Wispr raises $280M at $2B and previews its first speech model

Wispr announced a $280 million Series B on Monday at a $2 billion valuation, led by Menlo Ventures. The round takes the dictation startup’s total raised to $361 million.

The price is the part worth pausing on. Reuters reported that Wispr’s previous round, a $25 million Series A extension in November, valued the company at $700 million. That makes the new figure 2.9 times the November mark in nine months, our arithmetic from the two numbers Reuters gave. That’s a steep move for a tool that types what you say.

Round	When	Amount	Valuation
Series A extension	November 2025	$25M	$700M
Series B	August 2026	$280M	$2B
Total to date	Cumulative	$361M	n/a

Menlo led alongside existing backers Notable Capital, NEA, Neo Ventures, 8VC and MVP Ventures, all of which returned. Acrew, Forerunner, Goodwater, Peak XV, Together Fund and PLUS Capital came in new.

The money arrived with a model. Wispr previewed Canto, its first proprietary speech model, and the pitch is accuracy in rooms that aren’t quiet, so the claim is testable. Chief executive Tanay Kothari put a number on it.

In the hardest conditions, with background noise, wind, heavy accents or music, error rates fall from more than 30% of words to somewhere between 5% and 10%.

Tanay Kothari, chief executive of Wispr, via Reuters

The mechanism sits in the training data. Most voice models are trained on extremely clean voice assets, the company said in comments reported by SiliconANGLE, while Canto was fed dogs barking, traffic and other voices so it can pull one speaker out of the noise. The result, the company says, is 30% to 35% fewer dictations that need editing.

That matters because of when it lands. Several users had complained over the previous few weeks about a quality dip in Flow’s output, according to TechCrunch. Canto reads as a repair as much as a launch.

Competition explains some of the urgency. TechCrunch named Willow, Monologue, Aqua and Superwhisper as rival dictation apps, and noted developers shipping free or cheaper tools for prosumers. Meanwhile, Wispr’s new meeting notetaker puts it up against Granola, Fireflies and Read AI.

Scale is the counterweight. People have written more than 60 billion words with Flow, and the product is in use at more than 10,000 enterprises, including almost all of the Fortune 500. Fortune reported revenue growth north of 150% for four consecutive quarters.

Even so, growth off a small base is the easy part.

The value may be there if the growth rate survives, since 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.

Michael Ashley Schulman, partner at Cerity Partners, via Reuters

Wispr isn’t stopping at dictation, though. It is partnering with hardware makers such as the Oasis ring so people can dictate without speaking loudly, and last month it announced Wispr Interface Labs under Ariya Rastrow, who worked on Amazon Alexa in its early days.

Canto is a preview, not general availability, so those error rates rest on Wispr’s own testing rather than anything independent. The thing to watch is whether the 5% to 10% figure survives contact with real kitchens and real cars. The second thing to watch is the price, because what a $2 billion valuation actually buys is a claim on growth that has to keep arriving, which is the open question across the whole sector.