Pomegra
Voice AI startup Wispr closed a $280 million Series B at a $2 billion valuation led by Menlo Ventures, expanding from dictation into meetings and workplace productivity tools.
- Wispr's Series B brings total capital raised to $361 million, with most of that money arriving in under 14 months.
- New proprietary speech model Canto targets word error rates of 5-10%, down from over 30% in noisy real-world settings.
- A meeting note-taker puts Wispr in direct competition with Granola, Fireflies.ai, and Read AI.
Lead
Wispr closed its $280 million Series B on August 17, 2026, with Menlo Ventures leading at a post-money valuation of $2 billion. New backers Acrew, Forerunner, Goodwater, Peak XV, Together Fund, and PLUS Capital joined the round, while existing investors Notable Capital, NEA, Neo Ventures, 8VC, and MVP Ventures all increased their positions. Total capital raised since founding stands at $361 million.
The fundraising pace is hard to ignore. Wispr closed a $25 million round with Notable Capital in November 2025, then a $30 million Series A with Menlo Ventures in June 2025 - less than 14 months separate that Series A from a Series B nearly ten times its size.
What Does Wispr Flow Actually Do?
Wispr Flow is an AI dictation tool that runs across any application on Mac, Windows, and iOS, transcribing speech while removing filler words and formatting text automatically. The product does not require users to break their existing workflow. Tanay Kothari and Sahaj Garg, Stanford roommates who co-founded the company in 2021, originally built voice software to test a non-invasive wearable device; when the software showed clearer product-market fit than the hardware, they pivoted in mid-2024. The app launched on Product Hunt that year, hit number one for the day and the week, and has since reported 40% month-over-month user growth.
Why Does a Dictation App Command a $2B Valuation?
The valuation implies Wispr has moved past the dictation category, at least in investor framing. Menlo Ventures has argued that the text box is dying - that voice input will displace keyboard entry across more of daily computing. That thesis is coherent in the abstract. The harder question is whether a startup owns that transition or whether it eventually collapses into an OS feature. Apple, Google, and Microsoft control the microphone and the default voice layer on every major platform.
What Wispr brings is cross-platform distribution with an active user base dictating across real work contexts. Canto, the company's first proprietary speech-recognition model, is meant to extend that advantage by cutting error rates from 30%-plus to under 10% in noisy environments - conditions the training benchmarks for most commercial speech models do not capture.
How Does Canto Change the Competitive Picture?
Prior versions of Flow depended on third-party speech recognition. Canto gives Wispr direct control over accuracy and latency without paying API costs or accepting model constraints set by another vendor. In practical terms, the gap between 30% and 8% word error rates in open offices or on active calls is the gap between a tool professionals adopt and one they tolerate for a month and uninstall. Canto is the company's argument that it is not a thin wrapper on existing infrastructure.
What Competition Does Wispr Face as It Moves Into Meetings?
The new meeting note-taker puts Wispr against Granola, Fireflies.ai, and Read AI in a segment that is crowded and where enterprise sales motions are already established. Wispr's argument is that users who already run Flow across their devices are a natural install base. The strategic risk is that meeting notes is increasingly a feature, not a standalone product category, and that Microsoft Copilot and Google Workspace are not absent from the conversation.
Wispr Interface Labs, a research division announced alongside the funding, is exploring hardware partnerships and new voice-interface paradigms. The original wearable-hardware instinct has not left the company.
The Athlete-Investor List
The round includes a notable celebrity cohort via PLUS Capital: Shaun White, Dak Prescott, Joe Burrow, Klay Thompson, Paul George, DK Metcalf, Kyle Hamilton, Alex Caruso, Domantas Sabonis, Aaron Gordon, Trae Young, and gymnast Livvy Dunne. Celebrity participation does not change unit economics, but it signals a consumer brand ambition running in parallel with the enterprise push.
Outlook
Wispr enters late 2026 with capital to build out the enterprise meeting tool and a proprietary model stack. The $2 billion valuation puts real pressure on the next milestone - Menlo's text-box thesis needs a product that proves voice can replace input, not just supplement it. Canto addresses the accuracy complaint. The meeting note-taker tests whether dictation distribution converts into a second product category. What remains unresolved is whether a voice layer controlled by a startup can hold ground once platform incumbents decide the interface is worth owning.
