Google Cloud and Nvidia to Power Apple's Overhauled Siri

What happened

Apple approved plans for a redesigned Siri assistant that will rely partly on Nvidia Blackwell B200 chips hosted through Google Cloud infrastructure. The upgraded system is expected to launch in September 2026 alongside iOS 27, following multiple delays since its initial announcement at WWDC 2024. According to reports from The Information, Apple's proprietary server system, Private Cloud Compute, proved too slow to handle the advanced, trillion-parameter Gemini models required for the assistant. To address this limitation, Apple will route complex cloud queries through Google's data centers while using Nvidia's confidential computing technology to encrypt data and AI models during processing. Simple tasks will continue to be handled directly on devices, while heavier workloads will be routed to external compute partners. Apple is expected to outline its full AI strategy at WWDC 2026, which begins on June 8.

Why it matters

This partnership represents a significant shift in how Google manages AI infrastructure and cloud services. For Google, the arrangement strengthens its position as a critical cloud provider for one of the largest consumer AI launches in years, potentially expanding its enterprise AI footprint across demanding consumer applications. The setup creates a three-way technology stack linking Apple, Google, and Nvidia for consumer AI delivery. Google's Gemini models will power portions of Siri's cloud-based processing, while its data center infrastructure will handle compute workloads that Apple's own servers cannot efficiently process at scale. The collaboration highlights the growing interdependence among major technology platforms in AI infrastructure, even between direct competitors. For Google shareholders, the arrangement could signal expanded cloud services revenue and broader adoption of its AI models in consumer-facing applications.

Bigger picture

The arrangement reflects broader collaboration trends across major technology platforms in AI infrastructure. Apple's decision to rely on external partners represents a pragmatic shift from its traditional strategy of maintaining absolute control over both hardware and software. The move underscores the intense resource demands of scaling generative AI and the competitive pressure facing companies attempting to deliver advanced AI features. For the broader AI sector, the partnership demonstrates how even the most vertically integrated technology companies are finding it necessary to collaborate on infrastructure and compute resources. Nvidia's Blackwell B200 chips are becoming critical infrastructure for large-scale model training and fast inference across multiple platforms. The setup also highlights the importance of privacy-preserving technologies like confidential computing in enabling partnerships between competitors. As companies race to deliver consumer AI capabilities, hybrid architectures combining on-device processing with cloud-based compute are becoming the industry standard.

What to watch

Key signals include Apple's WWDC 2026 presentation on June 8, where the company is expected to outline its full AI strategy and potentially demonstrate upgraded Siri capabilities. The September 2026 launch timeline for iOS 27 and the redesigned Siri will be critical for assessing execution. Investors should monitor whether the hybrid architecture delivers the performance and privacy standards Apple has promised, and whether the arrangement translates into measurable cloud services revenue for Google. Broader adoption of Nvidia's confidential computing technology across other partnerships could signal expanding opportunities for secure cloud-based AI processing. Any updates on Apple's progress in developing its own infrastructure capabilities or changes to its reliance on external partners will also be important indicators of the company's long-term AI strategy.