Apple’s WWDC 2026 Siri Overhaul Runs on Google’s Gemini

Apple’s Worldwide Developers Conference (WWDC) 2026 opens today at Apple Park in Cupertino, California, with a keynote starting at 10 a.m. PT that delivers what the company first promised at the same event two years ago: a completely rebuilt Siri, powered by a custom 1.2-trillion-parameter Google Gemini model Apple is licensing for $1 billion a year. The rebuilt assistant launches from the Dynamic Island, functions as a standalone chatbot app, and processes complex queries through a model eight times larger than Apple’s own 150-billion-parameter cloud system.

For most of the past decade, Apple’s pitch was consistent: what happens on your iPhone stays on your iPhone. The company put it on billboards at trade shows. The Gemini deal, jointly announced by Apple and Google on January 12, keeps Apple’s Private Cloud Compute infrastructure in the chain; Google’s custom model handles everything that sits above it.

A New Siri, Starting Today

The redesigned Siri carries a new face from the moment it is invoked. Users say “Siri” or hold the power button; the assistant rises from the Dynamic Island with a dark, glowing animation that matches the “All systems glow” branding Apple chose for the event. Results appear as rich cards covering news, weather, sports, people, and personal data including messages, notes, contacts, and calendar entries.

A standalone Siri app arrives on iOS, iPadOS, and macOS 27, modeled on the chatbot interfaces of ChatGPT and Gemini: a main screen of past conversations, iCloud sync across devices, and an “Ask Siri” entry bar that accepts text, voice, and file attachments including images and PDFs. A second entry point, the “Search or Ask” panel, is accessible by swiping down from the top center of the iPhone screen.

Features expected at the keynote:

• Standalone chatbot app with full conversation history, summaries, and iCloud sync across iPhone, iPad, and Mac
• Dynamic Island anchor with a dark animation replacing the previous white glow border
• “Search or Ask” panel, swipe-accessible from the top center of the screen, with rich card results for news, weather, sports, and personal calendar and message data
• Multi-step personal queries using data from emails, calendars, contacts, and on-screen content, including drafting emails from web and account sources
• Camera app visual intelligence additions, including nutrition label reading and contact identification from photos
• Third-party AI extensions: Gemini, Claude, and ChatGPT expected as user-selectable options alongside the rebuilt native Siri

Internally codenamed “Campo” in pre-event reporting, the system is also expected to introduce “World Knowledge Answers,” a capability that synthesizes responses from multiple sources rather than routing users to Safari. Siri currently redirects complex queries to ChatGPT; the rebuilt architecture is designed to absorb most of that function. Bloomberg’s Mark Gurman described iOS 27 broadly as a “Snow Leopard” release, after Apple’s 2009 macOS update that prioritized stability and codebase cleanup. No new hardware is expected at the event.

The Architecture of the Deal

Three Tiers, One Experience

Apple and Google jointly announced the partnership on January 12. The deal is multi-year; Apple pays roughly $1 billion a year for a purpose-built Gemini model trained at 1.2 trillion parameters. The previous Apple Intelligence arrangement had run on OpenAI’s GPT models, and Gemini outperformed them on Apple’s internal evaluation suite for complex queries, tipping the selection.

The rebuilt Siri operates on three tiers, with the Gemini model at the top of the stack:

Nvidia B200 chips via Google Cloud infrastructure provide the compute for the Gemini tier, per a TechTimes report from June 5. Reporting from March 2026 also noted that Apple had asked Google to explore hosting dedicated Siri servers on Google’s own cloud with custom security configurations, a step beyond the self-managed model Apple described when the deal was announced. Google confirmed publicly it “will not receive Apple user data” through the arrangement. Apple’s contract prevents Google from using Siri queries to train future Gemini models.

Privacy on Paper and in Practice

Gemini’s branding appears nowhere in Siri’s interface. The integration is white-labeled; from a user’s perspective, this is still Siri. CEO Tim Cook told CNBC on January 29: “We’re not changing our privacy rules. We still have the same architecture that we announced before, which is on device plus Private Cloud Compute.”

Apple has built its brand around a values claim: Google is the company that monetizes user data; Apple is the company that does not. Google already pays Apple an estimated $20 billion a year to remain the default search engine in Safari, per reporting during the 2024 Google antitrust trial. The Gemini arrangement adds a new layer to that relationship. An independent analysis presented at an Association for Computing Machinery (ACM) conference in June 2026 confirmed Apple’s three core Private Cloud Compute privacy claims after examining publicly available technical documentation, but that analysis covers what Apple has disclosed, not the potential Google-hosted server infrastructure that the March reporting described.

