Perplexity revenue surges 50% as startup pivots from search to AI agents

Perplexity Revenue Surge Signals Major Strategy Shift

AI startup Perplexity has reported a dramatic 50% increase in Perplexity revenue over just one month, with annual recurring revenue climbing to more than $450 million in March 20251

1

. The surge reflects the company's accelerating strategic pivot to AI agents and away from its original chatbot-style search engine that once positioned it as a potential challenger to Google2

2

.

The revenue growth follows the February launch of Perplexity Computer, an agentic tool designed to carry out complex tasks with limited human supervision, and the introduction of a usage-based pricing model that charges top-tier subscribers beyond their allocated credits1

1

. This marks a significant shift from search engine operations to autonomous task execution, mirroring broader industry trends as companies experiment with new pricing structures to reflect the higher inference costs of running AI agents.

Rapid Growth Amid Competitive Landscape

The company now serves 100 million monthly active users across its search and agent tools, including tens of thousands of enterprise clients1

1

. Revenue streams from enterprise subscriptions and consumer tiers ranging from $20 to $200 monthly, with the usage-based pricing model adding variable revenue on top of these subscription fees3

3

.

Before implementing the new pricing system, AI startup Perplexity had already demonstrated impressive growth, scaling annual recurring revenue from $16 million to $305 million over two years1

1

. However, this trajectory remains smaller compared to competitors like Cursor, which reached $2 billion ARR from under $100 million in 2024, while Anthropic reported $19 billion ARR by February and OpenAI generated approximately $20 billion last year2

2

.

Product Expansion Beyond Search

Perplexity's shift from search engine functionality accelerated with the 2024 launch of Comet, an AI-powered web browser capable of acting on users' behalf through voice and text commands to handle shopping, social media summarization, and email management1

1

. CEO Aravind Srinivas described the subsequent Perplexity Computer launch as "the next big thing," while the company also introduced Model Council, which runs queries through different models and displays outputs side-by-side2

2

.

The platform offers access to multiple model providers including OpenAI's GPT, Anthropic's Claude, and open-source options from Chinese companies like DeepSeek's R1 and Moonshot's Kimi1

1

. One advantage cited by sources close to the company involves triaging requests to the most efficient model for each purpose—using OpenAI's Codex and Anthropic's Claude Code for coding tasks, GPT-5 for writing, or Opus for reasoning1

1

.

Legal Challenges and Market Position

The company's original search focus attracted controversy, with copyright infringement and plagiarism lawsuits filed by publishers including The New York Times and Britannica alleging illegal content copying1

1

. Additional privacy lawsuits claim the company shared user data with Google and Meta without consent, though Perplexity has denied wrongdoing on all claims1

1

.

Despite legal headwinds, investor confidence remains strong. Perplexity was valued at $20 billion in September 2024, up sharply from $500 million at the start of that year4

4

. Backers include Nvidia, SoftBank's Vision Fund 2, venture capitalists New Enterprise Associates and IVP, along with Amazon founder Jeff Bezos and Meta's former chief AI scientist Yann LeCun1

1

.

When Nvidia CEO Jensen Huang encouraged audiences at the company's annual conference to subscribe to Perplexity Computer and "pay the maximum amount," Srinivas responded that "there is no limit, you can spend as much as you want"1

1

. This exchange highlights how AI groups are exploring pricing models to cover heavier workloads driven by agentic systems.

The company remains lossmaking as it pays model providers like OpenAI and Anthropic to use their models, alongside inference costs of running queries1

1

. A Perplexity executive told the Financial Times that revenue retention was "strong" without providing specific figures1

1

. The usage-based pricing model, while potentially boosting revenue, creates more volatile metrics for estimating annual growth compared to traditional subscription models2

2

.