OpenAI Feels 'Burned' by Apple's ChatGPT Integration: Legal Threat, Missed Billions, and a Strained Two-Year Partnership

TL;DR: OpenAI has retained outside legal counsel and is weighing a formal breach-of-contract notice against Apple over the 2024 ChatGPT-Siri partnership. OpenAI expected the deal to generate billions of dollars per year in subscription revenue; Apple’s implementation reportedly buried the integration behind friction, including a requirement that users explicitly invoke the word “ChatGPT.” Apple is meanwhile moving Siri’s underlying intelligence to Google Gemini under a $1B-per-year deal. The dispute is now tangled with a parallel talent war: at least 12 senior OpenAI staff previously built Siri, Vision Pro, and Apple’s audio team — and Apple’s Ruoming Pang, former head of foundation models, just left to lead OpenAI’s “Device” team. To stem the bleeding, Apple has begun issuing out-of-cycle $200K–$400K retention RSUs to its iPhone product design staff. The fight has expanded from a software deal into a people-and-platforms conflict.

1. The deal that disappointed #

In June 2024, Apple and OpenAI announced one of the most-hyped AI partnerships in consumer technology. ChatGPT would be woven into Siri, Writing Tools, Image Playground, and Visual Intelligence across iOS, iPadOS, and macOS. Apple would take a cut of ChatGPT subscription revenue generated through its platforms. OpenAI would gain distribution to 2.5 billion active Apple devices.

Two years later, the partnership is collapsing. On May 14, 2026, Bloomberg reported that OpenAI has enlisted an outside law firm to explore options including a formal breach-of-contract notice to Apple. No lawsuit has been filed. OpenAI has indicated it would prefer to resolve the matter privately.

The core grievance is distribution. OpenAI projected the Apple channel as a billion-dollar-a-year subscription pipeline, with direct sign-ups through Apple’s Settings app converting a slice of more than two billion active Apple devices into paying ChatGPT users. The actual run-rate sits well below that projection.

“We have done everything from a product perspective,” an OpenAI executive told Bloomberg. “They have not, and worse, they haven’t even made an honest effort.”

2. How Apple buried the integration #

According to Bloomberg, OpenAI’s frustration is rooted in specific Apple design choices that suppressed the integration’s reach:

Explicit invocation required. Apple users who want to summon ChatGPT through Siri must specifically say or type the word “ChatGPT” when issuing a command. The friction makes the feature harder to discover and use.
Small, limited response windows. When ChatGPT responds, it does so inside a constrained window with limited information. Users can easily ignore the result rather than engage with it.
No prominent placement. The integration was not promoted meaningfully across Apple’s marketing, retail, or onboarding flows. Few iPhone owners ever learned the integration existed.
Limited functionality. Apple’s implementation exposes fewer features than the standalone ChatGPT app. Persistent memory, custom GPTs, advanced voice tools, and direct subscription management are not available through Apple’s built-in flows.

OpenAI’s internal studies found that users overwhelmingly prefer the standalone ChatGPT app over Apple’s built-in integrations. The same hardware, the same user, but a measurably worse experience through Apple.

3. The Safari deal that wasn’t #

Internally, the OpenAI deal was framed as a potential counterpart to Apple’s enormously profitable Google Search deal in Safari. That arrangement generates an estimated $20 billion per year for Apple. OpenAI executives believed a similar pattern of distribution would deliver recurring subscription revenue in the billions annually.

The comparison shaped OpenAI’s commercial expectations. Apple never directly promised a Safari-scale outcome, but the parallel was drawn explicitly when the deal was announced. An OpenAI executive told Bloomberg that Apple did not fully explain how the integration would work, so OpenAI took a “leap of faith” it now appears to regret.

“When we heard about this opportunity, it sounded amazing: being able to acquire a giant number of customers and have distribution in such a big mobile ecosystem,” the executive said.

The partnership was non-exclusive from day one. OpenAI’s bet rested on Apple’s installed base, paired with Apple’s marketing reach, making ChatGPT the default chatbot for hundreds of millions of users. That bet did not pay off.

