Apple v. OpenAI: The Lawsuit That Hardware Built on Borrowed Secrets Cannot Survive

Thesis

OpenAI spent two years building a hardware business meant to define the post-smartphone era. Apple's complaint, filed July 10 in the Northern District of California, alleges that business was seeded with stolen physical components, coached security evasion, and misappropriated manufacturing techniques. The specific allegations are damaging. The adversary is categorically different from anyone OpenAI has faced before. And the timing—six weeks from a September IPO window—is close to catastrophic.

This is not a column about whether OpenAI is guilty. Nothing has been proven. No court has ruled. Every accusation below belongs to Apple's complaint, and I will attribute it accordingly. But the allegations carry weight on their own merits.

What Apple Actually Alleges

Apple alleges that the misconduct was directed by OpenAI's senior leadership, including Chief Hardware Officer Tang Tan, who used Apple's confidential project codenames during recruiting, asked job candidates to bring Apple hardware components to their interviews, coached departing Apple employees on evading Apple's security procedures, and solicited details about unannounced products.

The specifics cut deeper than generalized allegations. According to the filing, Tan "directed job candidates still working for Apple to bring 'actual parts' from Apple to their interviews for 'show and tell' sessions in which he and his team at OpenAI can elicit still more Apple confidential information." Apple is alleging that its Chief Hardware Officer directed candidates to physically remove unreleased components—batteries, logic boards, system-in-package modules—from Apple facilities.

Apple's complaint also alleges Tan possessed and distributed an internal Apple "Need to Know" document to new OpenAI hires before they gave notice, a document that included Apple's departure security protocols. The offboarding security checklist allegedly became an evasion manual.

On Chang Liu: the complaint describes Liu, a senior system electrical engineer with more than eight years at Apple, as having obtained access to Apple's internal network through a former colleague's laptop in January 2026 and pulling dozens of confidential hardware files, including details on unreleased products, engineering presentations, and technical specifications. Liu messaged a former colleague still at Apple: "LOL, I found out I can access the [network storage], so funny."

The suit also alleges that OpenAI used confidential Apple information to approach Apple's trusted partners, "even having one carry out a specific trade secret metal-finishing technique for OpenAI, misleading the partner to believe they had Apple's permission to do so." The alleged scheme extended beyond hiring into Apple's supply chain.

Apple sent a letter to OpenAI in February raising its concerns and received no response, the company said in the complaint. In litigation, documented silence functions as evidence of notice and willful inaction.

Apple's complaint characterizes the misconduct as running "at every level, from members of its Technical Staff to its Chief Hardware Officer," and calls OpenAI's hardware business "rotten to its core by its illegal reliance on misappropriated trade secrets." The filing adds: "This is the tip of the iceberg. Apple lacks visibility into what's been happening behind closed doors at OpenAI, where such misconduct is normalized and exemplified by leadership."

Why This Adversary Is Different

OpenAI's legal track record is strong. A jury dismissed Elon Musk's lawsuit on statute of limitations grounds, and that case was the single most visible legal obstacle to going public. A California court dismissed a separate xAI trade secret claim in June 2026. Publisher copyright litigation continues but has not disrupted operations. OpenAI's legal team has been effective.

Apple operates at a different scale. The company conducted a forensic investigation of its own device logs and server records before filing suit. The complaint includes alleged message-level communications, specific file download records, and named internal codenames. That documentary foundation exists before discovery opens.

The real leverage lies in what comes next. By litigating, Apple gains access to the discovery process to learn the full extent of the alleged operation. Apple's own complaint calls this "the tip of the iceberg"—a strategic legal statement about what it expects to find once it can compel production of OpenAI's internal communications, recruiting records, and hardware development documentation.

Over 400 former Apple employees now work at OpenAI. Each signed an NDA. Each went through an exit process. Each is a potential deponent. Apple knows what its trade secrets are, their commercial value, and which former employees now sit in OpenAI's hardware organization.

The remedies Apple seeks compound the threat. Apple wants the court to bar OpenAI from using or disclosing its trade secrets, compel return of all confidential materials, preserve evidence, and—per Macworld's reporting—force product redesigns and destruction of proprietary materials. The suit seeks injunctive relief and damages as OpenAI works to bring its first consumer hardware device to market. An injunction is not a fine. It freezes a product.

