Sources: Perplexity's estimated ARR rose to $450M+ in March, jumping 50% in a month after the launch of its Computer agent and its shift to usage-based pricing
Context & Ripple Effects
Perplexity's revenue trajectory had already moved from roughly $10M in early 2024 to about $50M later that year, then toward $100M in early 2025 and $150M by mid-2025, alongside rising private-market valuations. The reported March step-up is therefore less a standalone growth datapoint than a test of whether its product expansion can convert usage into materially more revenue.
The company is pairing its agent offering with architecture intended to divide work between local and cloud models, while its earlier search product had demonstrated substantial query activity. That makes the shift from broad AI-search engagement to a large query base for its AI search product and paid task execution the key commercial transition.
First-order effects
- Perplexity's reported $450M-plus ARR run rate, following a 50% monthly jump, strengthens the immediate case that its Computer agent and usage-based pricing are monetizing demand more effectively than a primarily subscription-led model.
- Usage pricing makes the economics of each agent task more consequential: Computer's local/cloud task routing can directly affect both customer value and Perplexity's serving costs.
Second-order effects
- Rival AI search and agent products face added pressure to show not just adoption but task-level revenue, especially where expensive model inference makes flat subscription pricing difficult to sustain.
- The result also shifts attention from Perplexity's less successful commerce experiments—its difficulty scaling shopping and advertising efforts—toward agent usage as a potentially clearer monetization path.
Third-order effects
- If repeatable, this points to AI consumer products evolving from answer engines funded by subscriptions or ads into metered task platforms, where pricing and model-routing decisions become core product strategy.
- That model could favor providers that can demonstrate a low cost per useful completed task; it also makes reported ARR more sensitive to usage intensity and therefore potentially less comparable with conventional subscription revenue.
The trend: AI search companies are testing whether autonomous-task agents and consumption pricing can turn high engagement into durable revenue while containing inference costs.