Apple filed suit against OpenAI on Friday in federal court in Northern California, and if you've been watching this industry, you'll understand why it feels less like a shock and more like an inevitability.
The lawsuit names Tang Tan, OpenAI's hardware chief and former Apple VP, alongside Chang Liu, who spent eight years at Apple as a senior systems electrical engineer before moving to OpenAI earlier this year. The allegations are specific. Apple claims Tan was using the hiring process itself as an intelligence-gathering exercise, directing job candidates who were still Apple employees to bring "actual parts" into their OpenAI interviews for "show and tell" sessions. The idea, allegedly, was to extract proprietary information from people who hadn't yet handed in their badges. Liu, separately, is accused of failing to return an Apple-issued laptop after his departure and using it to download confidential technical documents.
OpenAI's response is measured: "We have no interest in other companies' trade secrets. We remain focused on building innovative technology that empowers people everywhere." The courts will work out what actually happened.
The partnership that was always going to fracture
This wasn't some ancient rivalry. As recently as 2024, Apple and OpenAI were announcing a headline partnership. ChatGPT was being baked into the iPhone operating system. It was a big deal, and a clear signal that Apple had decided the fastest way to ship AI features was to use OpenAI's models rather than build everything in-house.
That arrangement made sense as long as OpenAI stayed in its lane. Then OpenAI paid $6.4 billion for IO Products, the hardware startup co-founded by Jony Ive (yes, the same Jony Ive who spent decades designing the iPhone, the Mac, the Apple Watch). OpenAI isn't just a software and API company any more. It's building consumer devices. And that's when partnership became competition.
The comparison people are already drawing is to the Levandowski case from 2018. Anthony Levandowski left Google for Uber, allegedly bringing trade secrets with him. Uber settled for $245 million. Whether this case follows the same arc is something only the courts can answer, but the legal theory is similar: when senior talent moves between AI competitors, IP questions move with them.
What founders and product teams should take from this
There's a practical lesson here that goes beyond the lawsuit itself. The companies supplying your AI stack are not neutral infrastructure. They have their own ambitions. OpenAI has models, brand, and now a serious bet on hardware. Anthropic is building for enterprise. Google has its own devices and ecosystem. None of them are neutral.
When you build your product's core AI features through a single provider's API, you're placing a bet that their priorities and yours stay aligned. That's a reasonable bet to make early on. But it gets riskier as your product matures and as these providers expand into adjacent markets. Today's integration partner is tomorrow's competitor. Apple found that out the hard way.
The approach we take at Dakik when building AI features for clients is to give them a layer they actually control. That means custom agents running on your infrastructure, RAG pipelines built around your proprietary data, and vector search using Qdrant so your embeddings and content sit in a database you own. You still use foundation models for the heavy lifting (that's not changing any time soon), but the part that makes AI genuinely useful for your specific users is yours. When the relationships between big AI companies shift, as they clearly have here, your product keeps working.
We wrote recently about what it means that OpenAI's chief futurist has left the company and where he's headed next. The signals coming from inside these organisations are consistent: the AI stack is consolidating fast, and the companies at the centre of it are competing hard for the same real estate.
Hardware is now a serious part of OpenAI's roadmap. That changes the dynamic for everyone building on top of AI. Whether the lawsuit succeeds, settles quickly, or drags on for years, the underlying tension is real and it isn't going away. The AI stack you're relying on today may look very different in 18 months.
If you're thinking about how to structure your AI features so you're not entirely at the mercy of that, let's talk.