Dan Ives, the widely followed analyst at Wedbush Securities, has identified physical AI as Nvidia's next "holy grail" — the next leg of growth the chip designer is positioning itself to capture after the boom in AI data center accelerators.

Ives made the remarks in a Yahoo Finance interview, casting the transition from cloud-based generative AI to AI that acts in the physical world as the defining opportunity of Nvidia's next phase. The label "holy grail" signals his view that physical AI could rival or exceed the data center GPU cycle that has already transformed Nvidia's revenue base and made it the most closely watched supplier in semiconductors.

Physical AI refers to machine intelligence embedded in machines that perceive, reason about, and act in real environments — robotics, autonomous vehicles, and industrial automation systems, as opposed to the language and inference workloads that run in hyperscale data centers. For Nvidia, the category extends its platform beyond the GPU clusters powering large model training and inference and into the compute substrates of factories, warehouses, and vehicles.

Ives's framing matters because it addresses the central question facing Nvidia investors after two years of extraordinary AI-driven growth: where the next tranche of demand comes from once the initial wave of data center buildouts matures. His answer is that the same silicon, software stack, and developer ecosystem Nvidia built for AI compute can migrate into physical systems, opening an addressable market that is geographically broader and less dependent on a handful of hyperscale buyers.

The analyst's track record on the AI infrastructure cycle has been prominently bullish throughout, and his choice of language — reserving the term "holy grail" for physical AI rather than any further extension of the data center story — positions robotics-adjacent computing as the frontier he believes Wall Street has not yet fully priced.

Nvidia itself has leaned into the same theme, but Ives's comments put an analyst's valuation argument behind it: if physical AI adoption follows the trajectory of generative AI, the suppliers of the underlying compute stand to capture a disproportionate share of the value, and Nvidia's established position in accelerated computing gives it the pole position.

What remains unresolved in Ives's thesis is timing. The generative AI cycle moved from research curiosity to tens of billions in quarterly silicon demand in roughly two years; physical AI deployment, tied to robotics hardware cycles and industrial capital expenditure, is likely to unfold on a slower cadence. Ives did not offer shipment or revenue projections tied to physical AI in the interview, and no vendor has yet disclosed capacity commitments specific to the category.

For now, the claim stands as a directional bet rather than a forecast grounded in disclosed order books. If Ives is right, the competition to supply compute for machines that move — rather than servers that sit — will set the terms of the next semiconductor demand cycle.