Wispr AI's $280M Series B at $2B Valuation

Wispr didn't start out building a dictation app. Founded in 2021 by Tanay Kothari and Sahaj Garg, the company spent its first two years developing a non-invasive neural wristband meant to translate silent speech into text, before pivoting to software in 2023 ^[1]and shipping its first Mac app, Flow, in October 2024 ^[2]. Windows and iOS versions followed through 2025, and Wispr's fundraising accelerated in lockstep with the product's traction: a $30 million Series A led by Menlo Ventures in June 2025 ^[2], a $25 million add-on from Notable Capital five months later aimed at building what the company called a 'voice operating system' ^[3], and now, less than 10 months after that round, a $280 million Series B at a $2 billion valuation that lifts total funding to $361 million ^[4]. Going from an $81 million cumulative raise as of November 2025 to a $2 billion valuation nine months later tracks a company that pivoted twice, first from hardware to software, and now from a single dictation feature to a broader platform bet.

Wispr's pitch to investors goes well beyond cleaning up dictation. Menlo Ventures partner Matt Kraning, who joined Wispr's board after the firm's Series A investment, argues the constraint in AI has moved: 'The bottleneck in AI has moved from the model to the human interface, and Wispr is the interface' ^[5]. He frames dictation as just an entry point: 'Dictation is how you get in the door. What people pay for is not having to type, which puts you up against workflow tools, meeting tools, and eventually the text box in front of every AI model' ^[6]. CEO Tanay Kothari has echoed that framing, calling dictation 'the starting point for something bigger' with voice meant to become 'the foundational layer' of computing ^[5]. To back that ambition, Wispr is pushing into a meeting notetaker that competes directly with Granola, Fireflies, and Read AI, and has stood up Wispr Interface Labs, a research group led by former Amazon Alexa engineer Ariya Rastrow to design new human-computer interfaces beyond the keyboard ^[4]. Rastrow puts the underlying diagnosis bluntly: 'AI is an interface problem now' ^[7]. Reaction on X leaned into that same framing, with investors treating the round as validation of the interface thesis rather than as an isolated dictation story, grouping it alongside other large same-week AI funding news.

The technical centerpiece of the round is Canto, Wispr's proprietary speech-recognition model, unveiled the same day as the funding to address a specific complaint: dictation accuracy degrading in noisy, real-world settings ^[4]. In especially challenging acoustic conditions, Canto is designed to cut word error rates from more than 30% down to roughly 5-10%, which the company expects will translate into about 30-35% fewer dictations needing manual editing ^[8]. That reliability push matters because Flow's user base has moved well past a novelty tool: the app has been used to write more than 60 billion words ^[5], and Wispr now counts roughly 100,000-125,000 business customers, including nearly all Fortune 500 companies ^[8]. Revenue has grown more than 150% per quarter for four consecutive quarters, the kind of growth that underpins the jump to a $2 billion valuation less than a year after a $30 million Series A.

The expansion raises the stakes on two fronts at once. Competitively, Wispr is fighting a more crowded dictation market that now includes Willow, Monologue, Aqua, and Superwhisper, while its notetaker push adds a more crowded product surface on top ^[4]. On trust, co-founder Sahaj Garg has framed privacy and security as existential to the company's ambient ambitions rather than optional add-ons: 'Everything about these tools is about trust. Privacy and security aren't features for us: They're the foundation for earning ambient access to your life' ^[6]. That framing matters more with each step Wispr takes from transcribing what a user dictates to recording entire meetings and, eventually, building always-on interfaces, since each step widens the amount of a user's day the company can hear.