Big Tech's Broken Promises

Executive summary

The OpenAI-Apple partnership is fracturing in public — and most Americans saw it coming. A new survey of 182 adults finds that 43% were unsurprised the ChatGPT-Siri deal failed, reflecting a deep-seated cynicism about high-profile tech alliances that is backed up by the facts on the ground: Apple reportedly buried the integration, required users to say "ChatGPT" out loud to trigger it, and then quietly signed a rival $1 billion-per-year deal with Google to power Gemini-based Siri instead.

The fallout matters beyond two tech giants squabbling. OpenAI is eyeing legal action against Apple ahead of its developer conference, a dispute that could reshape how AI companies negotiate distribution deals — and who bears the risk when they go wrong. Meanwhile, 56% of respondents said lawyers are the biggest winners when tech companies fight in court, with only 13% believing consumers come out ahead.

Trust in big tech to honor its commitments is eroding: the 2025 Edelman Trust Barometer puts U.S. confidence in tech companies at 63%, down from 73% a decade ago. For OpenAI, which is targeting a $1 trillion IPO valuation while running at a loss until at least 2030, a billion-dollar subscription pipeline that never materialized is not a minor grievance — it is a strategic emergency.

Context

In mid-May 2026, Reuters broke the news that OpenAI was exploring legal options against Apple, including a potential breach-of-contract notice, over a Siri integration that never delivered the subscriber growth OpenAI had projected. That report landed just days before Apple's annual developer conference — a deliberate signal from OpenAI about the stakes it sees in the dispute.

The partnership dates to 2024, when Apple integrated ChatGPT into Siri and Apple Intelligence as the first — and, until recently, only — third-party AI model in the iOS ecosystem. OpenAI executives called it a transformational distribution opportunity, expecting it to generate billions of dollars a year in new subscriptions by reaching iPhone users who had never actively sought out ChatGPT.

What happened instead, according to Ars Technica reporting based on OpenAI insiders: Apple required users to explicitly say the word "ChatGPT" to invoke the feature, deployed it in a small output window, and provided no meaningful promotional support. OpenAI suspects the deliberate underpromotion may have actually damaged the ChatGPT brand rather than grown it. Renegotiation attempts have stalled.

Then came the pivot. In January 2026, Apple signed a deal with Google — reportedly worth roughly $1 billion per year — to integrate Gemini technology into a redesigned Siri. Google Cloud chief Thomas Kurian publicly confirmed the collaboration. Apple is simultaneously planning to open its AI platform to multiple providers in iOS 27, with users able to select their preferred AI model through Settings. The era of any single AI company holding exclusive real estate inside Apple's ecosystem appears to be ending.

To gauge how the public is processing this dispute, a 182-person survey was conducted in May 2026 — timed to the news cycle — with four questions probing reactions to the breakdown, perceptions of blame, trust in big tech partnership agreements, and beliefs about who wins when tech companies fight in court. The findings reveal a public that is more skeptical than surprised, and a market structure that is shifting faster than most AI partnership contracts anticipated.

Findings

Most People Aren't Shocked — And They Have Good Reason Not to Be

The single strongest signal in the survey is what didn't happen: outrage. When asked how they felt about the OpenAI-Apple breakdown, 43.4% of respondents chose "Not surprised, these deals often fail" — the plurality response by a wide margin. Only 29.1% expressed surprise that the partnership fell apart.

That cynicism is grounded in a well-documented pattern. McKinsey research finds that 62% of technology-weighted deals and mergers fail to meet the financial targets set at inception — a failure rate that reflects how hard it is to translate a promising integration agreement into actual revenue, especially when one partner controls the user interface. The OpenAI-Apple deal fits this template almost precisely: OpenAI had the product, Apple had the distribution, and Apple ultimately decided how prominently — or how obscurely — to feature it.

Reporting from Ars Technica makes the mechanics concrete. Apple required users to explicitly say "ChatGPT" to trigger the feature, used a constrained output window, and ran no meaningful promotional campaign. One OpenAI executive described the original pitch as sounding "amazing" — access to a massive mobile ecosystem — and the reality as something far more limiting. OpenAI now believes the underpromoted integration may have hurt the brand rather than built it.

For the 18.7% of respondents who said they were "concerned about the impact on users," the issue is not just corporate disappointment — it is what a buried AI feature means for the millions of iPhone users who may not realize they have AI tools available at all.

Lawyers Win. Consumers Don't. Most People Know It.

