Perplexity account plan
Answer engine powering curiosity
At a glance
Perplexity is a US AI answer engine and agent company (Perplexity Computer) that says it reached a $1 billion annualized revenue run rate in November 2026.
Perplexity, founded in 2022 by Aravind Srinivas (CEO), Denis Yarats, Johnny Ho and Andy Konwinski, built an AI answer and research engine and now sells Perplexity Computer, an agent that orchestrates over 20 frontier models across files, tools and connectors. It is privately held and independent, last valued at about $20 billion in a September 2025 funding round, having raised about $900 million in total as of mid-2026. Srinivas said it reached a $1 billion annualized revenue run rate in November 2026, up from below $250 million at the start of 2026, and that it could file for a US IPO as early as Q1 2027. The team was about 400 people in June 2026. It does not train its own flagship model and routes questions to outside models such as OpenAI's ChatGPT and Anthropic's Claude.
Business units
Search (answer engine)
The original product: an AI answer engine with citations, downloaded over 80 Mn times, with Pro and Max subscriptions. Srinivas says consumer search has been commoditized and that the answer engine was always a lead generator for the frontier products. India is its largest market, helped by an Airtel bundle of a free one-year Pro subscription for Airtel's 360 Mn users.
Perplexity Computer (agents)
An agent system that turns a request into teams of agents working across models, files, tools and connectors, including Personal Computer running continuously on a dedicated Mac mini. It has driven most of the 2026 growth and targets non-developer work: finance, corp dev, sales reps, data science and research analysts.
Enterprise
Perplexity Enterprise Pro is deployed at companies such as Tech Mahindra, which rolled it out in June 2026 across its sales and customer-facing teams, including customer partners and senior sales leaders.
Strategic priorities
- Agents that do the work
- Srinivas says the money is now in agents that go and do work, not in answering questions; Computer drove revenue from below $250 million at the start of 2026 to more than $750 million by August.
- Orchestration across models
- Perplexity positions itself as the orchestrator that routes each step to the best model, local or server-side, and calls token value per watt per user the key metric.
- Knowledge work beyond developers
- Computer targets finance, corp dev, sales, data science and research analysts; in a June 2026 study with Harvard Business School it cut estimated time by 87% and cost by 94% across 10,000 matched pairs of tasks.
- Lower cost per task
- Perplexity is post-training its own models on open-source models and moving some inference to local devices to bring down what it spends on frontier model tokens and raise margins.
- Public listing
- Srinivas said he would not rule out filing for a US IPO in the first quarter of 2027, earlier than the previous 2028 target.
Latest results
Srinivas said in November 2026 at a Semafor event that Perplexity had reached a $1 billion annualized revenue run rate, roughly 4x growth during 2026, after below $250 million at the start of the year and more than $750 million by August. In April 2026 he said revenue had grown fivefold from $100 million to $500 million with only 34% growth in team size, and that the company expected to double revenue again in 2026. In June 2026 he said the burn had been brought down by more than 50%. Perplexity says it spent 5x less than Anthropic to reach the $1 billion run rate.
Investments and deals
- Billion Dollar Build
- Perplexity gives a million dollars of compute credits to any group with a credible path to building a billion-dollar company, and Srinivas wants a thousand such companies.
- Own models
- Post-training its own models on top of open-source models, to serve existing product features on models it owns and serves itself.
- Local compute
- Personal Computer runs continuously on a dedicated Mac mini; Perplexity is moving some inference to local devices to cut cost and keep sensitive data local.
Challenges and risks
- Dependence on outside models
- Perplexity does not train its own flagship model and routes questions to systems such as ChatGPT and Claude, so part of its spending becomes revenue for Anthropic.
- Commoditized search
- Srinivas says consumer search has been commoditized and that Google's AI mode now looks like Perplexity.
- Cost of continuous agents
- Srinivas says the real concern with always-on agents is cost, which is why Perplexity is pursuing hybrid local and server inference.
- Promotional growth in India fading
- India downloads peaked at 22 Mn in October 2025 and fell to nearly 14 Mn by July 2026 after the Airtel promotional offer ended.
