Apple's Gemini Siri Bets Big on Google and Nvidia This Fall

For more than a decade, Apple sold a simple promise: the most important parts of its products were built in Cupertino, by Apple, on Apple's own terms. That promise is about to meet its most public exception. According to multiple reports, the company is preparing to launch a dramatically rebuilt Siri this September, and the intelligence behind it will not be Apple's. It will be Google's Gemini, running on Google's cloud, accelerated by Nvidia's chips. After years of stumbles, Apple is outsourcing the brain of its most personal product, and doing so in plain sight.

The arrangement is striking precisely because it cuts against everything Apple usually stands for. As MacRumors noted, citing The Information, the plan "diverges from Apple's usual strategy of attempting to control all the critical ingredients to its products." For a company that designs its own silicon, operating systems, and services, renting someone else's frontier model is a remarkable concession.

The September Launch and What Is Actually Reported

According to reporting summarized by MacRumors and 9to5Mac on June 3 and 4, the overhauled Siri is targeted for a September 2026 launch, aligning with Apple's traditional autumn iPhone and software cycle. It is worth being precise about what is confirmed and what is reported. Google has publicly acknowledged the broad partnership. At Google Cloud Next 2026, Google Cloud chief Thomas Kurian said the company is "collaborating with Apple as their preferred cloud provider to develop the next generation of Apple Foundation Models based on Gemini technology," adding that those models "will now power future Apple Intelligence features including a more personalized Siri coming later this year."

The granular details, the September timing, the specific chips, and the financial structure, come from press reporting rather than official Apple statements, and should be read as reported rather than confirmed. Apple is expected to preview the new assistant during its WWDC keynote, scheduled to begin at 10:00 a.m. Pacific on June 8, 2026, before a full consumer rollout in the fall.

Inside the Gemini Deal and Its Price Tag

The partnership was jointly announced by Apple and Google on January 12, 2026. According to Bloomberg's Mark Gurman, Apple agreed to pay Google roughly $1 billion per year for access to a custom Gemini model built specifically for Siri and Apple Intelligence. That model is reported to carry around 1.2 trillion parameters, roughly eight times the size of the 150-billion-parameter cloud models Apple had been running on its own.

The Financial Times reported that the agreement is structured as a cloud computing contract "which could lead to Apple paying several billion dollars to Google over time," with at least one analyst pegging the lifetime value near $5 billion. Crucially, the FT also reported that OpenAI was not the chosen partner because, according to a person close to the company, OpenAI made "a conscious decision to not become the custom model provider for Apple" in the autumn of 2025, preferring to focus on its own hardware ambitions with former Apple designer Jony Ive.

Why Apple Turned to Google After Siri Stalled

To understand why Apple is writing billion-dollar checks to a rival, you have to revisit a painful year. In 2025, Apple publicly conceded that the ambitious, deeply personalized Siri it had teased would slip well beyond its promised window. The fallout was organizational: Tim Cook moved Siri out from under longtime AI chief John Giannandrea and handed it to Vision Pro creator Mike Rockwell, who now reports to software lead Craig Federighi. Inside Apple, according to Bloomberg, Rockwell had spent years criticizing Siri and pitching executives on a more capable, personalized assistant.

The technical reality was unforgiving. According to The Information, Apple tried to run a modified version of Gemini on its in-house server system but found that it ran too slowly to be usable. When your own infrastructure cannot deliver a frontier experience on a deadline, and your archrival can, the calculus changes. Apple chose to ship something excellent on borrowed rails rather than ship something mediocre on its own.

The Chip and Infrastructure Arrangement

The hardware story is where the deal gets genuinely unusual. According to The Information's reporting, Apple will tap into Google's fleet of Nvidia Blackwell B200 data center chips to handle the heaviest Siri queries. The division of labor is meant to follow Apple's familiar logic: handle as much as possible on the iPhone, iPad, Mac, and other devices, and offload only the most demanding requests to the cloud.

That raises an obvious question about Private Cloud Compute, the Apple-silicon server system the company unveiled two years ago as its privacy-preserving answer to cloud AI. Reporting indicates Private Cloud Compute could not run the giant Gemini model fast enough, and it remains unclear, by 9to5Mac's account, exactly how that system fits into the September launch. For now, the most powerful Siri features appear set to live on Google's machines rather than Apple's.

The Privacy Tradeoff Apple Is Making

Privacy has long been Apple's marketing crown jewel, so processing user requests on a competitor's servers is delicate territory. The reported solution is technical: according to The Information, user data will be encrypted using Nvidia's hardware-based confidential computing, which encrypts information while it is actively being processed on the chip rather than only at rest or in transit. Nvidia describes the feature as preserving "the confidentiality and integrity of AI models" deployed on its GPUs.

The promise, then, is that Google would not be able to read what users ask Siri, and that Apple's privacy guarantees survive the move off Apple silicon. Whether that reassures regulators and privacy-conscious customers is a separate matter. Apple is asking users to trust a chain that now runs through Google's data centers and Nvidia's hardware, a meaningfully longer chain than "it stays on your iPhone."

What It Signals About the AI Power Balance

Step back and the deal reframes the entire competitive map. Google, fighting an antitrust battle over its search dominance, suddenly becomes the indispensable AI supplier to its biggest mobile rival, a position that strengthens Gemini's claim as a frontier model trusted by the most demanding customer in technology. OpenAI, by reportedly walking away, doubled down on becoming a hardware company rather than an infrastructure vendor. Microsoft, the early generative-AI front-runner through its OpenAI alliance, watches a billion-dollar relationship form without it.

For Apple, the stakes are existential in a quieter way. The iPhone's next decade likely depends on whether the assistant living inside it feels genuinely intelligent. If the Gemini-powered Siri lands well in September, Apple buys itself time to rebuild its own models behind the scenes while customers enjoy a competitive experience today. If it stumbles, Apple will have paid billions to discover that even rented brilliance cannot fix a product it could not build itself. Either way, the September keynote will be the moment we learn whether Apple's most pragmatic AI gamble was worth the price of admitting it needed help.