Stock Markets July 7, 2026 03:05 PM

Perplexity to Run Nvidia's Vera CPUs for Continuous AI Agent Workloads

Startup cites faster agent coding performance as Nvidia pushes into traditional CPU territory

By Caleb Monroe
Share
Twitter Reddit Facebook LinkedIn
INTC NVDA AMD

Perplexity said it will deploy Nvidia’s newly announced Vera central processing units to handle tasks for AI agents, noting faster coding performance versus conventional processors. Nvidia expects substantial revenue from the Vera line as it moves into a market historically dominated by Intel and AMD. Perplexity did not disclose purchase volume.

Perplexity to Run Nvidia's Vera CPUs for Continuous AI Agent Workloads
INTC NVDA AMD
Summarize with
ChatGPT Perplexity Claude Grok Gemini

Key Points

  • Perplexity will deploy Nvidia’s new Vera CPUs to run AI agent workloads, citing faster coding performance.
  • Nvidia projects approximately $20 billion in sales from the Vera CPU series by the end of its fiscal year.
  • The move highlights competitive pressure in the CPU market long led by Intel and AMD and affects semiconductor and cloud infrastructure sectors.

Perplexity on Tuesday announced plans to adopt Nvidia’s new central processing units to support AI agent workloads, signaling another customer win as Nvidia expands beyond its established GPU business.

Nvidia has framed the Vera CPU series as a key plank in its move toward more general-purpose computing hardware, and it has projected up to $20 billion in sales from Vera by the end of the current fiscal year. The company’s push arrives while AI-focused firms are increasingly developing specialized silicon tailored to their workloads.

The modern CPU market has been led for years by Intel and Advanced Micro Devices, whose processors power everything from consumer laptops to large-scale web servers. Many of those existing chips were designed before the rise of autonomous AI agents - software programs that, after receiving instructions, can carry out multiple linked tasks without human intervention.

AI agents typically operate continuously and hand off subtasks without the pauses common to human interaction. In describing Perplexity’s reasoning for the switch, Nate Kupp, the company’s vice president for compute enterprise and infrastructure, said Nvidia’s CPU completes AI agent coding tasks at roughly 1.5 times the speed of traditional CPUs.

Perplexity did not disclose the number of Vera processors it intends to acquire. Nvidia has previously said that other AI companies, including OpenAI, Anthropic and Oracle, will make use of its CPUs, indicating a broader commercial push.


Context and market implications

The rollout of Vera chips represents a strategic move by Nvidia into a sector long dominated by Intel and AMD. For companies running persistent, agent-driven workloads, the technical advantages Nvidia reports could translate into operational efficiencies, but the scale of adoption remains unclear pending buyer disclosures.

Data points referenced

  • Nvidia expects up to $20 billion in revenue from the Vera CPU line by the end of the fiscal year.
  • Perplexity reports Vera CPUs execute AI agent coding tasks about 1.5 times faster than traditional CPUs, according to Nate Kupp.
  • Perplexity has not revealed the size of its Vera CPU purchase.

This article presents the available details about Perplexity’s intended use of Nvidia’s Vera CPUs and related market statements. Where the original reporting left specifics unreported, such as the total purchase volume Perplexity plans, those gaps remain reflected here.

Risks

  • Perplexity did not disclose how many Nvidia CPUs it will purchase, leaving the scale and financial impact of the deployment uncertain - impacts cloud infrastructure and procurement planning.
  • Nvidia’s projected $20 billion in Vera sales is a forecast and its realization will affect semiconductor market dynamics and incumbent CPU vendors such as Intel and AMD.
  • Shifts in workload requirements for AI agents may alter demand patterns for traditional CPUs versus new architectures, creating uncertainty for server hardware and data center capacity planning.

More from Stock Markets

JPMorgan Highlights Three U.S. SMid Biotech Opportunities: BridgeBio, ORIC, Olema Jul 7, 2026 Fitch Rates Amazon’s New Debt 'AA-' as Company Eyes Large Capex Push Jul 7, 2026 Moody's Lowers Cable One Credit Grade, Cites Prolonged Broadband Weakness Jul 7, 2026 Anthropic widens Claude Cowork to mobile and web, enabling cross-device AI work sessions Jul 7, 2026 K2 Airways Cargo 737 Loses Contact After Reporting Navigational Fault en Route to Karachi Jul 7, 2026