Apple, Google, and NVIDIA: Strategic Alliance for AI

Apple recently confirmed that Google and NVIDIA play a crucial role in the development of its new generation of artificial intelligence models. By leveraging the technologies of these two giants, Apple aims to accelerate the development of Apple Intelligence without entering the massive investment race led by its competitors. This strategy was unveiled at WWDC 2026, where Apple presented its most advanced cloud model, the AFM Cloud Pro, which relies on technologies from Google and NVIDIA GPUs integrated into its private computing infrastructure.

An Ambitious AI Strategy

Long perceived as cautious in the face of the rise of generative AI, Apple revealed a more ambitious strategy at WWDC 2026. The company announced that its most advanced cloud model, AFM Cloud Pro, is based on technologies from Google and NVIDIA GPUs integrated into its private computing infrastructure. This approach aims to enhance the capabilities of Apple Intelligence and Siri while maintaining the protection of personal data, a key hallmark of the brand.

The Role of Google and NVIDIA in Apple’s AI Infrastructure

During the technical conference that followed the WWDC keynote, Apple executives presented the Apple Foundation Model Cloud Pro (AFM Cloud Pro), the most advanced cloud model developed by the company. Amar Subramanya, head of artificial intelligence at Apple, explained that this model achieves a performance level comparable to Google’s Gemini Frontier models. Apple confirmed that this cloud infrastructure relies on NVIDIA GPUs hosted in Google’s cloud environment.

This revelation clarified the nature of the partnership announced between Apple and Google earlier this year. Contrary to market speculation, Apple Intelligence does not directly use Gemini as a conversational engine. The models are designed by Apple, but their development benefits from the technologies and work done around the Gemini Frontier models.

Executives explained that the third-generation AFM models have been trained with proprietary data. They are enhanced through reinforcement learning techniques and refined based on the results obtained from Google’s most advanced models. This allows Apple to leverage recent industry advancements without directly depending on external models available to the public.

Apple also officially confirmed the use of NVIDIA chips in its private cloud computing infrastructure. This first confirmation is significant, as the company had previously remained discreet about its reliance on the ecosystem of the world leader in AI processors.

A Distinct AI Approach

While Microsoft, OpenAI, Google, and Meta invest tens of billions of dollars in building data centers and training increasingly powerful models, Apple prefers to follow a distinct trajectory. Craig Federighi, senior vice president in charge of software, emphasized the group's desire to build AI primarily designed for the user. The goal is not to develop the largest model on the market, but to offer features directly integrated into Apple devices while limiting exposure of personal data.

This philosophy is reflected in the technical architecture presented at WWDC. Apple has developed a system orchestrator capable of automatically directing each request to the most suitable model, whether it involves local processing on the device or execution in the cloud. This software layer constitutes one of the pillars of Apple’s AI strategy. It allows Siri to utilize information stored locally on the iPhone, iPad, or Mac (calendar, messages, or reminders) to provide more contextualized responses and limit data transfers to remote servers.

To extend this architecture to Google’s cloud infrastructures and NVIDIA GPUs, Apple had to solve a major challenge: preserving its privacy standards. According to Sebastian Marineau-Mes, vice president of software at Apple, "the advancements made around NVIDIA Confidential Computing technology have enabled the construction of an environment that meets the group’s security requirements."

By combining its own AI models, NVIDIA's computing capabilities, and certain technologies from Google, Apple thus gains access to the most advanced resources on the market without incurring the same expenses as its competitors in building giant infrastructures.