Apple has intensified its legal campaign against OpenAI by sending evidence-preservation letters to approximately 40 former Apple employees who now work for the artificial-intelligence company, according to the Financial Times. The letters reportedly direct recipients to retain documents and communications that could be relevant to Apple’s trade-secret lawsuit and ask them to meet with the iPhone maker’s legal team. The step expands the practical reach of a case that initially named two former employees and signals that Apple is examining whether the conduct alleged in its complaint involved a broader group of workers, recruiting contacts and business relationships.

The letters follow Apple’s lawsuit filed July 10 in the U.S. District Court for the Northern District of California. The complaint names OpenAI Foundation, OpenAI Group PBC, io Products and former Apple employees Chang Liu and Tang Yew Tan as defendants. Apple brought claims involving alleged trade-secret misappropriation and breach of contract and requested a jury trial. It is seeking measures that include preventing the defendants from possessing, using or disclosing Apple trade secrets, requiring the return or preservation of materials and obtaining damages tied to any proven losses.

No court has determined that OpenAI, io, Liu or Tan misappropriated Apple information. The allegations remain claims made by Apple at the opening stage of litigation, and the defendants will have opportunities to challenge both the factual assertions and the legal sufficiency of the complaint. OpenAI has said it has no interest in other companies’ trade secrets and remains focused on developing its own technology. The company has also indicated that it has not seen evidence supporting Apple’s broader characterization of institutional misconduct.

A preservation letter is not itself a lawsuit against its recipient, nor does it establish that the recipient engaged in wrongdoing. Such communications generally warn potential witnesses or custodians of evidence not to delete emails, messages, files, photographs, device backups or other records that may become relevant. Recipients can include people who merely observed events, participated in hiring discussions, worked on related projects or possessed communications with individuals named in a complaint. The approximately 40 people contacted by Apple could therefore occupy very different positions in the developing case.

The scale of the reported outreach is nevertheless significant. Apple said in its complaint that it had uncovered what it described as a pattern involving former employees who moved to OpenAI. According to reporting on the filing, Apple estimates that more than 400 of its former employees now work at the AI company. Contacting about 40 of them would give Apple access to a potentially large body of material concerning departures from Apple, OpenAI’s recruiting procedures, hardware interviews, supplier communications and the formation of its consumer-device team.

Apple’s most detailed allegations concern Liu, who worked at Apple for more than eight years as a senior system electrical engineer before joining OpenAI in January 2026. Apple claims Liu did not complete required exit procedures or promptly return an Apple-issued computer. The complaint further alleges that after joining OpenAI, he discovered and used an authentication flaw that allowed access to shared Apple network folders. Apple says Liu then downloaded dozens of confidential hardware-related files, including technical specifications, engineering presentations, proprietary project data and information about unreleased products.

The complaint also alleges that Liu communicated with an Apple employee he was recruiting and advised that person about preparing for an OpenAI interview involving Apple technology. Apple claims he discussed ways to avoid attention from its security team and suggested using a separate messaging service. Liu has not been found responsible for those allegations, and the complaint represents Apple’s account of communications and technical records that will be tested through the litigation process.

Tan is a more senior figure in the dispute. He spent about 24 years at Apple and most recently served as vice president of product design for the iPhone and Apple Watch, according to the complaint. He later co-founded io Products and is now OpenAI’s chief hardware officer. Apple alleges that before leaving, Tan sent himself supplier information and internal industry summaries and later used confidential knowledge when interviewing Apple employees for positions connected to OpenAI’s hardware operations.

Apple and OpenAI logos represent an escalating trade-secret dispute involving former employees and consumer hardware development.

Among Apple’s most consequential assertions is that Tan instructed some candidates still employed by Apple to bring physical components to interviews for demonstrations. The complaint says OpenAI interview processes sought information involving design artifacts, prototypes, subsystem choices, integration tools, vendor selection and supplier collaboration. Apple also claims an internal document describing its employee-departure security procedures was circulated among workers preparing to leave for OpenAI. The defendants have not been found liable for these alleged activities.

Apple is attempting to frame the case as more than an isolated dispute involving files carried away by departing engineers. Its complaint argues that OpenAI and io used information involving Apple’s product-development methods, manufacturing capabilities and supplier network to support a coordinated entry into consumer hardware. Apple further alleges that confidential knowledge was used in discussions with trusted manufacturing partners, including in connection with a metal-finishing technique that it considers proprietary.

The preservation letters could help Apple test that institutional theory. Evidence from a wider group of former employees may show whether the disputed actions were unusual decisions by particular individuals, common practices within a team or conduct directed by managers. Communications surrounding interview preparation, offboarding, access permissions, supplier meetings and the storage of technical documents could become central. The parties may also dispute whether particular information qualified as a legally protected trade secret, was already known in the industry or represented general expertise acquired through years of engineering work.

That distinction is especially important in California, where technology companies compete aggressively for skilled workers and employee mobility is a defining feature of the labor market. Employers can protect specifically identifiable confidential information, but they do not own an employee’s general knowledge, abilities or professional experience. Apple will therefore need to identify the information it believes was secret, demonstrate that it took reasonable steps to protect it and connect any acquisition or use of that information to the defendants’ conduct.

