San Francisco voice AI startup Wispr Flow has closed a $280 million Series B at a $2 billion valuation, roughly tripling its worth in nine months as investor appetite for enterprise-grade voice interfaces intensifies. The round, announced on August 17, 2026, brings total funding to $361 million and lands alongside the preview of Canto, Wispr's first proprietary speech recognition model.
Menlo Ventures leads with heavy insider participation
Long-time backer Menlo Ventures led the financing, joined by existing investors Notable Capital, NEA, Neo Ventures, 8VC and MVP Ventures. New investors include Acrew, Forerunner, Goodwater, Peak XV, Together Fund and PLUS Capital. The round also drew an unusual roster of athletes and cultural figures, from Klay Thompson and Paul George to Joe Burrow, Dak Prescott and Livvy Dunne. The financing arrives less than a year after Wispr's $25 million round in November, underscoring how quickly capital is redeploying into AI companies with real enterprise pull.
Flow reaches the Fortune 500 as usage crosses 60 billion words
Wispr says its dictation product Flow has now been used to generate more than 60 billion words and is deployed across nearly all Fortune 500 companies and more than 10,000 enterprises. Flow works on Mac, Windows, iPhone and Android with AI-powered auto-edits, tone matching, a command mode for AI-assisted editing and support for over 100 languages, integrating natively into tools like Cursor and Replit. The company follows a broader pattern seen in this year's enterprise AI rollouts, where distribution is proving to be as valuable as raw model performance.
Canto targets real-world conditions where transcription usually breaks
Alongside the funding, Wispr previewed Canto, a proprietary speech recognition model built for the noisy environments where consumer dictation typically collapses. CEO Tanay Kothari said Canto is trained for cars, streets, wind, music and open offices — conditions where competing systems see error rates above 30% but where Wispr's new model drops them to between 5% and 10%. Much of the $280 million will go into pushing the company's internal "zero edit rate" metric higher, so more dictated text can ship without correction. The bet echoes the broader shift toward vertically integrated stacks that recent AI infrastructure moves — like Google's Gemini 3.7 Flash and Z.ai's GLM-5.3 — have reinforced through 2026.
Reporting based on coverage from Reuters, TechCrunch, TechStartups and Wispr Flow's Series B announcement.