Nvidia is in discussions to invest in Perplexity as part of an equity round that would value the AI search startup at more than $30 billion, according to a report by The Information1,2,3.
The deal would mark a valuation jump of more than 50% from Perplexity's previous financing — a round at $20 billion in September 2025. A subsequent round in January 2026 valued the company at $23 billion.
Perplexity's annualized revenue has surged past $750 million, fueled largely by Perplexity Computer, a cloud-based AI agent that automates multi-step professional tasks. At $750 million in annualized revenue, a $30 billion valuation would price Perplexity at close to 40 times sales.
Nvidia is not a new backer. The chipmaker participated in Perplexity's $18 billion extension in July 2025, alongside SoftBank Vision Fund 2, NEA, and IVP. Nvidia also backed Perplexity's Series B in January 2024 and returned for the company's unicorn round in March 2024.
Before the current talks, Nvidia had weighed alternative approaches to deepening its relationship with Perplexity, including a technology licensing deal and staff hires.
Perplexity was founded in August 2022 by Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski. The company's fundraising trajectory has been steep: $73.6 million at a $520 million valuation in January 2024, $63 million at $1.04 billion in March 2024, $500 million at $9 billion in late 2024, $500 million at $14 billion in May 2025, and the $100 million extension to $18 billion in July 2025. CEO Aravind Srinivas has floated an IPO within the next few years.
ANALYSIS The progression from roughly $500 million in valuation to more than $30 billion in under two years reflects the pace at which private-market capital has repriced AI-native search and agent products. Nvidia's repeated participation across multiple Perplexity rounds — and its consideration of licensing and hiring alternatives before settling on another equity investment — suggests the chipmaker views Perplexity as a strategically significant distribution surface for AI inference workloads, not merely a financial bet.