Apple plans to move Siri workloads to Google Cloud amid internal infrastructure issues

Apple is facing significant efficiency issues with its Private Cloud Compute (PCC) platform.

Current AI-dedicated capacity is reportedly utilized at only around 10%, while a substantial portion of purchased hardware remains idle in warehouses.

According to The Information and 9to5Mac, the core problem stems from internal fragmentation. Teams within Apple operate on isolated technology stacks, making it difficult to flexibly redistribute resources across projects. This is compounded by pressure from the finance department, which is dissatisfied with rising costs tied to duplicated infrastructure and is not yet ready to allocate billions for a large-scale overhaul.

The technical side is also under strain. Modified M2 Ultra chips are reportedly falling short for large language model workloads, and software update cycles on Apple’s in-house servers are proving too resource-intensive.

As a result, the company is acknowledging a technological gap and may rely on a competitor’s infrastructure to ensure stable performance of its voice assistant.