4. Apple’s pivot to Google Gemini #

While OpenAI’s concerns mounted, Apple was already lining up a different model partner. In January 2026, Apple signed a multi-year deal with Google to use Gemini models as the foundation for its own AI capabilities. The custom 1.2-trillion-parameter Gemini model will power the next generation of Siri, expected to ship with iOS 27 in September 2026.

The Gemini deal is reported to cost Apple roughly $1 billion per year. That figure has not been publicly confirmed by either company. Apple’s next generation of foundation models is expected to be built on Gemini and on Google’s cloud infrastructure, deepening a dependency OpenAI once hoped to claim for ChatGPT.

The financial scale of the Gemini arrangement shifts the highest-value Siri queries — complex multi-step requests, summarization, planning — away from the OpenAI stack entirely.

5. The Jony Ive factor #

Behind the scenes, the OpenAI-Apple relationship was further strained by OpenAI’s 2025 acquisition of io, the AI hardware startup founded by Jony Ive, Apple’s former chief design officer. The deal positions OpenAI as a potential competitor to Apple’s iPhone hardware business.

Bloomberg reported that Apple was “rankled” by the Ive partnership. Apple’s internal motivation to promote ChatGPT through its ecosystem reportedly declined after learning of OpenAI’s hardware ambitions. Reuters and Bloomberg both cited sources describing a loss of motivation inside Apple to help supercharge ChatGPT integration.

The OpenAI executive told Bloomberg that OpenAI’s potential legal action has nothing to do with Apple’s other AI partnerships, emphasizing that the deal was never intended to be exclusive.

6. The talent war: Apple’s brain drain to OpenAI #

The Jony Ive acquisition is the visible part of a much larger shift. Behind the OpenAI hardware push is a deliberate, multi-year recruiting campaign targeting Apple’s most senior engineers and product designers — a campaign Apple is now visibly losing.

Ruoming Pang, head of foundation models at Apple, has left to lead OpenAI’s new “Device” team. Pang spent more than five years at Apple building the on-device AI models that powered Apple Intelligence and Siri. His move, first reported by the Financial Times, is the highest-profile AI hire by OpenAI from Apple to date. At OpenAI, he will work on AI agents capable of interacting with smartphones and PCs — a direct overlap with the Jony Ive hardware project.

Pang is not alone. According to a public LinkedIn tally, at least 12 current OpenAI staff previously held senior roles at Apple, with most of the moves happening in the last 24 months. The list includes:

Engineering Program Management Leader, Vision Pro and Future Platforms (Feb 2026)
Manager, Mac, Audio and Home (Feb 2026)
Staff Software Engineer, Vision Product Group (Nov 2025)
Senior Manager, Product, Siri AIML (Aug 2025)
Embedded Software Engineer, Vision Product Group (Jun 2025)
Acoustic Transducer Engineer (May 2025)
Senior Engineering Manager / Tech Lead, Acoustics ML (May 2025)
Vice President, iPhone and Apple Watch Product Design, Acoustics, Materials and Interconnects (Feb 2024)

The pattern is unmistakable. OpenAI is hiring from Apple’s audio, voice, and on-device ML teams — the exact groups that built Siri, Vision Pro, and the HomePod. The hires are not random; they line up with OpenAI’s stated product direction: a voice-first, on-device AI companion.

A Wall Street Journal study of LinkedIn profiles in late 2025 found that dozens of Apple engineers and designers had moved to OpenAI and Meta in the preceding months. The departures spanned audio engineering, watch design, and robotics. The PYMNTS summary noted the moves were happening “at a time when competitors are working to challenge the market share held by the iPhone.”

6.1. Apple’s defensive play: golden handcuffs #

In March 2026, Apple broke its own compensation norms. The company began issuing out-of-cycle retention bonuses to its iPhone Product Design team — RSUs valued between $200,000 and $400,000 per person, vesting over four years. The structure is a textbook golden handcuffs arrangement: to realize the full value, designers must stay at their desks in Cupertino until at least 2030.

Apple occasionally issues retention packages, but doing so outside the standard annual review cycle is a rare admission of vulnerability. The timing lines up exactly with the wave of senior OpenAI hires from Apple’s hardware and audio teams. The message: Cupertino has noticed the departures and is willing to spend to slow them.

