Perplexity AI Statistics (2026): Revenue, Users, Queries and Valuation
You can find reports describing 30 million users, tens of millions of users, hundreds of millions of monthly queries, rapidly rising revenue, and a $20 billion valuation. Those numbers can all be useful, but they do not measure the same thing.
The clearest 2026 picture is this: Perplexity’s annual recurring revenue exceeded $450 million in March 2026, before CEO Aravind Srinivas said in April that the company had grown revenue from $100 million to $500 million. Its latest reported financing valued the company at $20 billion in September 2025. Perplexity does not publish an official global monthly-active-user count.
This report separates revenue, users, queries, funding, app usage, and web traffic so each Perplexity AI statistic keeps the context needed to cite it correctly.
Perplexity AI statistics at a glance
Metric | Latest useful figure | What it means |
Annual recurring revenue | More than $450M, March 2026 | FT estimate of ARR after rapid growth |
Revenue milestone | $500M, April 2026 | CEO statement; his post did not separately define the accounting period |
Valuation | $20B, September 2025 | Reported valuation in a $200M funding round |
Total funding | $1.5B, September 2025 | PitchBook total reported by TechCrunch |
User base | “Tens of millions,” February 2026 | TechCrunch description; Perplexity does not publish MAU |
Earlier user milestone | ~30M users, April 2025 | CEO figure reported by the Financial Times |
Queries | 780M in May 2025 | CEO-reported monthly query volume |
India mobile MAU | ~14M, July 2026 | Sensor Tower estimate after an October 2025 peak of ~22M |
Computer models | 19 models | Company-stated architecture for Perplexity Computer |
These figures should not be combined into one growth rate. Revenue, monthly queries, user counts, web visits, and mobile MAU describe different parts of the business.
Get the Mobile Testing Playbook Used by 800+ QA Teams
Discover 50+ battle-tested strategies to catch critical bugs before production and ship 5-star apps faster.
How much revenue does Perplexity make?
Perplexity’s reported revenue trajectory accelerated sharply between 2024 and 2026.
The Financial Times reported that the company was generating more than $35 million in annualized revenue in August 2024. By mid-2025, FT reporting put annualized revenue at approximately $150 million.
By September 2025, TechCrunch reported that Perplexity’s annual recurring revenue was approaching $200 million.
The jump became much larger in early 2026.
The Financial Times reported in April that estimated annual recurring revenue had climbed to more than $450 million in March 2026, following the launch of Perplexity Computer and a greater emphasis on usage-based and agentic products.
Days later, CEO Aravind Srinivas wrote that Perplexity had increased revenue from $100 million to $500 million while increasing headcount by 34%. The Economic Times reported the statement on April 14.
There is an important wording distinction here.
The FT explicitly describes the March figure as annual recurring revenue. Srinivas' April post uses the broader word revenue. It is therefore safer to write:
Perplexity's ARR exceeded $450 million in March 2026, and CEO Aravind Srinivas subsequently said the company had reached a $500 million revenue milestone.
It is less defensible to silently turn the $500 million CEO statement into audited full-year revenue.
Annualized revenue is not completed annual revenue
ARR or annualized revenue extrapolates a current run rate.
If a company generates revenue at a particular monthly pace and maintains it for 12 months, the annualized number estimates what that pace would produce.
That is useful for a fast-growing private company, but it is not the same as revenue already earned during a completed financial year.
This distinction matters especially for Perplexity because its run rate has been changing quickly.
How much is Perplexity worth?
Perplexity’s reported valuation rose rapidly as investors backed its expansion from AI search into browsers, enterprise products, and agents.
By early 2024, the company was valued at roughly $500 million. By May 2025, CNBC reported that Perplexity was completing a $500 million funding round at a $14 billion valuation.
In July 2025, further financing reportedly valued the company at $18 billion.
Then, in September 2025, TechCrunch reported that Perplexity had secured $200 million in new capital at a $20 billion valuation, citing reporting from The Information. Reuters separately reported the same financing terms.
TechCrunch also reported, citing PitchBook, that Perplexity had raised $1.5 billion in total funding since its founding.
That makes $1.5 billion the cleaner cumulative-funding figure to use.
Do not reconstruct a higher total by manually adding reported rounds unless you can establish that the source's cumulative figure excludes the latest transaction.