What leadership says
“The frontier in AI is not about answering questions anymore. It's about actually going and doing work for you.”
“We spent 5x lower than the company that's going to IPO this week to get to that number”
“The rate at which our business is growing, I wouldn't rule out filing something in the first quarter of 2027.”
“Everyone here works at a small business, and everything we build is for people who build.”
Where a supplier can help
- Connectors and data for Computer
- Computer's value rests on connectors to the tools and data a team already uses; providers of business data, apps and integrations fit its push into finance, sales and research work.
- Inference and compute cost
- With the plan to post-train its own models and move inference to local devices, suppliers of inference capacity, GPU compute or model optimisation map to its cost-reduction priority.
- IPO readiness
- A possible US IPO filing in Q1 2027 implies work on finance, reporting, governance and investor relations, where advisers and software vendors can help.
- Enterprise go-to-market
- Enterprise Pro deployments such as Tech Mahindra's and the targeting of non-developer functions point to needs in enterprise sales, onboarding and security tooling.
Tamtam's reading of the priorities and challenges above, not the company's own words.
Selling to companies like Perplexity?
Tamtam finds the companies that match your ideal customer, researches each account and tells you who to contact and when.
Find similar companiesKey documents
Read in full by Tamtam; each summary links to the document itself.
- Perplexity May Launch IPO As Early As Q1 CY27: CEO Aravind Srinivas (Inc42)
- Perplexity has brought forward its IPO plan and could file draft papers for a US listing as early as Q1 2027, CEO and founder Aravind Srinivas told CNBC.
- His words: "The rate at which our business is growing, I wouldn't rule out filing something in the first quarter of 2027."
- Srinivas says ARR (annual recurring revenue) has reached $1 Bn, above the $750 Mn ARR expectation it had set ahead of an IPO; the earlier target date was 2028.
- In 2025 Srinivas had said Perplexity would not go public before 2028, whatever happened to the OpenAI and Anthropic IPOs.
- Founded in 2022 by Srinivas, Denis Yarats, Johnny Ho and Andy Konwinski; the answer and research engine has been downloaded over 80 Mn times.
- India is its largest market, per the founder; an Airtel deal bundled a free one-year Perplexity Pro subscription for Airtel's 360 Mn users.
- India downloads rose 625% month on month to 5.9 Mn in July 2025, peaked at 22 Mn in October 2025 and fell to nearly 14 Mn by July 2026 after the offer ended; estimated monthly India revenue rose from $34,000 (January 2025) to $156,000 (July 2026), per TechCrunch.
- In June 2026 Tech Mahindra began deploying Perplexity Enterprise Pro across its sales and customer-facing teams.
- Perplexity CEO touts $1B run rate, jabs at Anthropic in IPO week (Cryptopolitan)
- CEO Aravind Srinivas said at Semafor's Silicon Valley & The World event (Monday, interviewed by editor-in-chief Ben Smith) that Perplexity reached a $1 billion annualized revenue run rate that same day.
- He claimed capital efficiency: "We spent 5x lower than the company that's going to IPO this week to get to that number" (Anthropic, listing on Nasdaq that week).
- Annualized revenue was below $250 million at the start of 2026 and passed $750 million by August, driven by the Perplexity Computer agent; roughly 4x growth in the year.
- Perplexity does not train its own flagship model; it mostly routes questions to outside systems including OpenAI's ChatGPT and Anthropic's Claude, so part of its spend becomes Anthropic revenue.
- Srinivas says users only "really care if the work actually got done", not which model answered.
- Last valued at about $20 billion in a September 2025 funding round; it had raised about $900 million in total as of mid-2026.
- 20VC with Harry Stebbings: How Export Controls Helped Not Hurt China & Power is the Bottleneck to AI | Perplexity CEO
- Introduced as a $20 billion company with 45 million users and over a billion searches a month, built in 3 years; Srinivas puts the team at about 400 people and says it may be 800 or 1,000 in two years.
