Product1 distinct publisher3 min readPublished
Apple's July suit says departing engineers carried hardware designs to OpenAI. OpenAI's answer is that Apple's own habit of pushing work into personal iCloud accounts blurred the boundary it now has to prove was crossed.
The Product Desk · Product desk

Compiled by The Product DeskSomething wrong?How this is made
Follow any of these and your For You feed starts watching them — no settings page required.
product
Greg Brockman, not Sam Altman, is the one running OpenAI day to day1 distinct publisher
invest
OpenAI hands Cursor developers 76 days to migrate off its models2 distinct publishers
build
Grok 4.6 lands in Copilot two days after launch, and the model picker becomes a procurement problem1 distinct publisher
invest
OpenAI kept the nonprofit and deleted the cap: Brockman's stake and the missing guardrail1 distinct publisher
Someone photographs a whiteboard on a managed iPhone and the image lands in a personal library, because that is the account signed in for Photos. Users tend to keep work in whichever account syncs without a support ticket. Teams tend to assume the opposite: that work stays in the corporate tenant while the personal account holds nothing but holiday pictures. The BYOD pitch is usually about meeting people where they already work. The thing being done is moving the account of record outside the company.
The two practices OpenAI attributes to Apple push information both ways at once, work product out into personal iCloud accounts and personal messages in to corporate monitoring [4][15]. The inbound half is the awkward one for an employer. Monitoring produces records, records are discoverable, and what they document is an environment where the line between company and personal was crossed by configuration long before anyone resigned.
Trade-secret law generally asks whether the owner took reasonable measures to protect the information [8]. TNW's read is that Apple's policies on devices, accounts and data could become an important part of the case rather than background to it [9]. That is OpenAI's August theory stated plainly in its motion to dismiss: "Apple cannot transform the foreseeable consequences of its own systems and policies into evidence of theft" [5]. California leaves Apple little else to work with, since non-compete agreements are largely unenforceable there [10], which means preventing the transfer of confidential information has to run through trade-secret law rather than through limits on where a former employee can work [17].
The documented moves so far are the July complaint, the August motion to dismiss, Monday's filing, and a public post arguing Apple is getting this wrong, three of them court filings [14][7]. The fourth is aimed at the hardware engineers deciding whether to return a recruiter's call, and OpenAI has been hiring exactly those people for the device family Greg Brockman has confirmed it is building [11]. TNW also reports the case is already creating delays in that programme [13], so the injunction fight matters to OpenAI's schedule whatever the merits turn out to be.
For whoever owns the device profile, the useful exercise is smaller than a policy rewrite. Take the five artifacts you would least like to see at a competitor and name, for each, the account of record and the person who can enumerate every copy. Then plot two axes: does work product live only in accounts the company controls, and does personal content end up in logs the company keeps. The single quadrant where you can later prove what left is yes on the first and no on the second. Convenience puts you in one of the other three, and the invoice arrives as a discovery request, years after the admin who approved the setup has moved on.
Ranked by verification strength, evidence, and original report placement.
In a filing lodged late Monday in the US District Court in San Jose, OpenAI said: "This dispute is a mess of Apple's own making, and it is trying to blame everyone else," according to Reuters as reported by TNW.
Apple's lawsuit, brought in July, accused departing employees of taking hardware designs to OpenAI, including prototypes that Apple alleged were shown during job interviews.
OpenAI argues that Apple encouraged employees to use personal iCloud accounts for work and monitored personal messages on company devices, creating a boundary between corporate and personal information that was never particularly clear.
In its August motion to dismiss, OpenAI argued that "Apple cannot transform the foreseeable consequences of its own systems and policies into evidence of theft"; Monday's filing follows the same argument with considerably stronger language.
OpenAI is not necessarily disputing that work-related information may have ended up in personal accounts; it is challenging what that movement of information means.
Distinct publishers with included, body-backed reporting in this cluster.
1 article · September 1, 2026
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Documented docket, one-sided substance
The skeleton is verifiable: a July complaint, an August motion with a quotable line, a Monday filing, an injunction request, a public post. The flesh is not. The Next Web reaches the Monday filing through Reuters rather than the document, and every statement about how Apple actually handled devices and accounts arrives as OpenAI's own framing of them, with no Apple reply and no independent look at the policies.
Nothing shipped to count
There is no product in the world here. OpenAI is hiring engineers for a device family Greg Brockman has confirmed exists; no units, customers, dates or deployments appear anywhere, and the one claim about the programme's trajectory is about delay rather than uptake. Counting adoption from this material would mean inventing it.
Framing slightly ahead of the paper
The restraint is real — The Next Web says outright that a striking complaint is not proof of misappropriation, and that OpenAI's argument may not survive legally. What runs ahead of the record is the surrounding scenery: an AI-native device cast as the first credible challenge to the iPhone's role, and a hardware programme said to be slipping on the strength of the outlet's own prior story. Small gap, and it sits in the strategic framing rather than the legal reporting.
Almost everything traces to a litigant
Read the sourcing and it is two adversaries plus a wire service. The sharpened language in Monday's filing is a tactical choice, the "Apple is getting this wrong" post is advocacy published outside the court, and Apple's allegations were drafted to persuade a judge. Add the setting — California bars the non-compete Apple might otherwise have used, so trade secret law is the only lever — and both sides have a strong reason to describe Apple's data hygiene in whichever direction helps them. Apple's answer to the latest filing is simply absent.
Confident about the exchange, not the facts
What we would defend: these filings happened, in this order, with that quote in them. What we would not: any characterisation of Apple's device and account policies, which no one outside OpenAI has described here. One publisher, one side, one relayed quote. The docket is public, so this is a story whose confidence should rise quickly — or not at all — once Apple answers.