Apple says OpenAI recruited hires to siphon hardware trade secrets — discovery will decide
Apple‘s federal lawsuit against OpenAI alleges a targeted campaign to extract confidential hardware information through recruited employees, naming former Apple VP Tang Yew Tan and ex-engineer Chang Liu. The case combines concrete accusations about interview practices and file transfers with the larger context of OpenAI’s 2024 io Products buy and a 2025 shift by Apple away from OpenAI models for Siri.
Specific accusations: show-and-tell requests and retained files
Apple’s complaint says Tang Yew Tan directed candidates to bring Apple parts to “show and tell” interview sessions so OpenAI could learn manufacturing details, supplier relationships, and unreleased designs. The suit also alleges Tan advised departing staff on bypassing security protocols, and that Chang Liu kept an Apple laptop and downloaded dozens of confidential engineering documents after joining OpenAI.
OpenAI’s hardware push versus Apple’s protected know‑how
OpenAI’s $6.4 billion acquisition of io Products in 2024 — Jony Ive’s startup — signaled a concrete pivot into consumer hardware that puts it in direct overlap with Apple. Apple’s move in 2025 to switch Siri’s backend from OpenAI models to Google’s Gemini illustrates how commercial and technical ties between the companies have frayed as hardware stakes rose.
What discovery needs to establish to separate signal from noise
The crucial evidentiary questions are specific: do device files or internal notes showing Apple supplier part numbers, manufacturing tolerances, or unreleased product drawings appear in OpenAI systems; can access logs tie those files to named individuals; and do product design elements at OpenAI trace back to the allegedly transferred materials? If discovery finds chain‑of‑custody records, timestamped downloads from Apple servers by named defendants, or the use of Apple supplier contacts in OpenAI procurement, Apple gains traction for injunctions and damages. OpenAI will counter with lawful hiring norms, claimed independent development, and a lack of overlap in final product designs — a defense that makes detailed forensic evidence decisive.
| Allegation | Evidence Apple cites | What discovery can confirm | Consequence if proven |
|---|---|---|---|
| Directed candidates to bring Apple parts | Interview notes, witness statements | Emails, calendars linking requests to Tan; interview photos/files | Injunction on use; limits on OpenAI’s hardware hiring practices |
| Retention and downloading of confidential files | Alleged laptop retention, copies of engineering docs | Forensic images, server logs, metadata linking files to Liu | Monetary damages; potential criminal exposure for theft |
| Use of supplier or manufacturing know‑how | Overlap in supplier lists or specs | Procurement records, vendor communications, prototype comparisons | Contract disputes with suppliers; reputational harms |
Practical checks companies and regulators should expect
For corporate counsel and in‑house security teams the immediate decision lens is whether existing exit controls, device‑wipe procedures, and non‑disclosure enforcement are sufficient. Apple says it warned OpenAI in February 2025 about misuse and received no response; that allegation raises a compliance question for every partner: did prior warnings produce remediation or just silence?
Regulators and investors will watch two constraints: first, whether courts issue injunctive relief that limits talent mobility or specific hiring channels; second, whether discovery discloses supplier or manufacturing links that attract broader contract or antitrust scrutiny. For OpenAI, the timing matters — an IPO process intensifies scrutiny of unresolved litigation and the risk profile tied to hardware ambitions.
Short Q&A
Will this stop OpenAI’s hardware products? Not immediately — Apple seeks injunctions but courts weigh harm and evidence; detailed forensic discovery will determine whether specific products must be altered or blocked.
How long until discovery clarifies the facts? Expect months: complex device forensics, metadata analysis, and depositions typically expand pretrial timelines, especially where cross‑border supplier data is involved.
What should suppliers and hiring teams do now? Reinforce device return/exchange policies, audit access logs, and treat vendor contacts as potential vectors for litigation until discovery resolves whether supplier details were misused.