The talent war and the OpenAI legal fight are not separate stories. They are the same story viewed from two angles. The ChatGPT-Siri deal failed commercially; the people who were supposed to make it work have already left. OpenAI is now rebuilding the team that Apple assembled, on OpenAI’s payroll, with OpenAI’s product roadmap.

7. The legal landscape #

OpenAI’s options are constrained by what is already on the docket. The company is in the middle of its trial with Elon Musk, a co-founder whose suit alleges that OpenAI abandoned its nonprofit founding mission. Any major escalation against Apple is likely to wait until that proceeding wraps.

Musk’s separate antitrust suit against both Apple and OpenAI added a layer of public hostility around the same partnership. Musk alleged that the deal integrating ChatGPT with Apple’s products violated antitrust and propped up OpenAI’s effort to dominate the chatbot market and Apple the smartphone market.

A May 2026 ruling in Musk’s case added a new wrinkle to the OpenAI-Apple dispute. Magistrate judge Hal Ray Jr. denied Musk’s request to see Tim Cook’s internal messages discussing the deal but ordered Apple to share documents by mid-June from Craig Federighi, Senior Vice President of Software Engineering. The judge noted that Federighi “made high-level, strategic decisions about the Apple-OpenAI Agreement” and “may have unique relevant evidence not already produced relating to Apple’s integration of OpenAI into Apple Intelligence.” Apple was also ordered to provide any documents referring to “potential exclusivity clauses of the artificial intelligence provider for Apple products.”

The ruling matters because it forces Apple to disclose internal decision-making about the partnership at the same time OpenAI is preparing its own breach-of-contract case. If Apple promoted ChatGPT weakly to preserve exclusivity leverage with Google Gemini, that could be precisely the evidence both Musk and OpenAI are looking for.

OpenAI’s outside firm is examining whether to send a formal breach-of-contract letter to Apple’s general counsel. The letter would put the alleged 2024-integration breach on Apple’s desk in writing, as a precursor to negotiation or a future lawsuit. Any larger escalation waits behind the Musk verdict.

7. What Apple did next #

iOS 27, expected to be unveiled at WWDC 2026 on June 8, will introduce a system called Extensions that allows users to install rival AI chatbots from the App Store and route Siri queries, writing tasks, and image generation through whichever model they choose. Apple is testing integrations with both Anthropic’s Claude and Google’s Gemini.

The strategic message: AI in iOS becomes a layer, not a single product. ChatGPT, Claude, and Gemini will all be options inside Siri and Apple Intelligence. None of them will be the default. Apple is building an AI architecture designed to reduce dependence on any single provider.

For OpenAI, the new Extensions system might even benefit ChatGPT by giving it more prominent placement through a model-picker interface. But the broader trajectory is clear: OpenAI is moving from privileged partner to one option among several.

8. What the deal cost OpenAI #

The partnership’s deterioration is not just a financial story. It is a lesson in the limits of distribution without control. Being inside the iPhone sounded like a growth engine. In practice, it meant accepting Apple’s design choices, Apple’s revenue terms, and Apple’s willingness — or unwillingness — to promote the product.

The same camera, the same hardware, the same billion users — but the user lost choice in how to invoke AI. Apple didn’t improve ChatGPT. Apple constrained it. The same logic that drove Apple’s walled-garden approach to the App Store, iMessage, and the Safari search deal has now been applied to AI.

For Apple, the lesson cuts the other way. The $250 million class-action settlement over delayed Siri features, the Google Gemini dependency, the OpenAI dispute — each one is a forced compromise. The company that built its reputation on vertical integration has been forced, repeatedly, to depend on others for the capability its customers expect.

9. Bottom line #

Three things to watch:

The breach-of-contract letter. If OpenAI’s outside counsel sends a formal notice, it will put the dispute in writing and force Apple to respond publicly.
WWDC 2026 on June 8. Apple’s iOS 27 keynote will show whether the Extensions system addresses OpenAI’s complaints by giving ChatGPT more prominent placement — or accelerates the shift away from OpenAI.
The Musk trial. OpenAI’s ability to escalate against Apple is constrained until the Musk case concludes. A Musk loss in the antitrust suit would also weaken OpenAI’s hand on its own claims.

The deal that once symbolised mutual need now illustrates why the two largest ambitions in consumer technology — owning the device and owning the intelligence — may be fundamentally incompatible as long-term partnerships.