Table of Contents
🌏 中文版
Funding Details
| Item | Value |
|---|---|
| Company | Wispr (San Francisco, USA) |
| Round | Series B |
| Amount | $280M |
| Lead | Menlo Ventures |
| Participants | Notable Capital, NEA, Neo Ventures, 8VC, MVP Ventures (existing investors); Acrew, Forerunner, Goodwater, Peak XV, Together Fund, PLUS Capital (new investors) |
| Valuation | $2B (previous round: Series A2 in November 2025 at $700M, ~2.9x growth) |
| Total raised | $361M |
| Founded | 2021 |
| Headcount | ~50 (GetLatka, October 2025 data) |
What the Company Does
Wispr builds AI-powered voice dictation — replacing typing with speech, letting users input text by talking in any application.
Its core product Flow covers 162 countries and 100+ languages, claiming to be 4x faster than typing while automatically adjusting tone, removing filler words, and adapting formatting to context. Alongside this round, Wispr released Canto, a proprietary voice model that cuts recognition error rates in noisy environments from 30%+ down to 5–10%, addressing recent user complaints about declining transcription quality. The company also established the Advanced Interfaces Lab in July, led by Chief Scientist Ariya Rastrow (formerly involved in early Amazon Alexa development), with the goal of evolving from "voice-to-text" to "voice-to-outcome."
Flow is already used by employees at most Fortune 500 companies and deployed across over 125,000 enterprises, with revenue growing 150%+ quarter-over-quarter for the past four quarters. Wispr is also expanding into meeting notes (a notetaker competing with Granola, Fireflies, and Read AI) and wearables (partnering with Oasis smart ring), aiming to grow from a single dictation feature into a full voice interface platform.
What This Funding Signals
Implications for the Agent Ecosystem
Menlo Ventures lead partner Matt Kraning's thesis is straightforward: frontier models have solved the "intelligence" problem, but not the interface problem of "how ordinary people convey intent to models" — the text box itself is a tax on thinking. Wispr aims to be the layer connecting agents to human intent: an engineer controlling a coding agent by voice, a sales rep finishing a call and having the CRM update automatically — all applications of the same "voice-to-outcome" logic. The funds will go toward expanding the enterprise sales team, deepening the enterprise product line, and enabling the Advanced Interfaces Lab to incorporate intent channels beyond voice (e.g., what you're looking at, what you're working on).
What Investors Are Betting On
Menlo first invested in Wispr 14 months ago with a public bet that "voice will kill the keyboard." This follow-on $280M round is "one of the largest investments the firm has ever made in an AI company." Kraning's logic: voice dictation itself can be replaced by free features (OS-level dictation from Apple/Google/Microsoft), but "zero-correction, works-everywhere" plus "executes your intent rather than just transcribing" requires a company's worth of systems integration — not something an OS checkbox can deliver. Menlo previously invested in infrastructure-layer AI companies like Anthropic; this bet positions Wispr as "the voice-layer OS interface," not just another dictation app.
Numbers Worth Watching
- Valuation jumped from $700M in November 2025 to $2B — nearly 3x growth in 9–10 months, faster than most AI companies at a comparable stage;
- Of the $361M total raised, $280M (~78%) is concentrated in this single round, suggesting Wispr chose to take one large raise rather than multiple smaller rounds while riding 150%+ quarterly revenue growth;
- The new Canto voice model claims to cut error rates in challenging acoustic environments (background noise, wind, accents) from 30% to 5–10% — a direct product response to recent user quality complaints. Whether it delivers will show in retention numbers going forward.
Watchlist Status
Wispr is not yet on the watchlist. Recommend adding to section C5 (Voice AI / Voice Agent), tracked alongside Vapi, ElevenLabs, and Cartesia. Key tracking points: voice input interface + Advanced Interfaces Lab's "voice-to-outcome" roadmap, $280M Series B, $2B valuation.
Today's Takeaway
I originally assumed "AI dictation" was a red ocean carved up by iOS/Android built-in features, but Menlo's back-to-back heavy bets prove otherwise: the battlefield isn't "who has the highest recognition accuracy" — it's "who can turn voice into an intent channel that triggers real actions." Dictation is just the foot in the door to earn user trust and data; the real moat is the execution layer behind it that translates "what you said" into "what the system should do."
References
Loading...