Apple May Use Google Servers for AI Siri

Apple is reportedly considering using Google's cloud infrastructure to store data for its upcoming upgraded, AI-powered Siri. The potential move highlights the immense computational demands of next-gen AI, forcing even tech giants like Apple to rely on rivals for critical cloud services.

This partnership is years in the making, following significant internal struggles at Apple to advance Siri. Efforts to integrate modern AI have been plagued by a complex infrastructure, with engineers reportedly splitting Siri's backend into two parts, complicating development and leading to delays for promised features. Former employees have cited a lack of ambition and risk-taking within the AI/ML group, which was internally dubbed "AIMLess." The move to Google's servers marks a significant shift for Apple, which has long aimed for cloud independence with initiatives like "Project McQueen," a plan to build out its own data center infrastructure and reduce reliance on third-party providers like AWS and Microsoft Azure. This long-term strategy kicked off after Microsoft reportedly couldn't keep up with the growth of Apple's services. At the heart of Google's advantage is its custom-designed hardware. Google's Tensor Processing Units (TPUs) are specifically built for the massive matrix calculations required by large language models like Gemini, offering significant efficiencies in performance and energy consumption compared to more general-purpose GPUs. This specialized infrastructure allows for the high-throughput processing essential for a global-scale service like Siri. To address privacy concerns, Apple is extending its on-device security principles to the cloud with a system called Private Cloud Compute (PCC). This architecture is designed to ensure that any user data processed in the cloud is handled in a "stateless" manner, meaning it is not stored or made accessible to anyone, including Apple, after a request is completed. The collaboration with Google comes after Apple reportedly evaluated models from other top AI firms, including OpenAI and Anthropic. While OpenAI's models are widely adopted and Anthropic's are known for a safety-first approach, Google's Gemini stands out for its deep integration into a vast ecosystem and its native ability to handle multiple data types like text, video, and audio simultaneously. This deal also deepens the existing financial relationship between the two tech giants. Google already pays Apple an estimated $20 billion annually to be the default search engine in the Safari browser. This new arrangement for Siri's AI infrastructure adds another significant financial tie between the two competitors.

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.