Apple asks Google to host Siri as its own AI servers sit idle in warehouses

Apple Turns to Google as Private Cloud Compute Struggles

Apple has asked Google to investigate setting up servers for a new version of Siri powered by Google Gemini, marking a significant shift in Apple's AI cloud strategy1

1

. The move comes as Apple's own Private Cloud Compute infrastructure operates at only 10% capacity on average, with some already-manufactured AI servers sitting dormant on warehouse shelves2

2

. This underutilized cloud infrastructure highlights deeper challenges within Apple's approach to scaling its artificial intelligence capabilities.

The discussions between Apple and Google center on the possibility of deploying servers inside Apple's own data centers or having Google run Siri directly on its infrastructure5

5

. While Apple had already announced in January that Google's Gemini AI models would help power the upgraded Siri, this latest development suggests Apple may lean even more heavily on Google to catch up in the AI race1

1

.

Technical Limitations Expose Infrastructure Weaknesses

Private Cloud Compute faces multiple technical hurdles that make it unsuitable for hosting advanced generative AI capabilities. The system is described as underpowered, with chips believed to comprise mostly M2 Ultra processors that are not powerful enough to run frontier models like Gemini2

2

. These chips were designed for consumer devices and are not optimized for AI workloads, leaving them ill-equipped to handle the demands of large language models3

3

.

Updating the software on Private Cloud Compute is reportedly trickier and takes more time compared to other server systems2

2

. The initial crop of Apple Intelligence features hasn't been used as much as Apple expected, casting the entire Private Cloud Compute buildout in a negative light internally2

2

.

Fragmented Cloud Technology Stack Creates Inefficiencies

Apple's cloud infrastructure is highly fragmented, with different teams running technologies independently rather than drawing from one centralized pool2

2

. This siloed approach leads to inefficiencies where parts of the company's technology stack sit idling while other departments would use the available server capacity if they could access it. The Apple finance team has reportedly been frustrated about the costs of this duplicate infrastructure2

2

.

According to Igor Naverniouk, a former Apple engineer who worked on the new Siri, "Apple and Google have two very different engineering cultures. At Google, most things are centralized. Everybody is using the same supercomputer. At Apple, technology choices are siloed"4

4

. Apple has reportedly attempted to unify everything several times since at least 2013, but those projects have stalled repeatedly over the last decade2

2

.

Privacy Standards Remain Non-Negotiable

Despite the shift to host Siri on Google servers, Tim Cook has emphasized that Apple will not relax its privacy systems. "We're not changing our privacy rules," Cook stated. "We still have the same architecture that we announced before, which is on device plus Private Cloud Compute"4

4

. For years, Apple banned its AI engineers from using Google's cloud technology because of privacy concerns, with software chief Craig Federighi repeatedly vetoing Google Cloud as an option3

3

. In 2023, Google made changes to its security systems that satisfied Apple privacy standards, allowing Apple to adopt Google's cloud infrastructure for artificial intelligence3

3

.

Apple currently uses Private Cloud Compute for stateless data processing with encryption, where user data remains on the server only until the response is returned to the device, with no retention via logging or debugging5

5

. Apple continues to offer security experts the opportunity to test its privacy systems4

4

.

Years of Under-Investment Catch Up With Apple

Apple has suffered from a cultural reluctance to bolster its cloud infrastructure for years, leading to the departure of key cloud experts like Patrick Gates, who pioneered bringing Apple chips to data centers3

3

. The company still strongly focuses on hardware devices and consumer features rather than their supporting cloud technology, despite the growth of services3

3

.

When Apple realized it needed to use the cloud to support its AI efforts, its internal AI infrastructure was "beginning to decay," with the company in the process of decommissioning old Nvidia-powered servers3

3

. Combined with financial pressure, this led Apple to increasingly turn to third-party providers like Amazon. Even back in 2021, Apple was Google's largest corporate cloud customer4

4

.

Preparing for Surge in Demand

Apple wants to be prepared for a potential surge in AI use on its devices when the more powerful, Gemini-based version of Siri debuts later this year3

3

. While current Apple Intelligence usage remains low, Apple expects much higher demand for the new Siri chatbot features whenever they finally land2

2

. Google already has extensive experience with mass LLM server buildouts thanks to Gemini, making it a natural partner for scalability2

2

. The company already relies on Google's cloud for some iCloud features, including cloud storage2

2

. The ever-changing landscape of AI may force Apple leadership to invest more heavily in-house going forward, but implementing those changes represents a much longer-term trajectory2

2

.