On 10 July, Apple filed a lawsuit against the OpenAI Foundation, OpenAI Group PBC, io Products, and two former Apple employees in a federal court in Northern California.
In the lawsuit, Apple claims that OpenAI systematically obtained confidential data about its products, suppliers, and manufacturing processes to accelerate the development of its own mass-market AI device.
OpenAI has denied the allegations, stating that it has no interest in others’ trade secrets and remains focused on developing its own technology. The case was brought several months before the expected unveiling of OpenAI’s first device.
Apple is seeking to protect the technical know-how developed in the design and production of the iPhone, Apple Watch, and other products. OpenAI is trying to move beyond the role of a software provider that reaches users only through other companies’ screens, operating systems and app stores.
The outcome could determine how quickly the leading AI company becomes a device maker, and how much knowledge its new hires are allowed to bring with them from a company that has spent decades building the world’s most profitable consumer hardware ecosystem.
The allegation goes beyond a few stolen files
The most serious aspect of the lawsuit concerns the specific evidence that Apple claims to have discovered. Former senior system electrical engineer Chang Liu allegedly failed to return a company-issued work laptop after leaving in January 2026.
Apple claims he then exploited a previously unknown authentication bug to access an internal data network, where he downloaded dozens of documents on hardware, development projects and engineering procedures.
The lawsuit also describes his communications with then Apple engineer Yu-Ting Peng, who moved to OpenAI in April, alleging that she continued to pass him data on projects, suppliers, and technical issues.
Tang Yew Tan, Apple's former Vice President of Product Design, occupies an even more prominent place in the lawsuit. After 24 years at Apple, Tan moved to io Products and now leads hardware development at OpenAI.
Apple claims he sent himself supplier data and internal industry and supply chain analyses before leaving, and then asked Apple engineers he was trying to recruit to OpenAI's hardware team to bring components and product samples from their current projects to interviews.
OpenAI will try to show that the disputed actions were personal decisions made by employees
The documents mention batteries, compact modules that integrate the processor, memory and other electronic components, the device’s main electronic boards and metal shields against electromagnetic interference, as well as presentations containing confidential information about Apple projects.
The lawsuit seeks to extend liability from individuals to the company. According to Apple, OpenAI used knowledge of procedures to advise employees on how to avoid security checks when leaving, and allegedly used confidential data in dealings with suppliers.
Apple claims that one of its suppliers, at OpenAI's request, implemented a special metal finishing technique in the belief that Apple had authorised the use of the technology.
According to the lawsuit, other suppliers received questions from OpenAI that used Apple's internal component names and technical details known only to a narrow circle of employees and partners.
The key question is whether the confidential data reached OpenAI through the actions of a few employees or through the way the company organised its hiring and hardware development.
If Apple can prove that executives knew about the acquisition of that data, encouraged it, or used the information obtained in dealings with suppliers, liability will not rest solely with those individuals.
OpenAI will therefore try to show that the disputed actions were personal decisions made by employees, that the company was not aware of them, and that it developed its device independently.
What belongs to the engineer, and what belongs to Apple?
Apple says that more than 400 of its former employees now work at OpenAI. That information alone proves nothing. Employees in California have the right to move to a competitor and use the experience, skills and knowledge they have acquired throughout their careers.
It becomes controversial only if they bring confidential documents, technical data, prototypes or other information belonging to their former employer. Apple must show what information is a trade secret, how it was protected and how it was obtained or used without permission.
A blanket claim that ex‑engineers “know the iPhone” will not suffice. A digital trail showing file downloads after termination of employment, messages about the delivery of physical components, and communications with suppliers could, if confirmed, constitute much more serious evidence.
OpenAI is likely to argue that some of the data Apple considers secret is already common knowledge within the industry and that its technical solutions were developed independently.
Monetary compensation comes later and can hardly offset the advantage a competitor would gain by entering the market several months earlier
At the same time, it will require Apple to show exactly what was downloaded, who had access to it, and how OpenAI allegedly used it for each disputed piece of information.
Such a request could narrow the lawsuit, but it would also trigger a detailed examination of business computers and phones, user accounts, messages, project documents, and communications with suppliers.
For Apple, the primary objective is securing an injunction. Monetary compensation comes later and can hardly offset the advantage a competitor would gain by entering the market several months earlier.
A targeted ban on the use of certain documents, processes or supplier relationships could force OpenAI to change the design, replace components or prove that development was carried out separately. A broader ban on ex‑Apple employees would be much more difficult to sustain in California.
OpenAI wants direct access to the user
In 2024, OpenAI partnered with Apple, making ChatGPT available through Siri and other features on Apple devices. For a company without its own hardware or distribution network, the iPhone was the quickest route to hundreds of millions of users.
However, Apple retained control over the entire relationship with the user, deciding if, when and how ChatGPT would be included.
