Grammarly Had Revenue on Day One. The Number Everyone Quotes Is Two Different Numbers.

Quick answer (2026): Grammarly's annual recurring revenue runs $350M (2021), $500M (2022), $650M (end of 2024) and more than $700M (May 2025), per Sacra's analyst estimates and the company's own May 29, 2025 financing announcement, which states "more than $700 million" in annual revenue and "more than 40 million users rely on Grammarly daily." A second and much smaller series also circulates widely, running $43.5M (2019), $88.7M (2021), $178.9M (2023) and $251.8M (2024). Those two ledgers are not competing estimates of the same quantity, and the gap is not growth. One is recurring run rate, the other is recognised revenue, and only the first has a named source. Two more facts that get lost underneath the number: Grammarly took no outside capital for its first eight years (founded 2009, first round May 2017) and was already profitable with 7 million daily users by 2017. And since October 29, 2025, the company that owns it is not called Grammarly at all. It is called Superhuman Platform Inc.

Why this search returns two incompatible numbers

Run the search yourself and watch Google contradict itself inside a single answer box. On the day this was written, the AI Overview for "grammarly revenue" led with "more than $700 million" in annualized revenue and then, four lines later, listed "Prior Revenue: Reported around $251.8 million in 2024."

Read literally, that is a company that grew 178% in about five months, at a scale where nobody grows 178% in five months.

It did not. The two figures are different accounting bases wearing the same word. Once you separate them, the numbers stop fighting and the more interesting story underneath becomes visible.

The sourced ledger

Here is the series that has a name attached to it. Sacra tracks private company revenue and reports Grammarly on an annual recurring revenue basis:

Source: Sacra company page for Grammarly, retrieved August 2026, cross-checked against the company's own May 29, 2025 announcement for the 2025 figure. The 2023 value is not separately published; Sacra states the 2023 to 2024 growth rate rather than the level.

The final row is the only one that is not an estimate. Grammarly itself put "more than $700 million" in writing on May 29, 2025. Slator reported the same period more precisely: "As of Q2 2025, Grammarly is a profitable at-scale business with an annualized revenue of more than USD 700m, supported by an individual user base estimated at 40 million and over 50,000 Grammarly Business accounts."

Note the word in both. Annualized. That is a run rate, not a year of collected revenue.

The other ledger, and where it comes from

The smaller series is everywhere. It appears on Growjo, which currently models "$626.6M per year," on Electro IQ, which reports "In 2024, the company grew revenue-wise to US$251.8 million, at a 40.7% growth rate from US$178.9 million in 2023," and in a widely shared Startup GTM breakdown that puts revenue at $43.5M in 2019 and $88.7M in 2021.

Try to trace any of those figures to a filing, a disclosure, or a named analyst and the trail stops. None of those pages cite a source for the revenue series. There is no filing to check, because Grammarly has never been a public company and has no statutory accounts in the way that, say, JetBrains does in the Czech register.

There is one thread worth pulling. SQ Magazine, which does cite its sources, attributes a figure of approximately $250 million in 2024 GAAP revenue to Sacra, alongside the approximately $650 million ARR figure for the same moment. That single line explains most of the confusion on this SERP: the same analyst, the same company, the same year, two numbers roughly 2.6 times apart, because they measure different things.

The honest caveat is that this reconciliation does not fully close. At 2021 the two series sit about four times apart, $350M ARR against $88.7M, and an ARR-versus-recognised-revenue gap alone does not usually stretch that far. So the correct read is not "one series is ARR and the other is GAAP, mystery solved." It is narrower and more useful: one series is attributable and the other is not, and where the unattributed series can be checked against a named source, it is off by multiples rather than percentages.

What ARR and recognised revenue actually do at this company

The distinction is not academic here, and it is the part an operator can use.

Annual recurring revenue takes the contracted subscription base at a moment and annualises it. Recognised revenue counts what the accounting rules let you book as earned in a period. A subscription business with heavy annual prepay and expanding enterprise seats will systematically show ARR above recognised revenue, because cash arrives and contracts are signed before the service is delivered. Grammarly is exactly that shape: 40 million daily individual users on consumer subscriptions, plus more than 50,000 business accounts on seat-based contracts.

That is why the gap widens rather than closes as the enterprise side grows. It is a structural feature of how the company sells, not an accounting irregularity.

The part the revenue number hides: eight years, no capital

Underneath the metric argument is the fact that makes Grammarly worth writing about at all.

Max Lytvyn, Alex Shevchenko and Dmytro Lider founded Grammarly in 2009 in Kyiv. They did not raise a round until May 2017. That is eight years, and it was not a comfortable eight years. Speaking to CNBC in January 2025, the founders described funding it from the proceeds of their previous company, MyDropBox, a plagiarism detection tool sold to Blackboard in 2008.

