- Cook said heavy AI users could get “upgrade possibilities” through iCloud+, but described Apple’s plan for covering compute costs as incomplete. [1] - Apple has not announced a separate Siri subscription, AI-specific prices, numerical usage caps or the additional access included with each iCloud+ tier. [1][2] - Siri AI remains scheduled as an English-language beta later in 2026; Apple has not provided a specific consumer release date. [2] - Apple’s five-model architecture combines on-device processing, Apple-silicon servers and a high-end model running on Nvidia GPUs in Google Cloud.
[3] Apple expects to let customers buy higher artificial-intelligence usage through iCloud+, but it has not decided exactly how much additional access subscribers will receive or what they will pay, CEO Tim Cook said during the company’s fiscal third-quarter earnings call Thursday.[1]
Cook’s comments were narrower than a declaration that Apple will put its rebuilt Siri behind a subscription. Apple has separately disclosed that some server-intensive Apple Intelligence features will have daily limits and that most iCloud+ plans will provide increased access. It has not announced a standalone Siri plan or new AI-specific prices.[1][2]
An iCloud+ upgrade path, not a completed price plan #
Asked about the computing expense created by Apple’s expanded AI services, Cook said it was “early going” and cautioned that Apple did not have a complete plan. He added that people who use the services extensively could receive “some kind of upgrade possibilities on iCloud+,” allowing them to move higher in Apple’s subscription lineup.[1]
Apple disclosed part of that structure at its Worldwide Developers Conference in June. Some Apple Intelligence features, including image generation, will carry daily limits because they depend on powerful server models. Apple said most iCloud+ subscriptions would provide increased access, but it did not publish the limits, identify the qualifying plans or announce AI-specific pricing.[2]
Apple has not said that the disclosed daily limits apply to every Siri AI function. The assistant spans local and cloud processing: some work runs on the device, while more demanding requests can be routed to models operating through Private Cloud Compute.[2][3]
Cook’s answer identifies iCloud+ as Apple’s intended route for selling heavier AI usage. It does not establish that ordinary Siri commands, personal-context searches or other core assistant functions will require payment. Apple also has not said whether different limits will apply to Siri conversations, image generation and other cloud features.
Siri still has no exact consumer release date #
Cook’s pricing answer did not provide a firmer Siri release date. Apple’s official commitment remains an English-language beta for supported devices later in 2026. Developer testing began June 8 across iOS 27, iPadOS 27, macOS 27 and visionOS 27, with support for additional languages due after the initial beta.[2]
That timetable is distinct from the broader Apple Intelligence release. Apple says its other new AI features will reach users with its operating-system updates this fall, while Siri AI is separately described as a beta coming later in the year.[2]
Availability will vary geographically. Apple has said Siri AI will initially be withheld from iPhones, iPads and Apple Watches in the European Union, although supported Macs and Vision Pro devices there can receive it. Siri AI and the other new Apple Intelligence features will remain unavailable in China while Apple addresses regulatory requirements.[2]
Five models split work across devices and cloud infrastructure #
Apple’s third-generation foundation-model family consists of five models custom-built in collaboration with Google. AFM 3 Core is the successor to Apple’s 3-billion-parameter on-device model. AFM 3 Core Advanced is a 20-billion-parameter sparse model that activates between 1 billion and 4 billion parameters for a request, reducing the active-memory requirement on supported consumer hardware.[3]
Three additional models run through Private Cloud Compute: AFM 3 Cloud for general server workloads, ADM 3 Cloud for image generation and editing, and AFM 3 Cloud Pro for complex reasoning and agentic tool use. Four of the five models were designed for Apple silicon. For Cloud Pro, Apple worked with Google and Nvidia to extend Private Cloud Compute to Nvidia GPUs hosted in Google Cloud.[3]
The arrangement follows the multiyear agreement announced by Apple and Google in January. The companies said Apple’s next-generation foundation models would be based on Gemini models and cloud technology while Apple Intelligence continued to operate through Apple devices and Private Cloud Compute.[4]
Apple reported $109.4 billion in fiscal third-quarter revenue, up 16% from a year earlier. [5] The July 30 call was Cook’s last as CEO before John Ternus takes over on September 1.
Apple still has not disclosed the numerical free allowance, the additional capacity attached to each iCloud+ plan or whether heavier AI usage will bring new subscription prices.
[6]## Companies mentioned
Further sources #
[1] Axios, “Tim Cook says heavy Siri users may have to pay extra,” July 30, 2026. ↗ [2] Apple Newsroom, “Apple Intelligence brings powerful AI capabilities into everyd… ↗
[[3] Apple Machine Learning Research, “Introducing the Third Generation of Apple’s F… ↗](https://machinelearning.apple.com/research/introducing-third-generation-of-apple-foundation-models)
[[4] Google and Apple, “Joint statement from Google and Apple,” January 12, 2026. ↗](https://blog.google/company-news/inside-google/company-announcements/joint-statement-google-apple/)
[[5] Apple Newsroom, “Apple reports third quarter results,” July 30, 2026. ↗](https://www.apple.com/newsroom/2026/07/apple-reports-third-quarter-results/)
[[6] The Associated Press, “Strong iPhone sales propel Apple’s fiscal Q3 results pas… ↗](https://apnews.com/article/apple-earnings-revenue-iphone-ai-94102918cb3592ebc1d2a38c4d7d819a)
The stories that matter, in one email. Free — unsubscribe anytime.