Perplexity has confirmed that it plans to use Nvidia’s new Vera CPU, a move that adds another customer to Nvidia’s effort to push beyond GPUs and into the CPU market dominated by Intel and AMD.
Nvidia says the chip is built for AI agents and other data center workloads, while Perplexity says it fits the type of work its systems already need to handle.
Perplexity Vice President for Computer Enterprise and Infrastructure Nate Kupp said Nvidia’s CPU ran AI agent coding tasks about 1.5 times faster than traditional CPUs. He also said Vera was “a dead-on fit” for a number of the company’s core workloads.
Based on Perplexity’s comments, the company appears to have evaluated Vera on workload fit and task speed, rather than on brand value or market hype.
Nvidia’s press release describes Vera as “the first CPU built for AI agents” and says it is now in full production. The company says the chip delivers 1.8 times faster task completion than x86 CPUs and is aimed at agentic AI, reinforcement learning and data processing.
Nvidia also says Vera uses 88 Olympus cores and LPDDR5X memory bandwidth of up to 1.2 TB/s. The company says systems using Vera will be available from system builders and cloud partners starting this fall.
Nvidia is trying to extend its role from AI accelerators into the CPU layer, especially for agentic workloads that involve tool use, code execution and orchestration.
AI agents do not behave like typical human users of CPUs; they run continuously, call tools, and complete chained tasks. According to Reuters, that is one reason Nvidia is targeting this market, and Perplexity’s comments indicate the company sees measurable benefit in that design.
[ALSO READ: AI Workloads Are Shaping FinOps Priorities: Redefining Cloud Economics in 2026 ]