How much did they put in? Shevchenko: "It was sub-$1 million." Lytvyn: "It was close to all we could do at the time."

Within a year they were, in Lytvyn's words, "on the edge of running out of money." The founders did not take salaries until 2011, at the point the company reached 300,000 paid subscribers. The first version of the product required users to upload a document, press process, and wait five to ten minutes. They hit one million daily users in 2015 after switching to freemium, and by 2017 Sacra records them at 7 million daily users and already profitable, which is the year they finally took the $110M Series A.

The line that does the most work is this one, from Lytvyn:

"We started selling the product right away. Essentially, we had revenue on Day 1, and used it to fund further development. Our users paid for building the rest of Grammarly."

The $1 billion that was not a funding round

The May 2025 headline read like a mega-round. It was not one.

The $1 billion from General Catalyst came out of its Customer Value Fund. Slator describes the structure plainly: "The funds will operate like a loan or credit line with capped returns tied to revenue, rather than an equity stake." Grammarly's own announcement says the capital will "scale sales and marketing and for strategic acquisitions."

So a profitable company with more than $700M of run rate took a billion dollars of growth capital without issuing a share. Its last equity valuation remains the $13 billion set in the November 2021 round led by Baillie Gifford and BlackRock, four years earlier and never marked since.

The company put that capital to work fast. It acquired Coda, announced in December 2024 and closed in January 2025, a deal that also delivered its next chief executive: Coda's Shishir Mehrotra took over as CEO. On July 1, 2025 it acquired Superhuman, the email client, reported by Reuters and TechCrunch. In June 2026 it added GPTZero, the AI content detector, in a deal reported at around $88 million.

The company you are searching for was renamed

On October 29, 2025, Grammarly Inc. became Superhuman Platform Inc. The announcement, "Becoming Superhuman", was written by Mehrotra. Fast Company covered the decision the same day and noted the trade being made: the Grammarly brand survives as one agent inside a four-product suite alongside Coda, Mail and Superhuman Go.

This matters for anyone reading a revenue figure. A search for "Grammarly revenue" now returns numbers for an entity that no longer carries that name, in a portfolio where the writing assistant is one line of business among several. Any figure quoted after October 2025 without saying whether it covers Grammarly the product or Superhuman Platform the company is ambiguous by construction, and most of them do not say.

What operators should take from this

Pick your metric before you need it to say something. Grammarly reports annualized revenue because that is the number a growth investor prices. A company with statutory filings reports recognised revenue because a regulator requires it. Neither is dishonest. But if you run your own business on ARR while your bank account runs on cash collected, the gap will find you, and it is widest exactly when you are growing fastest and selling the most annual prepay. That is the trap in going annual-first, and it is the same mechanic at $32K MRR as at $700M.

Customer-funded is a real financing strategy, not a consolation prize. "Our users paid for building the rest of Grammarly" is a sentence describing eight years, one previous exit spent down to nearly nothing, and no salaries until year two. It worked. It also very nearly did not, repeatedly, by the founders' own account.

The counter-frame, and it is the important one: the eight bootstrapped years did not produce the $13 billion valuation. They produced a profitable company with 7 million daily users, and that is what made a $110M round available on good terms in 2017. The sequencing is easy to get backwards. Grammarly did not stay unfunded out of principle and get rewarded for purity. It stayed unfunded until it had leverage, then used it. Compare that with Zoho, which never took the money at all and built a larger business more slowly. Both are legitimate. They are different bets, and only one of them ends with a $1 billion capital facility from General Catalyst.

And be honest about the base rate. Two founders who had already built and sold a company, with sub-$1 million of their own proceeds and deep domain knowledge of the exact problem, is not the median bootstrapper's starting position. The lesson transfers. The starting conditions do not.

The honest bottom line

The best-supported answer to "what is Grammarly's revenue" is more than $700 million on an annualized basis as of May 2025, stated by the company, with roughly $650M ARR at the end of 2024 and a recognised-revenue figure closer to $250M for 2024 on a different basis. The smaller year-by-year series circulating on stat aggregator sites has no traceable provenance and should not be quoted as though it does. And since October 2025, all of it belongs to a company called Superhuman Platform Inc.

This is an editorial reconciliation assembled from company announcements, press reporting and third-party analyst estimates, not a company disclosure. Grammarly has never published audited financial statements. Figures labelled estimate are exactly that, and the accounting basis is stated wherever it is known.

Keep reading

• The month I went annual first at $32K MRR: a founder diary on exactly the ARR-versus-cash gap described above, at a scale where getting it wrong ends the company.
• The month my gross margin scared me at $38K MRR: what it feels like when the headline number keeps climbing and the number that pays you does not.
• JetBrains files its real numbers: the opposite case, a company whose figures are genuinely filed and audited, and what changes when you can actually check.