The purchase of io Products for about $6.5 billion showed that OpenAI wanted much more than a slot inside the iPhone. The company was founded by Jony Ive and a group of former Apple designers and engineers, including Tan.
Ive is not named in the lawsuit, but his involvement in the project confirms that OpenAI is developing a serious mass-market device.
Apple is trying to preserve control over the device and user access, while OpenAI wants to stop depending on iOS and Android
The goal is for the company to reach users directly, through its own product, rather than through Apple’s devices and rules. Such a device would allow OpenAI to determine for itself how people use its services and to develop an assistant that understands their requirements, circumstances, and daily habits.
OpenAI is developing its own device at a time when Apple is trying to close the gap with leading companies in generative artificial intelligence. Apple still controls the iPhone, the operating system and access to a huge user base, but OpenAI and Google have already launched AI products that are now part of the daily habits of millions of people.
The new version of Siri arrived after several delays and, to handle more complex requests, relies on remote servers instead of doing everything on the device itself. OpenAI has ChatGPT, but it does not yet have its own device to reach users directly, without an intermediary.
Apple is therefore trying to preserve control over the device and user access, while OpenAI wants to stop depending on iOS and Android.
The real advantage is created in the factories
OpenAI knows how to develop a model and software, but manufacturing a device requires years of experience. Battery life, heat management, microphone and camera quality, choice of materials, component reliability, and the organisation of large-scale production all need to be addressed.
Apple has built a distinct advantage here: it knows how to turn a good design into a product that can be manufactured in millions of units without any loss of quality.
That is why allegations concerning suppliers and production processes are particularly serious. Information about who can produce a given component, at what price, within what timeframe and with what level of quality can save a competitor years of development time.
OpenAI intends to present its first device by the end of 2026, and the court case may slow its development at the final stage
Data on failed solutions are equally valuable, as they show what should be avoided. If Apple proves that OpenAI used such information, the dispute will no longer be limited to employee transfers but will hinge on whether OpenAI accelerated the development of its device by relying on Apple’s confidential knowhow.
The timing of filing the lawsuit has a clear business logic. OpenAI intends to present its first device by the end of 2026, and the court case may slow its development at the final stage.
Examining computers and messages, questioning employees and suppliers, and determining whether disputed data were used in the project can delay production and entry to the market.
A few extra months would also be valuable for Apple to improve Siri, strengthen its AI systems, and prepare a response to a device whose final form and use have not yet been publicly announced.
The court battle begins before the market battle
OpenAI will likely try to limit the dispute to specific documents and data that Apple can show actually ended up in its project. Apple, by contrast, will seek prompt access to computers, messages, project documentation and communications with suppliers to determine whether confidential information was used in developing the device.
It is unlikely that a court would immediately halt the entire project without solid proof that Apple’s secrets are embedded in the product.
It is far more realistic that work on certain parts of the device would be temporarily limited until OpenAI demonstrates that they were developed without using the disputed data.
Apple is in no rush to settle now. First, it wants to determine how much of its data has been transferred, who has used it, and whether it has already influenced device designs, prototypes, or deals with suppliers.
OpenAI will almost certainly not abandon the project, but the introduction of the device by the end of 2026 is now much less certain
OpenAI will try to keep as many product details as possible out of the court proceedings, especially as launch nears. Because of this, the conflict is likely to escalate at first, and serious talks about a deal will only begin once both sides see what the evidence actually shows.
The partnership between Apple and OpenAI is unlikely to end immediately, as ChatGPT is already used by millions of Apple device owners and an abrupt termination would harm both companies.
However, their relationship has fundamentally changed. Apple will increasingly invest in its own AI systems and in collaboration with Google, while OpenAI will accelerate the development of a device that allows it to reach users without Apple’s mediation.
Even if the dispute ends in a settlement, it would likely impose strict rules on hiring Apple staff, working with suppliers, returning confidential materials, and limiting the role of former Apple employees in developing OpenAI’s device.
OpenAI will almost certainly not abandon the project, but the introduction of the device by the end of 2026 is now much less certain.
The most likely outcome is a delay, or a limited launch of the product, until it is verified whether questionable data, components or manufacturing processes were used in its development.
For Apple, such a delay matters more than any potential compensation. Each additional month without its own device keeps OpenAI inside iOS and Android, operating under the rules of Apple and Google.
If the lawsuit slows the product’s entry into the market, Apple will achieve an important business goal even before the court issues a final decision.
This dispute will demonstrate how far a company may use the knowledge of people it has recruited from a competitor, and where the unlawful use of confidential data begins.
OpenAI must now demonstrate that its device was developed in-house, without using Apple’s documents, technical solutions or supplier connections.
For Apple, however, the final verdict may not even be the most important outcome. It is enough that the process slows development, delays product launches and forces OpenAI to prove the origin of its solutions for months.
In the battle to create the device through which people will use artificial intelligence on a daily basis, even a delay like this can be a major victory.
