Apple Nears $1 Billion Yearly Deal With Google to Upgrade Siri

Apple is reportedly finalizing a multi-billion dollar agreement with Google to license a custom Gemini AI model, valued at around $1 billion annually, intended to power a significantly upgraded Siri voice assistant, according to Bloomberg's Mark Gurman. The tailored Gemini model, featuring 1.2 trillion parameters, is designed to operate on Apple's Private Cloud Compute servers, with a planned launch in spring 2026.

This strategic licensing deal is Apple's largest to date in the AI space. After comparing offerings from OpenAI, Anthropic, and Google, Apple opted for Google's Gemini primarily due to favorable pricing rather than performance metrics alone, as detailed by Bloomberg. The 1.2 trillion parameters in this custom model represent an eightfold increase over the 150 billion parameters in Apple’s current cloud-based Intelligence systems.

Apple will retain full control over its infrastructure by deploying Google's model exclusively on its own Private Cloud Compute servers, preserving its privacy-centric architecture. These servers utilize custom Apple silicon and employ end-to-end encryption, ensuring all user data remains secure during processing. Additionally, Apple will continue using its proprietary models for certain Siri functions, establishing a hybrid framework that prevents complete reliance on Google.

Neither Apple nor Google has publicly confirmed the partnership. Bloomberg notes that both companies intend to keep the arrangement confidential, with Apple marketing the enhanced Siri as its own proprietary advancement without highlighting Google's role. This strategy is similar to Samsung’s approach of integrating Google Gemini features into its Galaxy AI while presenting them under its own brand.

Strategic Shift Following Talent Losses

The potential agreement emerges as Apple confronts increased pressure from the departure of key AI talent, including Ke Yang, who previously led Siri’s AI search efforts before joining Meta in October. These departures have fueled concerns about Apple's ability to meet its ambitious March 2026 deadline for the Siri overhaul using only internal resources.

Apple's Private Cloud Compute infrastructure was introduced as a key component of its strategy to manage computationally demanding AI tasks while upholding strong privacy standards. The system relies on custom server nodes equipped with Apple silicon, operating a specialized OS derived from iOS and macOS. Each node’s Secure Enclave contains unique, non-replicable decryption keys, and the platform uses stateless processing to guarantee that no user data is retained after a request is completed.

The custom Gemini model is expected to manage advanced functions such as summarization and intricate planning that go beyond the scope of Apple's on-device models. Currently, Apple Intelligence processes personal data locally on iPhones and Macs using smaller foundation models, while more resource-intensive tasks are routed to cloud-based infrastructure.

Launch Timeline and Market Impact

According to Bloomberg, the redesigned Siri is scheduled for release in spring 2026 as part of iOS 26.4, although these plans remain subject to change. This collaboration positions Apple to better compete against advanced AI assistants from rivals like OpenAI and Anthropic, which have shown stronger capabilities in handling complex reasoning tasks.

The reported $1 billion annual investment underscores the strategic value of AI to Apple's ecosystem. In a related development, General Motors recently announced its intention to integrate Google’s Gemini across its vehicle lineup beginning in 2026, signaling broader adoption of Google's AI technology in consumer platforms.

Apple's move to integrate a third-party model for core Siri functionality represents a shift from its traditional in-house development of critical technologies. The hybrid architecture is designed to combine advanced AI performance with Apple's ongoing commitment to user privacy and data security.