Valuation is not revenue
A $20 billion valuation does not mean Perplexity produces $20 billion in sales.
A private-company valuation reflects the price investors place on the business during a financing transaction. Revenue measures money generated through operations.
They answer completely different questions.
How many people use Perplexity?
This is where Perplexity AI statistics become much less precise.
Perplexity does not publicly report a standard global monthly-active-user figure.
That is not an accidental omission.
During a February 2026 briefing, a Perplexity executive told TechCrunch that the company does not emphasize MAU as a core goal. TechCrunch described Perplexity's overall user base at the time as being in the “tens of millions.”
There is, however, an earlier company-attributed milestone.
The Financial Times reported that Srinivas said in April 2025 that Perplexity was serving approximately 30 million users.
Later reporting continued to reference a user base around that scale. But “users” and “monthly active users” are not automatically equivalent.
That makes the safest formulation:
Perplexity has publicly been described as serving tens of millions of users, but the company does not disclose a standardized global MAU figure.
Any precise global MAU estimate from an analytics company or statistics roundup should remain clearly labeled as an estimate rather than presented as a Perplexity disclosure.
How many queries does Perplexity process?
Perplexity has disclosed a much clearer query milestone.
CEO Aravind Srinivas said at Bloomberg's Tech Summit that Perplexity processed approximately 780 million queries in May 2025, up more than 20% from the previous month.
That works out to roughly 25 million queries per day on average during that month, although the monthly figure is the directly disclosed number.
The 780 million milestone represented a large increase from 2024.
The Financial Times had previously reported approximately 250 million queries in a month during 2024, compared with about 500 million across all of 2023.
Do not automatically extrapolate the May 2025 growth rate into 2026.
A 20% month-over-month rate maintained indefinitely would produce unrealistic projections. Unless Perplexity publishes a newer query milestone, 780 million monthly queries in May 2025 remains the stronger directly attributed figure.
What does current Perplexity traffic show?
Queries are not the same thing as website visits.
A single user session can contain multiple queries, while Perplexity activity can also occur through apps, browsers, enterprise products, APIs, and newer agentic interfaces.
Similarweb's August 2026 analysis shows why that distinction matters.
According to Similarweb, visits and unique visitors to Perplexity.ai peaked around October 2025 and declined in most subsequent months through May 2026. Among ChatGPT, Gemini, Claude, and Perplexity, Perplexity was the only platform in that comparison whose recent web trend was downward.
At the same time, Similarweb says Perplexity's worldwide app audience still showed substantial year-over-year growth.
Those statements can both be true.
A decline in website visits does not necessarily mean total platform usage is falling, particularly when a company is expanding into:
mobile apps
enterprise products
APIs
Comet
Computer
other agentic workflows
Web traffic should therefore stay labeled as web traffic, not total Perplexity usage.
How large is Perplexity in India?
India provides one of the strongest recent examples of how distribution can reshape an AI product's user base.
In July 2025, Perplexity partnered with Airtel to offer eligible customers a free 12-month Perplexity Pro subscription. TechCrunch reports that Airtel had roughly 360 million customers at the time of the offer.
Sensor Tower data shared with TechCrunch shows the immediate impact.
Perplexity recorded approximately 5.9 million app downloads in India in July 2025, up 625% from the previous month. Across the seven months when the offer was open to new users, Sensor Tower estimated 56 million downloads.
Monthly active users also surged.
India mobile MAU reached 8.9 million in July 2025, eventually peaked at approximately 22 million in October, and stood at nearly 14 million in July 2026.
That July 2026 level was down 37% from the peak but remained more than five times the roughly 2.6 million monthly users Sensor Tower estimated Perplexity averaged in India during the first half of 2025.
Downloads are not retention
The Indian data also demonstrates why downloads alone can create a misleading growth story.
Once new Airtel redemptions stopped, downloads dropped sharply. Yet a meaningful portion of the acquired user base remained active.
That makes the sequence more informative than any single number:
distribution → downloads → active users → retention → paid conversion
Each stage answers a different question.
For product teams evaluating user behavior, the same distinction matters outside AI search. Quash's guide to app retention statistics explains why installs and active usage should not be treated as equivalent product-health metrics.
Perplexity Computer and the shift beyond search
Perplexity's statistics are becoming harder to interpret partly because the product itself is changing.
