Is Perplexity AI in Trouble? The Truth About Its $500M ARR, Lawsuits, and the Agent Pivot

The growth is driven by two things: the February 2026 launch of "Computer," Perplexity's autonomous AI agent product, and a shift to usage-based pricing layered on top of existing subscriptions . Sacra estimates the company serves over 100 million monthly active users and tens of thousands of enterprise clients .

The Advertising Experiment That Failed

In February 2026, Perplexity abruptly abandoned its ad-based revenue model, stopping new advertisers and winding down existing placements . The reason was simple: ads were a rounding error. The company reportedly earned only about $20,000 in ad revenue in 2024, compared to $34 million in total company revenue — 0.06% of the total . Perplexity executives stated they have no plans to revisit the ad model .

A Major Partnership Collapsed

Snap confirmed in Q1 2026 that it had ended its $400 million partnership with Perplexity . The deal, announced in November 2025, had initially sent Snap's stock up 15%. Its collapse contributed to Snap's stock dropping ~10% in premarket trading after earnings . Snap's investor letter stated that Q2 guidance "assumes no contribution from Perplexity as we amicably ended the relationship in Q1."

Legal Exposure Is Real

Perplexity faces multiple high-profile lawsuits:

• The New York Times sued Perplexity in December 2025 for copyright and trademark infringement, asserting five counts including direct and contributory copyright infringement .
• Amazon filed a lawsuit to halt Perplexity's Comet browser from making purchases on behalf of users .
• Reddit also sued the company .

These cases could create significant liability or force operational changes.

Industry Skepticism Is Loud

At the Cerebral Valley Summit in San Francisco in 2025, over 300 AI founders and investors voted Perplexity the AI startup "most likely to fail" — ahead of even OpenAI . Critics point to Perplexity's sky-high valuation surge ($14B to ~$50B in a short period), thin margins, and reliance on other companies' models and data .

The CEO Says There's No Problem

CEO Aravind Srinivas has publicly denied financial distress. In responses on Reddit, he stated the company has "all the money" it needs and no IPO plans before 2028 . The rumors largely originated from a Reddit post noting that Perplexity had stopped marketing spend — which Srinivas attributed to product improvements, not cost-cutting .

Still Unprofitable — And Inference Costs Are a Structural Drag

Perplexity remains private and unprofitable . Every user query costs the company money for AI inference, and that structural headwind has not yet been overcome at scale . While agent products have dramatically increased revenue, the underlying unit economics remain a challenge.

Bottom Line: Not in Trouble, But Not Out of the Woods

Perplexity is not in imminent financial trouble. Its revenue acceleration is among the fastest ever seen in AI. But the company remains unprofitable, carries substantial legal risk, lost a major partnership, and its earlier ad-model failure raised real questions about long-run unit economics. The next 12–18 months will determine whether the agent pivot creates a sustainable business — or just buys time.