Perplexity to Use Nvidia's New CPU, Reinforcing Full-Stack AI Computing Push

What Perplexity's CPU decision changed

Perplexity, the AI-powered search and answer engine, said it plans to run part of its infrastructure on Nvidia's new data-center CPU rather than relying only on chips from traditional processor makers. Nvidia has spent the past two years building out its own CPU line to sit alongside its dominant AI accelerators, pairing the two so customers can buy a more complete rack of computing hardware from a single supplier instead of stitching together GPUs from Nvidia with CPUs from someone else.

For years Nvidia's business was almost entirely about the GPU, the chip that does the heavy lifting for training and running AI models. The CPU is the more traditional processor that manages a server, moves data around, and coordinates workloads. By selling both together, Nvidia is trying to capture more of the total dollar value in every AI data center that gets built, not just the accelerator slice.

Why it matters for AI infrastructure and chip stocks

A single customer commitment does not move Nvidia's revenue by itself. Perplexity is a fast-growing startup, not a hyperscaler spending tens of billions on data centers. What makes this notable is the signal it sends: an AI-native company chose Nvidia's own CPU over established alternatives when it had a real choice to make. Every one of these adoption decisions makes it a little easier for Nvidia to argue that customers want the whole system, not just the GPU, and that argument matters more to Nvidia's long-term growth than the revenue from any one deal.

It is also a read on competition. Data-center CPUs have historically been a market led by established processor makers, and Nvidia entering that space with its own designs is a direct challenge to that position. A customer like Perplexity choosing Nvidia's chip is a small but real data point in that contest, even though this particular story does not name a clear loser with enough certainty to justify a negative call on any other stock.

Which stocks, and why

Nvidia is the direct beneficiary here. The company's strategy of selling GPUs, CPUs, and networking gear as one bundle depends on customers actually buying the CPU piece and not just the accelerators. Each AI-native company that signs on, even a smaller one like Perplexity, adds evidence that the strategy is working and gives Nvidia a broader base of revenue per data center it supplies. This is a direct, structural piece of Nvidia's growth story rather than a one-quarter event, since infrastructure choices like this tend to stick once a company builds its software and operations around a given chip family.

No other listed company is named in this story with enough specificity to map a confident impact. The traditional CPU makers that compete with Nvidia in this space are not named in the report, and guessing at a negative impact on them from one startup's decision would be reading too much into a single data point.

What to watch

Watch whether larger cloud providers and AI labs follow smaller companies like Perplexity in adopting Nvidia's CPU line at scale, since that is what would actually move Nvidia's systems revenue in a meaningful way. Also watch Nvidia's own disclosures on what share of data-center revenue now comes from full systems versus GPUs alone, which is the clearest measure of whether this bundling strategy is gaining real traction across the industry rather than just picking up a handful of early adopters.