not about secrecy for its own sake. It is about making sure the investigation helps your case rather than becoming evidence in it. Consider — but do not rush — the early public filing Ex parte seizure orders, temporary restraining orders, and preliminary injunctions all have their place, and an early injunction can be the single most valuable outcome in the whole case. 1 But these motions require a level of evidentiary support you rarely have on day three. It is almost always better to spend a week building the record than to file fast and lose the first motion, which hands the defendant a narrative and hands you a hole to climb out of. The triage checklist If you want a single page to keep by the phone, it looks like this: litigation hold issued in writing within 24 hours; outside forensics firm retained and imaging scheduled; cloud and platform log retention extended; a list of suspected individuals plus their managers and peers; a list of the specific secrets that may have been taken; copies of every NDA, employment agreement, and IP assignment covering them; a review of the last 90 days of their calendar for unusual patterns; and outside trade secret counsel engaged under privilege. Do the first 72 hours well and you preserve every option that follows — injunction, damages, settlement leverage. Do them poorly and you may spend the next two years litigating around evidence you could have saved in an afternoon. If you are in the opening days of a matter and want a fast, confidential read on what to do now versus later, that is exactly the conversation we have most often. Go deeper with the Playbook. This post covers one piece of a much larger picture. For the full framework — what the law requires, what a strong pre-filing case looks like, how damages experts value these matters, how counsel fee structures change your economics, and how litigation finance fits in — read Certum Group's Trade Secret Litigation Playbook , our field guide for business owners and the counsel who advise them: certumgroup.com/the-trade-secret-playbook . And if you are evaluating a live dispute — or simply want to pressure-test what a matter is worth and how it might be funded — get in touch. A confidential conversation with Certum is free and carries no obligation, whether or not you ultimately seek funding. Reach us at certumgroup.com/contact-us . Sources 1. The ex parte seizure procedure is authorized by the Defend Trade Secrets Act, 18 U.S.C. Section 1836(b)(2), and is available only in extraordinary circumstances.
By Patrick Dempsey • September 15, 2026
This is the second post in Certum Group's seven-part series bringing our Trade Secret Litigation Playbook to the blog. It draws on Part II of the Playbook, Trade Secret Law in Plain English. Read or download the full Playbook here . Here is a statistic that surprises most executives: in federal trade secret cases that reach a verdict, plaintiffs win roughly 84% of the time. 1 That is dramatically better than the plaintiff win rate in commercial litigation generally. It does not mean every case is easy — the cases strong enough to reach a jury are a selected group — but it tells you something important about what happens when a well-built trade secret claim gets in front of a fact-finder: courts tend to enforce the rights the statute was designed to protect. So why do good claims still fail? Usually not at trial. They fail earlier, on assumptions the claim holder never stopped to test. In more than a decade of evaluating these matters, the same handful of misconceptions come up again and again. Here are six worth clearing up before they cost you a case. Myth 1: "It's only a trade secret if we stamped it CONFIDENTIAL." Marking helps, but it is not required. What matters is whether your overall secrecy program is reasonable under the circumstances — a holistic look at contractual, physical, and technical controls. A perfect stamp on an otherwise open system is worth less than a coherent program with a few gaps. Myth 2: "If part of it is public, none of it is protected." Courts routinely protect a combination of individually public facts when the particular combination delivers competitive advantage. The recipe can be assembled from ingredients anyone can buy. What you protect is the assembly. Myth 3: "Our employees signed NDAs, so we're covered." NDAs are a foundation, not a program. The full set of reasonable measures a serious claim holder is expected to have includes access controls, badging, egress monitoring, exit procedures, and technical segmentation. An NDA in the drawer and nothing behind it is exactly the gap a well-resourced defendant will press on first. Myth 4: "We didn't sue the last person who left, so it's too late now." Trade secret protection is evaluated case by case. Declining to act on one departure does not forfeit your rights as to the next one. Every matter stands on its own facts. Myth 5: "We're too small to enforce against a big company." This one gets the economics backwards. A well-funded claim against a large, solvent defendant is often easier to win — and easier to collect — than a disorganized claim against a small one. Resources can be added to a strong case; facts cannot be added to a weak one. The right capital partner exists precisely so that a smaller plaintiff can stand toe-to-toe with a much larger adversary and neutralize the outspend-them tactics that used to decide these fights. Myth 6: "It's just know-how — courts don't protect that." Courts protect integrated know-how constantly. The question is never whether know-how is capable of protection; it is whether you can identify it with enough particularity to describe what was taken. 2 That is the single most consequential early decision in the case, and it is the one most claim holders do not realize they are making when they plead "our proprietary software" instead of the specific, described combinations that actually give them an edge. The through-line Notice what these myths have in common: each one leads a claim holder to under-invest in a case that the numbers say is very winnable. The 84% figure is not a promise. It is an invitation to take the early work seriously — the secrecy program, the identification, the evidence — because that work is what turns a strong set of facts into a strong case. Get those right, and the law is on your side more often than in almost any other kind of commercial dispute. Go deeper with the Playbook. This post covers one piece of a much larger picture. For the full framework — what the law requires, what a strong pre-filing case looks like, how damages experts value these matters, how counsel fee structures change your economics, and how litigation finance fits in — read Certum Group's Trade Secret Litigation Playbook , our field guide for business owners and the counsel who advise them: certumgroup.com/the-trade-secret-playbook . And if you are evaluating a live dispute — or simply want to pressure-test what a matter is worth and how it might be funded — get in touch. A confidential conversation with Certum is free and carries no obligation, whether or not you ultimately seek funding. Reach us at certumgroup.com/contact-us . Sources 1. Stout, Trends in Trade Secret Litigation (2024), reporting an approximately 84% plaintiff-favorable outcome rate across 271 federal trade secret cases reaching a verdict since 2017. 2. Both the federal Defend Trade Secrets Act (18 U.S.C. Section 1836) and the state Uniform Trade Secrets Act define a trade secret as information that derives independent economic value from not being generally known or readily ascertainable, and that is the subject of reasonable measures to keep it secret.
By Kevin Skrzysowski • September 10, 2026
For most companies, the legal department is viewed purely as a cost center—a line item to be managed and minimized. But many organizations are sitting on significant, unrealized value in the form of affirmative claims: lawsuits they could bring against suppliers, vendors, or competitors who have breached a contract, stolen trade secrets, infringed intellectual property, or otherwise caused recoverable damages. All too often these valuable claims go unpursued because litigation is expensive, unpredictable, and competes with the business for budget and headcount. Certum Group, in conjunction with the Corporate Counsel Business Journal, created this one-hour webinar to show in-house counsel how litigation finance changes that calculus— allowing companies to pursue meritorious claims with little or no out-of-pocket cost and on a non-recourse basis, so the downside risk shifts to the funder while the company retains the upside. Our panel brought together leaders in commercial litigation, intellectual property litigation, and legal academia and demystified how litigation finance works and walked through practical, real-world uses for the corporate legal team. We also took a close look at claim monetization: the ability to receive cash today against the value of a pending or contemplated claim, rather than waiting years for a judgment or settlement. Monetization can take the form of an upfront advance secured by the expected recovery, or an outright sale of the claim to a specialized organization that then prosecutes the case and collects the judgment. Attendees left with an understanding of when litigation finance and monetization make sense, how to evaluate and pitch a case, what to expect from the process, and how these tools can transform the legal department from a cost center into a genuine contributor to the bottom line. Watch the full webinar replay HERE . Supporting Materials: Litigation Finance Guide In-House Survey Research Brief Trade Secrets Playbook