OpenAI Prepares Legal Action Against Apple Over Siri's Broken ChatGPT Promise
OpenAI is exploring breach-of-contract claims against Apple after the 2024 Siri integration failed to deliver billions in anticipated subscription revenue. As Apple quietly adds Anthropic and Google as rival AI providers, the flagship consumer AI partnership of last decade is unraveling — and it may do so in court.
Two years after Apple CEO Tim Cook and Sam Altman took a joint bow on stage at WWDC 2024 to announce a landmark AI partnership, the relationship has curdled into one of Silicon Valley’s most consequential business disputes. OpenAI has quietly enlisted an outside law firm to evaluate its legal options against Apple, according to reporting by Bloomberg and confirmed by multiple sources at both companies — a move that signals just how badly the once-celebrated Siri-ChatGPT integration has disappointed its architects.
A Deal Built on Subscription Dreams
When Apple announced it would weave ChatGPT into iOS 18, iPadOS 18, and macOS Sequoia in June 2024, the financial logic seemed elegant: Apple would gain cutting-edge AI without having to build a frontier model; OpenAI would gain access to hundreds of millions of devices and, critically, a new subscription funnel worth potentially billions of dollars annually. No money changed hands up front. Instead, OpenAI negotiated a revenue-sharing arrangement where Apple would pass along a portion of ChatGPT subscriptions driven through the integration.
Inside OpenAI, executives believed the deal could eventually generate “billions of dollars per year in subscriptions” as iPhone users organically upgraded to ChatGPT Plus after experiencing the assistant’s capabilities through Siri. Those projections, according to people familiar with the discussions, “haven’t come close to happening.”
The problem, OpenAI alleges, is not user apathy but design choices. The ChatGPT integration has been buried behind multiple taps, requires users to explicitly invoke the “ChatGPT” name within Siri’s flow, and offers a narrower feature set than the standalone app. Revenue from the arrangement is described internally as “financially disappointing” and nowhere near what either party originally modeled. OpenAI has grown increasingly frustrated, believing Apple has not done enough to surface the integration or drive upgrades.
Apple’s Calculated Hedge
Apple’s motive for the underwhelming rollout may not be negligence. Over the past year, Apple has been quietly testing integrations with both Anthropic’s Claude and Google’s Gemini as alternatives to ChatGPT within Siri’s forthcoming “any model” architecture. The strategic message is clear: Apple intends to position itself as an AI-agnostic conduit rather than a long-term partner to any single provider. Under this model, Apple captures value by owning the user relationship while AI companies compete on the underlying intelligence layer.
This shift represents a direct threat to OpenAI’s distribution thesis. When Altman agreed to the deal, exclusivity wasn’t formal, but preferential placement was implied. Now ChatGPT faces the prospect of being just one of three competing AI providers inside Siri — alongside Anthropic and Google, two companies with their own billion-dollar consumer and enterprise ambitions.
The dynamic mirrors, in some ways, how browser vendors eventually shifted from default search deals to multi-engine auctions. For OpenAI, losing that premium position could have lasting consequences for user acquisition and subscription conversion.
The Legal Calculus
OpenAI’s legal team has been working through options that range from a formal breach-of-contract notice — a softer opening that often precedes negotiation — to full-scale litigation. Lawyers are analyzing what Apple contractually committed to in terms of feature prominence, marketing support, and subscription targets, though the precise terms of the 2024 deal have not been made public.
For now, any legal action is on hold. OpenAI’s trial with Elon Musk, which wrapped closing arguments earlier this week, consumed significant executive bandwidth and legal resources. Sources say the company would prefer to resolve the Apple friction through direct negotiation rather than a courtroom battle, particularly given the complex commercial interdependencies that still tie the two companies together — ChatGPT remains available via Siri, and Apple still benefits from the partnership’s AI-forward marketing narrative.
But the mere fact that OpenAI has engaged outside counsel signals that patience is wearing thin. The company is also dealing with a broader strategic urgency: rivals including Anthropic and Google have surged in both capability and distribution, and a failed Apple channel is a hole OpenAI cannot afford to ignore.
Not the First Partner to Feel Burned
OpenAI has a pattern of high-profile partnerships that generate early excitement and later friction. Microsoft, despite being OpenAI’s most significant backer with a $13 billion investment, has seen its relationship evolve into something more competitive than collaborative, with Redmond building its own models and OpenAI competing directly for enterprise customers. Salesforce and other API partners have quietly reduced their dependence on OpenAI models as open-source alternatives mature.
The Apple dispute, however, is different in scale and symbolism. The Siri partnership was supposed to prove that OpenAI could be the AI layer for consumer hardware — the ambient intelligence built into billions of devices. If that thesis fails, it raises fundamental questions about whether ChatGPT’s consumer growth story is as durable as the company’s $300+ billion valuation implies.
What Comes Next
Apple is expected to announce the next iteration of Apple Intelligence at WWDC 2026, likely in June. How prominently ChatGPT features — or whether the Anthropic and Google integrations are elevated alongside it — will be read as a signal of where the relationship stands. Industry analysts expect Apple to formalize its multi-model strategy, which would implicitly reduce OpenAI’s privileged status even without a legal resolution.
For the AI industry broadly, the dispute serves as a cautionary tale about distribution deals built on conversion assumptions. Placing an AI model inside a platform does not guarantee users will upgrade — and platform owners retain enormous discretion over how deeply they surface any particular integration. The age of landmark AI partnership announcements may be giving way to an era of harder commercial terms, clearer performance commitments, and, when necessary, litigation.
OpenAI vs. Apple may well become the defining business dispute of the AI era’s first chapter.