Nvidia (NVDA) Stock Gets Apple Boost — But There’s a Catch

TLDR

• Apple confirmed its most advanced AI model, AFM Cloud Pro, runs on Nvidia GPUs via Google Cloud
• Nvidia stock edged up around 0.9% to $210.44 in premarket trading on the news
• Apple is not buying chips directly — it’s accessing Nvidia hardware through Google’s cloud infrastructure
• AFM Cloud Pro is Apple’s first AI model to officially run on third-party, non-Apple chips
• Nvidia’s “ambiguous confidential compute” technology made the deal possible by meeting Apple’s strict privacy standards

---

Apple has officially confirmed that its most advanced AI model runs on Nvidia chips — but the market reaction has been pretty subdued, and there’s a reason for that.

Nvidia stock was up around 0.9% to $210.44 in premarket trading on Tuesday after the news broke at Apple’s annual WWDC event on Monday.

NVIDIA Corporation, NVDA

The model in question is called Apple Foundation Model Cloud Pro, or AFM Cloud Pro. Apple AI executive Amar Subramanya described it as comparable to Google’s Gemini frontier models.

It runs in the cloud on Nvidia GPUs — but here’s the catch. Apple isn’t buying those chips directly from Nvidia. Instead, it’s accessing them through Google Cloud, as part of Apple’s Private Cloud Compute infrastructure.

That arrangement likely explains the muted stock reaction. The deal may not translate into major new chip orders for Nvidia, at least not right away.

Apple VP of software Sebastian Marineau-Mes said Apple specifically wanted Nvidia’s latest chips, but needed them configured in a more private way — one where the chips couldn’t read what was on the servers.

How Nvidia’s Privacy Tech Made the Deal Possible

A recent Nvidia development called “ambiguous confidential compute” gave Apple and Google the technical foundation they needed to build a system that met Apple’s privacy standards.

“We wanted to avail ourselves of the latest technology from Nvidia, and so we set out to extend private cloud compute to third-party cloud,” Marineau-Mes said.

Apple software SVP Craig Federighi was clear that Apple Intelligence uses Apple’s own models — not Google’s off-the-shelf Gemini, as many had assumed. Google’s technology was used to help train Apple’s own models, not to replace them.

“These four models… are custom built for Apple Silicon, trained using proprietary data with reinforcement learning and refined using outputs from Gemini frontier models,” Subramanya said.

Apple and Google had already announced a broader partnership in January. Monday’s WWDC keynote was the first time Apple officially confirmed Nvidia’s involvement.

Apple’s Lean AI Strategy

Apple has deliberately avoided the multi-billion dollar AI infrastructure buildout that its rivals have pursued. Federighi took a subtle dig at that approach during the keynote.

“Some appear to be racing forward, seemingly pursuing AI for the sake of AI, without clear regard for the people — all of us — that it’s ultimately meant to serve,” he said.

Apple’s pitch is privacy-first AI: models that run locally where possible, and only hit the cloud when needed, with personal data like calendar entries and messages used to personalize responses.

For Nvidia, landing Apple as a customer — even indirectly — is still a win. It validates Nvidia’s chips as the go-to hardware for AI workloads, even as competition grows from Intel, AMD, and custom silicon makers.

Nvidia stock was up 1.73% on the day. Apple closed down 1.89%.