Image: Apple
Apple has revealed that Siri AI’s cloud-based requests come with real running costs, with the company now factoring those expenses into the economics of its newest devices and Services business.
During Apple’s latest earnings call, CFO Kevan Parekh said requests handled directly on a device are highly efficient and carry “no real costs” for Apple, while requests that rely on server-side processing create inference costs.
Apple says it is still too early to know how much Siri AI usage will ultimately happen on-device compared with Private Cloud Compute. As Siri AI expands to more users and languages, the balance between those two types of processing could become increasingly important to Apple’s operating costs.
The company is already accounting for the impact. Parekh said Siri AI’s associated cloud costs will flow through Services cost of sales, while Apple will also defer a portion of hardware revenue and recognize it over the life of devices that include the service.
Those costs are one factor behind Apple’s forecast for lower gross margins during the holiday quarter. Higher memory prices remain the biggest pressure, but Apple specifically included the ramp of Siri AI when forecasting gross margin of between 45% and 46%.
Siri AI has expanded quickly since its initial launch. iOS 27.2 expanded Siri AI to French, Japanese, Korean, Portuguese and Spanish, while Apple continues to develop broader Apple Intelligence availability across its platforms.
Apple is also preparing to monetize heavier use of some server-side AI features. The company has confirmed that certain Siri AI and intelligent photo-editing features will eventually have usage limits, with customers able to pay for expanded access through iCloud+ tiers.
That gives Apple two competing incentives: keep more AI processing on-device to reduce cloud costs, while using server-side features as a potential new source of Services revenue. Apple says AI can also help drive hardware upgrades by giving customers another reason to move to newer devices with the latest capabilities.