Pomegra
Wispr's Series B - led by Menlo Ventures - triples its valuation in nine months as its proprietary speech model Canto cuts word error rates to under 10% in noisy environments.
- Wispr closed a $280M Series B on Aug. 17, 2026, valuing the AI dictation company at $2B, up from $700M in November 2025.
- Its Canto speech model reduces word error rates from over 30% to 5-10% in real-world noisy conditions, targeting legal, healthcare, and finance.
- Users of its Flow platform have generated over 60 billion words; adoption spans nearly all Fortune 500 companies and 10,000+ enterprises.
Lead
Wispr closed a $280 million Series B on August 17, 2026, valuing the San Francisco-based AI dictation startup at $2 billion - nearly three times the $700 million it commanded just nine months earlier. Menlo Ventures led the round. Existing backers Notable Capital, NEA, Neo Ventures, 8VC, and MVP Ventures returned, joined by new entrants Acrew, Forerunner, Goodwater, Peak XV, Together Fund, and PLUS Capital. The round brings total capital raised to $361 million.
What Did Wispr Actually Build?
Wispr Flow is an AI dictation layer that runs across Mac applications, letting users write anywhere - emails, documents, Slack - by speaking rather than typing. The company was founded in 2021 by Tanay Kothari and Sahaj Garg, both former Apple and Meta engineers, who initially built a brain-computer interface wearable for silent mouthed typing. After connecting that prototype to large language models, they shelved the hardware and shipped Flow as a software product in late 2024. The pivot was fast; the growth was faster.
Between May 2025 and May 2026, Wispr reported 150x revenue growth and 200x user growth while headcount expanded from 7 to 60. Those numbers are self-reported and unaudited, but the investor appetite at a $2 billion valuation reflects external conviction that the trajectory is real, not a rounding error.
Why Does the Valuation Jump Matter?
The math here deserves a second look. Wispr was valued at $700 million in November 2025 when it raised a $25 million Series A extension. Nine months later, it is priced at $2 billion on $280 million of fresh capital. That is a 186% markup at a time when most enterprise AI valuations have been under pressure to show revenue multiples that justify their prior marks.
The implication: investors are pricing in something beyond current revenue. That something is Canto.
What Is the Canto Speech Model?
Canto is Wispr's first proprietary speech model, previewed alongside the Series B announcement. Its claim is specificity - the model was trained on messy, real-world audio conditions, including background noise, wind, accents, barking dogs, and overlapping voices. In those noisy environments, standard speech-to-text systems produce word error rates above 30%. Canto, Wispr says, brings that figure to between 5% and 10%.
The company also projects that Canto will reduce the share of dictations requiring manual corrections by 30% to 35%. That number matters more to enterprise buyers than headline accuracy figures. Fewer corrections means lower friction, and lower friction means more consistent adoption across verticals like legal, healthcare, and finance - fields where transcription errors carry compliance and liability risk, not just inconvenience.
Training and deploying a proprietary model also changes Wispr's competitive positioning. Vendors that rely on third-party transcription APIs - OpenAI Whisper, Google Speech-to-Text - face a structural ceiling on differentiation. Building in-house shifts the moat from product design to model quality, which is harder to replicate quickly.
How Does This Change Wispr's Market?
The company has been explicit that dictation is the entry point, not the destination. It recently launched a meeting note-taker, moving it into direct competition with established players in that segment. The $280 million gives Wispr the runway to staff up product and go-to-market functions for enterprise sales cycles, which run longer and cost more than the consumer-focused growth that got it to 60 billion words written on Flow.
Menlo Ventures framed its thesis around the death of the text box - the idea that keyboard-based text input will cede ground to voice across more workflows. That is a large claim supported by thin current evidence but consistent with long-running adoption trends in mobile voice search and smart speaker usage.
Outlook
Wispr enters its next phase with a model it controls, a platform with documented Fortune 500 penetration, and enough capital to pursue the enterprise market seriously. The near-term test is whether Canto's accuracy gains hold at scale and translate to lower churn and larger contract values. A $2 billion valuation at this stage of revenue means execution risk is high and the tolerance for a missed quarter is low.
