Apple has established a sprawling and intricate compute architecture, one that ropes in Google and NVIDIA to paper over its embarrassing AI-related shortcomings. Even so, Apple's WWDC 2026 keynote answered as many questions as raised new ones. Thankfully, the Cupertino-based tech giant is now issuing clarifications at the speed of lightning, resolving lingering uncertainties on a war footing of sorts.
Apple craftily obfuscates Google's contributions to its new Apple Intelligence architecture, taking pains to point out its own technologies at the core of this new paradigm
We already know that Apple Intelligence consists of a combo of on-device and cloud-based models. Even so, this distinction was not very granular.
Thankfully, Apple has just provided a critical update, noting that the gigantic cloud-based Apple Foundation Model (AFM) is its own creation, albeit distilled from an equivalent Google Gemini model. Of course, we already know that Apple licensed a 1.2-trillion-parameter Gemini model from Google a few months back.
It seems the iPhone maker had only licensed Google's technology for model distillation purposes. Apple also takes pains to note that it conducted its own pre-training and post-training operations on the AFM Cloud.
Apple has also detailed the architecture of its Private Cloud Compute (PCC) framework, going on to note:
- What's new with PCC on Google Cloud is the implementation: "NVIDIA Confidential Computing with NVIDIA GPUs, Intel CPUs with TDX, and Google's Titan chip."
- Apple states that while the AFM Cloud is hosted in Google Cloud, the arrangement comes with "the industry’s most comprehensive transparency guarantees that allow external security researchers to verify our privacy commitments."
- "To mitigate the risk of supply chain attacks, we maintain a cryptographically verifiable, append-only ledger of all Google Cloud hardware that is part of the PCC fleet."
- "PCC on Google Cloud leverages many of the same architectural security patterns as PCC on Apple silicon to implement these layered protections: initial network data parsing for each request happens in a dedicated process within its own namespace, shared inference software is recycled with a short time-to-live duration, and attested keys are held in a separate, dedicated confidential VM isolated from external inputs."
- Apple also says that it will "provide public research tooling, and access to live PCC nodes in research mode through the Apple Security Bounty Program."
Apple has further clarified that the AFM Cloud itself is divided into 2 categories: a Pro model that runs on NVIDIA GPUs within Google Cloud, and a vanilla model as well as an image generation one that runs on Apple's own servers.
As far as on-device Apple Foundation Models are concerned, the AFM Core Advanced has 20 billion parameters, but only needs the quantum of parameters strictly needed to process a given inference request. Critically, this model was entirely designed by Apple, and requires the A19 Pro chip to run on an iPhone.
Of course, Apple has also prepared a less powerful on-device model for generalized inferencing on older iPhones. When a user submits a request, for instance, via the Siri AI, a localized orchestrator calls the required tools, collects data, and then generates the prompt for the AFM Cloud. Critically, raw data is not sent to the cloud, just the structured prompt.
Of course, this comes as Apple spent the better part of the technical presentation downplaying Google's role within the new Apple Intelligence and Private Cloud Compute framework.
About the author: Writing is my one incontrovertible passion. Over the past six years, he has authored over 2,200 distinct articles on financial and tech-related topics, spanning nearly 1 million words. And he has been a member of Wcctech mobile team since 2025. As an alumnus of the University of Toronto, Rotman Commerce Program, I bring nuance, in-depth knowledge, and a unique perspective to every topic that I cover. When I'm not writing, I'm traveling the world, exploring hidden confectionaries and restaurants as an aspiring food connoisseur.
Follow Wccftech on Google to get more of our news coverage in your feeds.