When asked who usually benefits most when tech companies enter legal disputes, 56.0% of respondents picked lawyers and legal firms. Only 13.2% said consumers and users come out ahead. The companies themselves were seen as beneficiaries by 22.5%, while 8.2% said no one really wins.

This is not idle cynicism. Georgetown Law research on asymmetric stakes in antitrust litigation finds that dominant defendants — companies with large financial resources — systematically invest more in litigation than smaller rivals can match, driving plaintiff success rates below what an efficient legal process would produce. The Google-Epic antitrust settlement illustrates the problem: the presiding judge publicly questioned whether the settlement benefited Epic more than other developers, or whether smaller players and consumers saw much practical gain at all.

The survey data contains a telling behavioral pattern: respondents who were surprised the OpenAI-Apple partnership failed were 68% more likely to believe the companies involved — rather than lawyers — benefit most from tech legal fights. The minority who still believe in the system's fairness are also the minority who didn't see the failure coming. Skepticism about outcomes and skepticism about remedies travel together.

Apple's Gemini Pivot Turned a Disappointment Into a Crisis

The partnership's underperformance would be a significant problem for OpenAI under any circumstances. But Apple's January 2026 decision to sign a roughly $1 billion-per-year deal with Google to rebuild Siri around Gemini technology transforms the story from a slow failure into a competitive displacement.

OpenAI projected the Apple channel as a billion-dollar-a-year subscription pipeline. That pipeline is reportedly running far below target. Now the distribution slot OpenAI expected to occupy is being filled — at equivalent cost — by its primary rival. Google Cloud chief Thomas Kurian confirmed the Gemini-Siri collaboration publicly in April 2026, removing any ambiguity about where Apple's AI priorities are heading.

The timing of OpenAI's legal deliberations is not accidental. By surfacing breach-of-contract options before Apple's developer conference — where a Gemini-powered Siri redesign is expected to be announced — OpenAI is attempting to apply maximum reputational and legal pressure at the moment Apple most wants a clean, forward-looking narrative.

For OpenAI's financial position, the stakes are acute. The company crossed $25 billion in annualized revenue by February 2026, a fourfold increase from the prior year, but it is not projected to reach profitability until 2030. Its path to a targeted $1 trillion IPO valuation runs through enterprise contracts — because the consumer subscription pipeline, including the one Apple was supposed to deliver, has not scaled as planned. A billion-dollar channel running at a fraction of projections is not a rounding error for a company that needs to sustain its growth narrative for public markets.

Trust Is Low — and Detail-Oriented People Trust Less

Free-response answers to "How much do you trust big tech companies to honor their partnership agreements?" skewed skeptical, consistent with the broader cynicism the survey captured. The Edelman Trust Barometer puts U.S. confidence in technology companies at 63% — down from 73% a decade ago — and finds only 44% of people globally feel comfortable with businesses deploying AI at all.

A personality-level signal reinforces this pattern: respondents who score higher on the Prism Meticulousness trait — people who tend to be detail-oriented and deliberate — showed a measurably lower likelihood of trusting big tech to honor partnership commitments (r = -0.177). It is the people who read the fine print who trust it least.

The free-response data on what went wrong with the OpenAI-Apple deal reveals a polarized public. A meaningful share offered specific, informed critiques — citing Apple's control over the user interface, the Gemini pivot, and misaligned incentives. A roughly equal share admitted they had little awareness of the partnership at all. That split matters: if accountability for AI partnership failures only reaches already-informed audiences, the pressure on companies to honor their agreements — or for regulators to care — stays weak.

Conclusion

The OpenAI-Apple dispute is arriving at a moment when public trust in big tech's ability to honor its commitments is already running low — and the survey data suggests that distrust is neither irrational nor uninformed. The people most attentive to detail are the most skeptical. The people who followed the news cycle saw the failure coming. And a 56% majority already assumes that whatever legal process follows will benefit attorneys more than anyone who actually uses these products.

The immediate thing to watch is Apple's developer conference. If Apple announces a Gemini-powered Siri redesign without addressing OpenAI's grievances, the legal options OpenAI has been weighing become harder to defer. A formal breach-of-contract notice would mark the first major public legal confrontation between two of the most prominent forces in consumer AI — and would force a reckoning about what partnership agreements in the AI industry actually mean and what remedies exist when they fall short.

Longer term, Apple's reported iOS 27 plan to open its AI platform to multiple providers via user Settings could change the negotiating dynamics entirely. In a multi-model world, no single AI company gets guaranteed distribution — but every AI company gets a fair shot to earn it. That is a better structure for competition, though it offers OpenAI little consolation for the billion-dollar pipeline it expected and never received.