Voice-to-Text Startup Wispr Flow Triples Valuation to $2 Billion in Nine Months

Wispr Flow, the app that turns speech into text on your screen, just raised $280 million at a $2 billion valuation, according to Reuters. That's nearly triple the $700 million valuation it had in November 2025, when it closed a $25 million Series A extension.

Menlo Ventures led the new round, the same firm that led the last one. Existing backers Notable Capital, NEA, Neo Ventures, 8VC and MVP Ventures returned, and new money came in from Acrew, Forerunner, Goodwater, Peak XV, Together Fund and PLUS Capital, according to Reuters and Cryptopolitan. Total funding since the company's 2021 founding now sits at $361 million.

The round also pulled in a roster of athletes as investors, including Dak Prescott, Joe Burrow, Klay Thompson, Paul George, Trae Young and Domantas Sabonis, according to Cryptopolitan and Pulse2.

The Growth Numbers Behind the Price Tag

Wispr says revenue grew more than 150% in each of the past four quarters, according to Pulse2. Menlo Ventures partner Matt Kraning put that even higher, telling Fortune the growth rate runs "over 30 times year on year," according to TheNextWeb's reporting on Menlo's published investment memo.

The company says Flow is used across more than 10,000 businesses, with Reuters and AOL citing that figure directly. WTVBAM's version of the Reuters story goes further, reporting the company's claim that its software has generated more than 60 billion words and is used at "nearly all Fortune 500 companies." Pulse2 puts the enterprise count even higher, at more than 125,000 businesses.

Startup usage stats get rounded, updated and cited differently depending on when a reporter got the company's pitch deck.

The Accuracy Problem Nobody's Hiding

TechCrunch reported, per TheNextWeb, that Wispr users had been complaining recently about a dip in transcription quality. Cryptopolitan confirmed the same thing: "several users complained recently that Flow's dictation had gotten worse."

Wispr's own numbers back that up. CEO Tanay Kothari wrote in a company blog post that in tough conditions, background noise, wind, heavy accents or music, error rates currently run above 30% of words. That's roughly one in three words wrong before the company's new model ships.

In difficult conditions today, before Canto ships, the product gets more than three words in ten wrong.

The fix, called Canto, is Wispr's first proprietary speech-recognition model, built by a new research group called the Wispr Advanced Interfaces Lab and led by Chief Scientist Ariya Rastrow, a founding member of Amazon's Alexa team, according to Pulse2 and Cryptopolitan. Kothari says Canto cuts those error rates down to between 5% and 10% in the hardest conditions, a roughly fourfold improvement, and should reduce how often users have to edit a dictation by 30% to 35%.

Betting Beyond Dictation

Wispr isn't just trying to fix voice-to-text. It's trying to get out of a crowded lane entirely. Competitors including Willow, Monologue, Aqua and Superwhisper are chasing the same dictation market, according to Cryptopolitan, alongside a wave of cheaper and free tools.

So Wispr shipped a meeting-transcription tool called Notetaker about a week before the raise, putting it up against Granola, Fireflies, Read AI and Otter. CTO Sahaj Garg told Cryptopolitan that Notetaker records locally and doesn't join calls as a visible bot, which puts the burden on users to disclose recording in states that require consent. Two competitors, Otter and Granola, are already facing separate California lawsuits over privacy, according to Cryptopolitan, though neither Wispr nor its rivals have been charged with wrongdoing, and no similar suit has been filed against Wispr.

The Skeptical Take, On the Record

Not every voice in this story is cheering. Michael Ashley Schulman, a partner at Cerity Partners, gave Reuters the most direct caution available: "Triple digit quarterly growth off a small base is the easiest number in venture capital to generate for a few quarters and the hardest one to keep producing once the early adopters stop being the entire customer base."

A near-tripling in valuation in nine months, on a product with a documented accuracy problem it's still fixing, is a bet on a growth curve continuing, not a proven durable business. Menlo's own memo, quoted by TheNextWeb, frames this as a wager that AI's next bottleneck is the interface, not the model, comparing Wispr's raise to a much smaller but thesis-matching €7 million round for a German startup called amber built on the same argument.

Wispr hasn't disclosed profitability, churn rates, or how much of that revenue growth comes from expanding within existing enterprise accounts versus signing new ones. Those numbers, not the valuation headline, will determine whether Menlo's bet on the "interface layer" pays off or whether this becomes another AI-era valuation that outran the business underneath it.