Perplexity AI Revenue Soars 50% in 30 Days Following Major Product Launch

Key Takeaways Annual recurring revenue at Perplexity AI climbed beyond $450 million with a 50% increase in 30 days Growth accelerated following the introduction of “Perplexity Computer” agent platform and usage-based subscription model Monthly active users exceed 100 million, with enterprise clients numbering in the tens of thousands Latest valuation stands at $20–$21.2 billion, while markets indicate 13% odds of reaching $50–75 billion Legal challenges include copyright claims from major publishers and a privacy-related class action lawsuit Perplexity AI has reached a significant financial benchmark, with Annual Recurring Revenue (ARR) surpassing $450 million in March 2026. This represents a remarkable 50% acceleration within just 30 days, marking the most dramatic revenue expansion in the company’s history since its 2022 establishment. BREAKING: Perplexity's revenue has reportedly surged +50% in one month after shifting into AI agents, per FT. As a result, Perplexity's revenue has doubled in one quarter to more than $450 million in ARR. This follows Anthropic's push into the space which said its ARR hit $19… pic.twitter.com/nvT4nexVeH — The Kobeissi Letter (@KobeissiLetter) April 8, 2026 Prior to this recent milestone, the company had demonstrated impressive momentum by scaling ARR from $16 million to $305 million across a two-year period. While that trajectory was noteworthy, two strategic initiatives launched on February 25 catalyzed unprecedented growth. The primary catalyst was the introduction of Perplexity Computer, an autonomous agent infrastructure that integrates existing AI functionalities into a unified platform. This enables users to execute complex workflows without navigating multiple applications. The secondary driver was a fundamental restructuring of the company’s monetization approach. Perplexity transitioned to usage-based billing for premium service tiers. Users now receive allocated credits, with additional consumption requiring incremental...