Two years ago they shared a stage. On 10 July 2026, Apple filed suit against OpenAI in federal court in Northern California, alleging trade secret theft on a scale it described as running “at every level” of the company. The Apple OpenAI lawsuit is a spectacular fight between two of the most valuable names in technology — and buried in the complaint is a set of lessons that apply just as sharply to a fifteen-person consultancy as to a trillion-dollar hardware firm.
What the Apple OpenAI lawsuit actually alleges
Apple’s complaint centres on hardware. Since OpenAI announced its move into consumer devices — and bought Jony Ive’s design startup io Products for a reported $6.4 billion — more than 400 former Apple employees have moved across. Apple argues that was not a hiring spree but a harvesting operation, and that product designs, manufacturing processes and supply-chain strategy walked out with them.
Two named individuals anchor the case. Apple alleges that OpenAI’s hardware chief, a former Apple vice president, directed Apple staff sitting job interviews to disclose confidential material as part of the process. A second former employee is accused of failing to return a company laptop and later exploiting an authentication bug to reach Apple’s internal network and download dozens of confidential hardware files.
OpenAI’s response was brief: it has “no interest in other companies’ trade secrets” and remains focused on building its own technology. The commercial backdrop is hard to ignore — the two firms integrated ChatGPT into Siri in 2024, and Apple has since announced that its rebuilt Siri will run on Google’s Gemini instead.
Why a small business should care about a billionaires’ lawsuit
Because the mechanics are identical at every scale, and only one side of this fight can afford a litigation team. When a salesperson leaves your agency, they take a mental map of your pricing, your pipeline and your best clients. When a contractor’s access is never revoked, they keep a live door into your files. What Apple is describing — interviews used as extraction, devices never returned, credentials that outlive employment — is not exotic corporate espionage. It is ordinary offboarding failure, at extraordinary stakes.
What has changed is that AI raises the value of the thing being taken and lowers the effort of taking it. Your prompt library, your fine-tuned workflows, your annotated client data, the internal playbook you spent two years refining: that is now a genuine competitive asset, and it is trivially copy-pasteable.
Five controls worth an afternoon
- Write down what is actually secret. A trade secret only has legal standing if you treated it as one. Name the assets — pricing models, client lists, methodology docs, prompt and workflow libraries — and mark them.
- Fix offboarding before you fix anything else. Same-day revocation of email, cloud drives, Slack, AI tool seats and shared logins. Get the laptop back. Apple, with all its resources, is in court partly because a laptop went missing.
- Grant access narrowly. Contractors and AI agents alike should see one folder, not the whole drive. The permissions you hand an autonomous tool deserve the same scrutiny you would give a new hire.
- Read the vendor terms. Know whether your inputs train someone else’s model, and where the data sits. The same question sits underneath the crawler rules that decide whether your website becomes training data.
- Get the paperwork right early. Confidentiality clauses, IP assignment for contractors, a clean policy on what may be pasted into public AI tools. Cheap now; expensive to retrofit after someone leaves.
The wider signal
Strip away the personalities and this is a story about how fast the AI industry is consolidating talent, and how little of that movement is governed by anything more robust than an employment contract. The same restlessness shows up elsewhere in the market — in the collapse of switching costs that let cheaper open-weight models take a third of US enterprise usage, and in the pricing pressure that produced GPT-5.6’s three-tier reshuffle. Nothing is locked in for long.
For a small business, the practical conclusion is unglamorous. You will not out-litigate anyone. What you can do is make your know-how hard to walk out of the building with, and make sure the tools you invite in are the ones you chose, on terms you have read. Apple is spending a fortune to relearn that lesson in public. You can learn it on a Tuesday afternoon, for free.
