Perplexity AI confirmed this week that its annual recurring revenue hit $450 million in March, a 50% jump from $305 million in just one month — the fastest monthly gain in the company's history. The Financial Times first reported the figure, citing internal documents seen by the publication. Since its founding in 2022, Perplexity had taken two years to reach $305 million; now it added $145 million in annualized revenue within 30 days.
One Big Change: Computer Platform and Credits Model
The catalyst was a dual launch on February 25: Computer, an autonomous agent platform that orchestrates up to 19 specialized AI models from OpenAI, Anthropic, and Google to execute multi-step tasks, and a credits-based pricing model that charges users for consumption beyond a monthly allocation. CEO Aravind Srinivas described the system as “one reasons, another codes, another writes.” Perplexity also abandoned advertising entirely in February, citing concerns that ads would undermine trust in AI outputs, leaving subscriptions and usage fees as its sole revenue streams.
The revenue trajectory reflects a broader industry shift: users will pay significantly more for AI that does things than for AI that says things. Usage-based pricing ties income directly to compute consumed by agent workflows, aligning monetization with delivered value.
From Search Engine to Enterprise Automation Rival
Perplexity now competes not with Google but with enterprise automation platforms. Gartner projects that 40% of enterprise applications will include task-specific agents by end of 2026. As reported by crypto.news, AI integration is reshaping corporate spending, with budgets flowing to tools that produce outputs rather than answers. The company still faces copyright lawsuits from The New York Times and Britannica, plus a separate privacy suit it has denied.
Perplexity had set an internal target of $656 million ARR by end of 2026 — once considered aggressive. At the current monthly pace, that goal is within reach. The next challenge: whether enterprise retention holds as the novelty of autonomous agents fades and competitors deploy similar orchestration layers at scale.