Skip to content
Sign inSubscribe

Sections

  • News
  • Articles
  • Topics
  • Authors
  • Newsletter

Topics

  • Business
  • Regulation
  • Security
  • Infrastructure
  • Robotics
  • Work

Company

  • About
  • Contact
  • Privacy
  • Terms
  • Podcast

Newsletter

Daily AI news and weekly essays. Free, unsubscribe anytime.

Free. Unsubscribe anytime. Privacy


© 2026 ainews.today. Independent reporting and essays on artificial intelligence.

AI Search

Perplexity says it has crossed a $1 billion revenue run rate

Chief executive Aravind Srinivas gave the figure at a Semafor event, at least four times the level reported for the start of 2026.

Eduard Mur·Nov 19, 2026

Summary

Aravind Srinivas, in glasses and a gray Perplexity sweatshirt, gestures with both hands on a TechCrunch Disrupt stage
Aravind Srinivas, Perplexity’s co-founder and chief executive, at TechCrunch Disrupt in San Francisco on October 30, 2024. · Photo: TechCrunch / Wikimedia Commons, CC BY 2.0

Perplexity has crossed a $1 billion annual revenue run rate, a yearly projection based on current sales, chief executive Aravind Srinivas said on November 16 at a Semafor event, according to Semafor’s account.

The figure would be at least four times the company’s level at the start of the year. The Information reported in August that Perplexity’s annualized revenue was under $250 million in January and had passed $750 million by then.

“We just crossed $1B on an annual run rate today,” Srinivas told Semafor editor-in-chief Ben Smith at the Silicon Valley & The World event, in a clip Semafor posted on X. “We spent 5x lower than the company that’s going to IPO this week to get to that number.” Anthropic priced its initial public offering on November 18.

Semafor@semafor

"We just crossed $1B on an annual run rate today," @AravSrinivas tells @semaforben. "We spent 5x lower than the company that's going to IPO this week to get to that number."

November 16, 2026 · View on X

Srinivas attributed the lower spending to a capital-efficient model, Semafor reported. Perplexity mostly routes requests to models built by other companies, such as OpenAI and Anthropic, instead of training a flagship model of its own.

The company has spent the autumn extending Computer, its agent, software that carries out multi-step tasks for a user. Automations, launched on September 29, run recurring work on a schedule or in response to events and remember earlier runs. On November 6 Perplexity added Teammates, agents that take on ongoing jobs, and a mode that runs Computer entirely on Apple-silicon Macs.

On the day of the interview Perplexity released its Monitor API, which watches a topic or web page and sends a webhook, an automatic notification to a developer’s software, when something relevant changes. It also opened Frontier Research, a lab that it said has more than 10,000 GPUs and will publish papers, data, models, and code.

Perplexity is privately held and does not publish financial statements, and the run-rate figure has not been independently verified. The company was founded in 2022 and is based in San Francisco.

Topics

AI SearchBusinessAgents

Tags

PerplexityAravind SrinivasAI agents

Eduard Mur

News editor
More by Eduard →

Get the next one in your inbox

Daily news and weekly essays, free.

Free. Unsubscribe anytime. Privacy

Related articles

All related articles →
First-person view in Battlefield 1942 of a soldier holding a rifle on a sandy island with watchtowers and palm trees, shown inside a browser window
Regulation

Fans use Claude to port games to the browser, and publishers push back

Claude Opus 5.5 can turn a game’s machine code into a browser port in a day. Which ports stay online depends on what they ship and whose name they use.

Nov 15·10 min read
Mati Staniszewski, in a light shirt, smiles while looking to the right in a dimly lit room
Business

How ElevenLabs is turning its voice technology into enterprise software

Listeners rank its newest voice model first, by a narrow margin and at four times a rival’s price. The company now sells agents tuned on customers’ data.

Nov 15·8 min read
A small yellow car on a race track passes under a red-and-white inflatable arch, with a lap timer in the corner of the screen
Agents

How Three.js maintainers started writing documentation for AI agents

The 3D library has shipped instructions for language models since February, and its latest release credits Claude as co-author of 65 changes.

Nov 15·7 min read

Back to home