OpenAI, in turn, may seek to show that its hardware work was independently developed, based on publicly available knowledge or derived from expertise that employees were free to use. It could also challenge whether Apple’s descriptions of broad areas such as supplier management, product integration or manufacturing methods are sufficiently specific to constitute trade secrets. Protective orders are likely to be important because Apple may have to disclose highly sensitive information to explain its claims without exposing the very technology it says it is trying to protect.

The case reaches beyond legal doctrine because hardware has become a central strategic project for OpenAI. The company acquired io Products, the design and engineering business co-founded by former Apple design chief Jony Ive, Tan and others, in a transaction valued at nearly $6.5 billion. Ive was not named as a defendant in Apple’s complaint, but he leads OpenAI’s broader device-design effort. The acquisition gave OpenAI a team intended to create a new consumer interface for artificial intelligence rather than relying exclusively on phones, computers and third-party operating systems.

OpenAI has disclosed few concrete details about the planned product. Executives have described an effort to develop a new way for consumers to interact with AI beyond traditional screens and applications. Chief Financial Officer Sarah Friar said earlier in 2026 that consumer hardware was expected toward the end of the year. Reports have suggested a device incorporating sensors such as cameras and microphones, although OpenAI has not publicly confirmed a final design, feature set, price or commercial release schedule.

Apple has a direct competitive interest in how that project develops. Its business is built around tightly integrated devices, operating systems, services and custom silicon, while OpenAI is seeking a physical platform through which consumers could access ChatGPT and other AI services. A successful OpenAI device could create a new category, become an accessory to existing phones or compete for functions currently performed by smartphones, wireless earbuds, smart speakers and wearable products. The uncertainty surrounding the product makes Apple’s claims about unreleased designs and manufacturing knowledge particularly sensitive.

Apple and OpenAI logos represent an escalating trade-secret dispute involving former employees and consumer hardware development.

The confrontation is also notable because Apple and OpenAI remain commercially connected. Apple began integrating ChatGPT into its software ecosystem in 2024, allowing users to invoke the service for certain requests when Apple’s own systems could not provide an answer. That arrangement positioned OpenAI as a partner in Apple’s effort to improve artificial-intelligence features. The hardware initiative, employee migration and current lawsuit have transformed the relationship into one containing both cooperation and direct competitive tension.

For OpenAI, the immediate business risk is not necessarily a final damages award, which could be years away, but the possibility of operational disruption during discovery. Employees may need separate legal advice, devices and accounts may be examined, and product teams could face restrictions on handling disputed materials. Suppliers and prospective partners may receive subpoenas or requests to preserve their own records. Management attention could be diverted at a time when OpenAI is balancing model development, infrastructure investment, consumer products and preparations for a possible stock-market listing.

A request for preliminary injunctive relief would raise the stakes further. Apple could ask a judge to restrict the use of particular documents, processes or technologies before the case reaches trial. To obtain such relief, it would generally need to persuade the court that it is likely to succeed on relevant claims and faces harm that cannot be adequately repaired through money alone. There has been no public ruling establishing that OpenAI’s device program depends on Apple’s trade secrets, and any attempt to halt development would likely be vigorously contested.

The dispute may also influence hiring across the technology industry. Companies recruiting from direct competitors already use confidentiality reminders, clean-room procedures, restricted project assignments and legal reviews designed to prevent new employees from importing proprietary material. Apple’s allegations could lead OpenAI and other AI businesses to tighten those controls, particularly when hiring workers with access to unreleased devices, supplier pricing, engineering specifications or manufacturing processes. Former employees may also become more cautious about retaining devices, personal copies of work materials or communications concerning previous projects.

Investors will be watching whether Apple’s expanding evidence effort produces additional named defendants, new claims or a narrower case concentrated on the two individuals already identified. Preservation letters are often a precursor to interviews, subpoenas, depositions and forensic review, but they can also be routine measures designed to ensure records are not lost. Their significance will depend on the evidence ultimately obtained and whether it supports Apple’s assertion that the alleged conduct extended into OpenAI’s organizational practices.

The next procedural steps are expected to include formal responses from the defendants, potential motions challenging the complaint and negotiations over how confidential information will be handled. Apple may be required to identify its asserted trade secrets with greater precision before obtaining broad discovery into OpenAI’s hardware program. OpenAI is likely to seek limits on requests it considers overly expansive or aimed at revealing its own confidential designs. Disputes over personal devices, encrypted messaging and records held by former employees could become an early test of the case’s scope.

Apple’s decision to contact dozens of former workers indicates that it is preparing for a sustained legal contest rather than a narrowly contained disagreement. For OpenAI, the challenge is to defend its hiring and development practices without slowing a hardware project that is intended to reduce its dependence on other companies’ platforms. For Apple, the objective is to protect product and manufacturing information while showing that its claims concern identifiable secrets rather than an attempt to restrict legitimate employee movement. The outcome could shape both companies’ device strategies and establish a closely watched precedent for the expanding competition between AI laboratories and established consumer-hardware manufacturers.