Two Years of Missed Deadlines

At WWDC 2024, Apple announced Apple Intelligence with demos showing Siri juggling flight data, calendars, and personal contacts to answer queries like “When does Mom’s flight land?” Those features, Personal Context Awareness and In-App Actions, were never shipped. Apple delayed them through 2024, revised the timeline to “the coming year,” then pulled television advertising that had promoted those capabilities to consumers who bought iPhone 16 devices partly on the strength of those promises.

Bloomberg reported in March 2025 that Apple executives described the delay internally as “ugly” and “embarrassing.” Apple’s senior vice president Craig Federighi later acknowledged publicly that the V1 architecture “wasn’t getting us to the quality level that we knew our customers needed.” At WWDC 2025, he told developers the company had a V2 working in-house but would announce no ship date until it met Apple’s standards. Apple scrapped V1 entirely and rebuilt from the foundation.

The June 2024 announcement bundled over 20 AI features into the Apple Intelligence launch, but the advanced Siri capabilities at the pitch’s center had not arrived eight months later. Sixty-nine plaintiffs filed Landsheft v. Apple Inc. in the Northern District of California, arguing Apple had sold iPhone 16 devices on AI promises it could not keep at the time of sale. The case is ongoing.

Evaluating an in-house alternative, Apple concluded it would take several years at current investment levels to build a model competitive with Gemini. Licensing allowed the company to accelerate immediately and redirect its AI budget toward on-device inference and the privacy infrastructure connecting users to the Gemini tier.

Where the Rivals Land

Apple’s choice of Gemini over OpenAI or Anthropic’s Claude shifted the competitive picture beyond Cupertino. Three companies absorbed the implications differently:

• Samsung’s Galaxy AI already runs Google Gemini models, which Samsung had positioned as a differentiator against Apple. Both flagship ecosystems now share the same underlying model, collapsing that distinction.
• Amazon accelerated its integration of Anthropic’s Claude into Alexa after the Apple deal was announced, staking a claim to a distinct AI lineage.
• Microsoft, which invested $13 billion in OpenAI, is doubling down on its in-house MAI (Microsoft AI) models to reduce dependence on any single AI partner, including OpenAI.

ChatGPT integration, which Apple added to Siri in 2024 for text and image generation tasks, remains available for now. iOS 27 is also expected to let users select Gemini, Claude, or ChatGPT as extensions within the standalone Siri app, per Bloomberg’s Mark Gurman reporting from March. That architecture positions Siri as a routing layer across AI providers, with the rebuilt native assistant as the default.

A Second Pipeline for Google

U.S. District Judge Amit Mehta’s September 2025 antitrust remedies order in the Google Search monopoly case explicitly prohibited Google from entering exclusive distribution contracts covering Gemini, alongside Google Search, Chrome, and Google Assistant. The Department of Justice filed an appeal of parts of that ruling in February 2026, arguing the prohibition was designed to prevent Google from replicating its search-default strategy in AI markets.

The Apple-Gemini deal landed directly in that contested space. Mogin Law LLP, in a published antitrust analysis of the arrangement, argued the deal is “precisely what Judge Mehta’s antitrust opinion was designed to prevent,” describing it as a new default AI distribution mechanism with the same structural effect on competition that the search defaults had. Google’s position is that the arrangement is contractually non-exclusive and that Apple chose Gemini on benchmark merit after a competitive evaluation.

The Gemini deal essentially creates a second exclusive pipeline between Apple and Google. If the courts are serious about restoring competition in AI-driven search and assistants, they cannot ignore this arrangement.

Rebecca Haw Allensworth, professor of antitrust law at Vanderbilt University, made that assessment to the Financial Times after the deal’s announcement. The European Commission is reviewing the arrangement under the Digital Markets Act, where both Apple and Google carry gatekeeper designations. Any injunction from the DOJ appeal process could force a restructuring of the Gemini integration before Apple completes the full iOS 27 rollout in the fall.

Two billion Apple devices represent the largest single AI distribution event in history, per chatforest.com analysis of the deal’s implications, allowing Gemini’s usage to expand dramatically without Google winning end-users through its own consumer apps.

Developer betas of iOS 27, iPadOS 27, macOS 27, and all other platform updates go to developers today after the keynote. Public betas follow in July; stable releases are expected in the fall alongside Apple’s next iPhone lineup, the same cadence Apple originally promised when it first demonstrated a rebuilt Siri at Apple Park in June 2024.