Aravind Srinivas and the climb to a $22.6 billion valuation
Aravind Srinivas leads Perplexity with a research background that informs its product architecture. He holds a PhD in Computer Science from UC Berkeley and previously held research positions at OpenAI and Google DeepMind. Srinivas co-founded the company in August 2022 with Denis Yarats, Johnny Ho, and Andy Konwinski. The company reached a $21.21 billion valuation following its Series E-6 funding round in early 2026, a trajectory that places it among the fastest-valued AI companies in history. By June 2026, Perplexity raised a cumulative $1.72 billion across eleven funding rounds. Investors like Jeff Bezos, Nvidia, and SoftBank have backed the firm. The company’s scale results from his decision to build a retrieval-augmented generation layer. This architecture fetches live web content during inference, which anchors factual claims to clickable URLs. Perplexity processed 780 million queries in May 2025 alone. His academic career focused on machine learning and computer vision. At the University of Montreal, he interned under Yoshua Bengio in 2016. During his time at UC Berkeley from 2017 to 2021, his supervisor was Pieter Abbeel. His research included BoTNet, which incorporated self-attention into ResNet backbones for image classification. He also co-authored work on Contrastive Predictive Coding with Olivier Henaff, Jeffrey De Fauw, Ali Razavi, Carl Doersch, and Aaron van den Oord. Srinivas’s work on reinforcement learning included Decision Transformers, which demonstrated how to apply transformer architecture to offline reinforcement learning.
Products and the subscription model
The product lineup expanded from a simple search engine into a multi-tool ecosystem. Perplexity released Comet, a Chromium-based browser, in July 2025. This browser allows users to manage tabs, fill forms, and book flights. In February 2026, the company released Perplexity Computer, an AI agent for multi-step tasks. This agent uses various large language models and specialized subagents to function. The company also launched the Shopping Hub in November 2024 with backing from Amazon and Nvidia. This platform generates product recommendations and allows purchases through the Perplexity interface. In October 2024, the company added finance-focused features that track stock prices and company earnings. These features source data from Financial Modelling Prep.
| Tier | Monthly Cost | Key Capabilities |
|---|---|---|
| Free | $0 | 5 Pro Search queries per day, base Sonar model |
| Pro | $20 | Unlimited Pro Searches, 20 Deep Research queries, multi-model access |
| Max | $200 | 10,000 Perplexity Computer credits, Model Council, priority routing |
| Education Pro | $10 | Pro feature set for verified academic accounts |
| Enterprise Pro | $40 | SSO, admin controls, Internal Knowledge Search |
| Enterprise Max | $325 | Unrestricted access, dedicated infrastructure, priority SLA |
You already know that search is changing, so I will focus on the specific costs. I find the $200 price point for the Max tier extremely steep, but the inclusion of 10,000 Perplexity Computer credits and priority routing through the Comet browser provides utility for professional users. The company also provides an API for developers. The Sonar Pro model costs $3 per million input tokens and $15 per million output tokens. The scale of the user base reached 45 million monthly active users. In January 2026, Perplexity signed a three-year, $750 million agreement with Microsoft to use Azure cloud and Foundry services. This agreement provides access to models from OpenAI, Anthropic, and xAI. Amazon Web Services remains the primary cloud provider. The company also expanded into mobile in 2026 through a partnership with Samsung. The technology powers search and reasoning for Bixby and Samsung Internet features on Galaxy S26 devices.
Legal battles and the subscription shift
Perplexity faces significant legal pressure from media organizations. In May 2026, CNN sued, claiming Perplexity scraped 17,000 stories. The New York Times filed a lawsuit in December 2025, alleging the company copied millions of articles without permission. Reddit also sued in October 2025, alleging unlawful data scraping. The constant legal threats from major publishers make the company’s operational stability look precarious. To address trust, leadership transitioned to a subscription-first model in February 2026. This move ended the AI-integrated advertising strategy to prioritize objective results over ad revenue. Srinivas maintains that Perplexity only aggregates information rather than plagiarizing it. In June 2024, Forbes accused Perplexity of using its original reporting without proper credit. Srinivas responded that the feature had "rough edges" but maintained the company only aggregates information. In January 2025, Perplexity Solved Solutions sued the company for trademark infringement. This lawsuit claims the name "Perplexity" violates a federally registered trademark. In August 2025, the Japanese newspaper Yomiuri Shimbun filed a lawsuit for the "free-riding" use of 120,000 articles. This was followed by lawsuits from The Asahi Shimbun and The Nikkei. How many more lawsuits will the company face before it settles its data disputes? Amazon also sued Perplexity over the Comet shopping agent, and a court ordered the company to stop using it for purchases on the platform.
