Apple servers are unable to handle the Siri service, so the company is turning to Google for hardware support

Apple is forced to seek external solutions for AI

The company Apple discovered that its own cloud infrastructure, Private Cloud Compute, cannot handle the demands of processing artificial intelligence queries. Specialized servers that are supposed to serve Siri models and other AI applications run at only about 10 % of their available capacity on average, and some equipment sits idle and even dusts in warehouses.

Why Apple is turning to Google

According to materials from 9to5Mac (citing The Information), Apple approached Google with a request to host its new Siri models in the company’s data centers. The reason is that within Apple there is fragmentation of systems, and different teams use isolated technology stacks, which makes flexible resource reallocation difficult. At the same time, the finance department expresses dissatisfaction with rising costs for duplicate infrastructure and is not yet ready to invest billions in rebuilding it.

Technical limitations

The Private Cloud Compute platform also lags behind in technical specifications: modified M2 Ultra processors cannot provide the necessary performance to run large language models (LLMs). Software updates require significant effort and time. Meanwhile, low user interest in Apple Intelligence features increases doubts about the feasibility of further deploying proprietary data centers.

Conclusion

Thus, Apple is forced to rely on Google’s experience, which has already proven its effectiveness in large-scale deployment of servers for LLMs thanks to the Gemini project. This solution will allow the company to respond quickly to AI demands and avoid additional costs for its own infrastructure.