In February 2026, the company launched Perplexity Computer, an agentic system designed to execute longer, multi-step tasks rather than simply return an answer to a search prompt.
Perplexity says Computer can independently coordinate 19 different AI models, selecting different models for different tasks and creating subagents when needed.
At launch, Computer was available through Perplexity's $200-per-month Max subscription.
The important part for interpreting Perplexity statistics is not the model count itself.
It is the shift in the unit of value.
Traditional search is easy to describe with:
user → query → answer
An agentic product may instead look like:
user → goal → plan → many model calls → tools → actions → result
One agent task can therefore involve significantly more underlying activity than one visible query.
That makes metrics such as website visits and search queries less complete as Perplexity expands beyond its original answer-engine interface.
For QA teams, the same measurement issue appears in autonomous testing. Quash's overview of agentic QA and autonomous test execution explains why evaluating a full agent loop is different from evaluating one generated response.
Why Perplexity statistics disagree
Most disagreement in Perplexity statistics comes from four measurement problems.
1. Different populations
A metric may count:
registered users
monthly active users
app users
website visitors
paying subscribers
enterprise accounts
API activity
Those populations overlap, but they are not identical.
2. Different channels
Perplexity now operates across its website, mobile app, Comet browser, enterprise products, APIs, and Computer.
A measurement covering only Perplexity.ai cannot describe every one of those surfaces.
3. Different time periods
A monthly query count from May 2025 should not be combined with a July 2026 app-MAU estimate and treated as a single 2026 user snapshot.
Always keep the date with the number.
4. Different evidence types
There is a meaningful difference between:
company disclosure → reported private-company data → measurement provider → modeled estimate
A CEO-reported query milestone is not the same type of evidence as an analytics company's estimated MAU.
Both can be useful, but they need different labels.
What Perplexity's numbers actually tell us
The strongest Perplexity statistics show a company whose economics have been growing much faster than its original search product alone would suggest.
ARR moved from tens of millions of dollars in 2024 to more than $450 million by March 2026. Investors valued the company at $20 billion in its September 2025 financing.
Its scale is still much smaller than the largest consumer AI platforms, and Similarweb's web data shows that its browser-based reach has not followed the same recent trajectory as ChatGPT, Gemini, or Claude.
But that comparison is becoming less complete as Perplexity pushes toward enterprise subscriptions and agentic products.
The correct conclusion is therefore not simply that Perplexity is “growing” or “shrinking.”
Its revenue, app adoption, web traffic, and product mix are moving differently.
That is why they should be measured separately.
The same principle applies when evaluating AI products in software development. Adoption alone does not show whether an AI system performs reliably in production. Quash's AI testing statistics report separates testing adoption from execution and quality outcomes, while its guide to AI-powered QA covers the broader evaluation problem.
Methodology and sources
This report prioritizes sources according to the metric being measured.
For company and financial milestones, it uses company-attributed statements and established financial or technology reporting. For web traffic, it uses Similarweb's own dataset. For mobile activity in India, it relies on Sensor Tower measurements reported by TechCrunch.
Key sources include:
Third-party model-based estimates for global MAU, subscriber totals, citation rates, and 2026 query volume were excluded where stronger direct or measured evidence was not available.
That means the article may contain fewer headline statistics than some roundups.
The trade-off is that the remaining numbers are easier to defend and cite.
Conclusion
The strongest Perplexity AI statistics currently available tell a clear but nuanced story.
Perplexity's annual recurring revenue exceeded $450 million in March 2026, and CEO Aravind Srinivas subsequently said revenue had reached $500 million. Its most recent reported financing valued the company at $20 billion, with TechCrunch reporting $1.5 billion in total funding.
For usage, the evidence needs more care.
Perplexity reported 780 million queries in May 2025, while its global user base has been described as being in the tens of millions rather than through an official MAU disclosure. In India, Sensor Tower measured mobile MAU at nearly 14 million in July 2026, after a peak of approximately 22 million.
The useful rule is simple: keep revenue separate from ARR, queries separate from users, app MAU separate from global usage, and website traffic separate from the entire Perplexity platform.
That context is what turns a Perplexity statistic into evidence instead of another AI-growth headline.
Continue reading
More articles from the Quash editorial desk.