The IPO Timing

On May 22, 2026, OpenAI filed a confidential S-1 with the SEC for what could become the largest technology IPO in history at a valuation between $852 billion and $1 trillion. Goldman Sachs, Morgan Stanley, and JPMorgan lead the deal. OpenAI pulls in roughly $2 billion monthly while losing $1.22 for every dollar earned.

The public S-1 amendment, arriving weeks before the roadshow, will disclose audited financials and full risk factors. A federal trade secret suit naming the Chief Hardware Officer, alleging institutional misconduct, and seeking injunctive relief against the company's primary unreleased hardware product must be disclosed. It appears in risk factors. It sits in front of every institutional investor on the roadshow. It becomes the first question in every analyst meeting.

At Davos in January, OpenAI's Chief Global Affairs Officer Chris Lehane told Axios the hardware device would arrive in the first half of 2026. That window has already passed without a launch. An injunction motion could functionally freeze a product that has not shipped. The entire $6.5 billion IO Products acquisition thesis depends on that hardware reaching market.

The Counter-Arguments

Intellectual honesty requires engaging OpenAI's strongest defenses.

First, OpenAI spokesperson Drew Pusateri stated: "We have no interest in other companies' trade secrets. We remain focused on building innovative technology that empowers people everywhere." That is categorical denial.

Second, trade secret law demands more than a complaint suggests. Apple must establish at trial that each piece of information constitutes a protectable trade secret: information it took reasonable steps to protect and that derives independent economic value from secrecy. The wider Apple distributed technical specs internally, the harder that showing becomes before a jury.

Third, California rejects the inevitable disclosure doctrine. Employees have a statutory right to change employers and carry their general skills and expertise. OpenAI will argue that Tan's knowledge of Apple's supply chain is inseparable from 24 years of professional expertise, not a discrete stolen artifact.

Fourth, the "show and tell" allegation—the most vivid and damaging claim—may hinge on job candidate testimony from people with their own complex interests. Proving Tan directed the conduct, rather than that candidates volunteered information, is a factual question a jury decides.

Fifth, OpenAI's hardware program draws on Jony Ive's independent design work, OpenAI's own model capabilities, and a team with established methodology. Apple cannot enjoin all of that.

Trade secret litigation is fact-intensive, slow, and expensive for both sides. These are real defenses.

Sharpened Thesis

The counter-arguments are genuine. OpenAI may ultimately prevail. But the calculus here is asymmetric in ways the company has never faced.

Apple is not a freelance litigant with a grievance. It has already conducted a forensic investigation, possesses documentary evidence specific enough to name individual messages and file downloads, has severed its commercial relationship with OpenAI entirely, and has every institutional incentive to press this case to conclusion. The partnership is dead. OpenAI grew dissatisfied with how Apple surfaced ChatGPT within its products, and Apple announced in June that a rebuilt Siri would run on Google Gemini instead. Nothing remains to protect in a settlement.

OpenAI carries this case into an IPO targeting valuations above $850 billion, with a hardware product that has not launched, a Chief Hardware Officer named as a defendant, and a discovery process Apple's own complaint characterizes as likely to reveal far more than currently visible. That is a structurally untenable position for a company trying to price shares to public markets.

What to Watch

Three factors will determine whether this case reshapes OpenAI's trajectory or becomes a manageable disclosure:

Hardware launch timing. If OpenAI ships a device before any injunction issues, the litigation dynamic shifts substantially. If the product slips again, speculation about court-ordered redesign will compound the S-1 narrative.

What discovery surfaces. Apple told the court this is the tip of the iceberg. If discovery produces internal communications showing Tan or other officers directing the alleged conduct, enterprise liability becomes significantly harder to defend. Isolated individual conduct improves OpenAI's position.

S-1 risk factor language. When the public amendment arrives, read the litigation disclosure carefully. How OpenAI characterizes the suit's materiality will reveal more than any press statement about what the company's lawyers actually believe is at risk.