- Revenue "has more than tripled since the beginning of the year" and the burn was brought down by more than 50%; he declined to share the figure but said it is "far far more" than $500 million ARR.
- Strategy: the answer engine "was always a lead genen" for frontier products; people pay for Pro or Max for deep research and agents, and the money is now in agents that do work (Perplexity Computer).
- Positioning: Computer orchestrates across models (Claude and GPT inside one harness), tools, files and connectors; the target market is non-developer work (finance, corp dev, sales reps, data science, research analysts), "clot code multiplied by 10".
- Key metric: "the most important metric in AI is token value per what per user"; plan to cut costs by post-training its own models on open-source models and moving inference to local devices.
- Bearish on ads in chat interfaces: they hurt the trust users place in accurate answers.
- One Computer user spends upwards of about $10,000 a month; continuous agent loops (cron jobs) are where the frontier is.
- Billion Dollar Build: $1 million of compute credits for any team with a credible path to a billion-dollar company; he wants a thousand such companies.
- IPO: had said 2028, hopes it "can be sooner than that"; revenue growth matters more than profitability today.
- How Computer is reshaping knowledge work (Perplexity blog)
- Perplexity published, with researchers at Harvard Business School, a study of how people use Perplexity Computer in its first three months.
- Computer orchestrates over 20 frontier models across files and tools, turning a request into teams of agents that work asynchronously.
- On near-identical requests, Computer ran an average of 26 minutes of machine work versus 33 seconds for Search (a 48x increase) and made about twelve times as many calls to external apps and connectors per session.
- Across 10,000 matched pairs of tasks, Computer cut estimated time by 87% and cost by 94%; a 269-minute workflow dropped to 36; savings held across 18 domains.
- In interviews the median user reported work going 25x faster; some reported speedups over 300x.
- 23% of Computer requests were things never asked of Search; 76% required analysis, evaluation or creation (55% on Search); 71% involved abstract, non-routine work (53% on Search).
- Across 8,000 users in eight occupation groups, people worked outside their primary occupation 59% of the time with Computer versus 50% with Search.
- Next question for Perplexity: how these individual gains aggregate at team and organisation level.
- Perplexity AI Reports 5x Revenue Growth to $500 Million (Analytics India Magazine)
- CEO Aravind Srinivas said Perplexity grew revenue fivefold, from $100 million to $500 million, while the team grew 34%.
- His post: "We just 5X'ed revenue from $100M to $500M with only 34% growth in team size."
- The company expects to double revenue again in 2026 while keeping a relatively small team.
- Growth has come mainly from Computer, an integrated AI system combining tools, files, memory and models that can allocate tasks across 19 different models.
- Personal Computer runs on a dedicated Mac mini, operating continuously and linking a user's local applications to Perplexity's servers.
- Srinivas frames the pivot to Computer as a return to roots serving founders and small businesses: "Everyone here works at a small business, and everything we build is for people who build."
- The Financial Times reported ARR rose from $16 million to $305 million over two years and exceeded $450 million as of March 2026.
- Founded in 2022; last raised about $200 million in 2025 at a valuation of roughly $20 billion.
Questions about Perplexity
What does Perplexity do?
Perplexity runs an AI answer engine and sells Perplexity Computer, an agent that orchestrates many AI models across files, tools and connectors to carry out knowledge work such as research, analysis and building documents.
What is Perplexity's revenue?
CEO Aravind Srinivas said Perplexity reached a $1 billion annualized revenue run rate in November 2026, up from below $250 million at the start of 2026. It is private and publishes no accounts.
Who leads Perplexity?
Aravind Srinivas is co-founder and CEO. He founded the company in 2022 with Denis Yarats, Johnny Ho and Andy Konwinski.
How many people work at Perplexity?
Srinivas put the team at about 400 people in June 2026.
Is Perplexity going public?
Srinivas told CNBC he would not rule out filing for a US IPO in the first quarter of 2027; the earlier target had been 2028.
Written by Tamtam from the company's own published documents listed above, read in full. Figures are as those documents state them, for the periods they name. (2026-11-19)