이미지: TechCrunch AI 화면 갈무리
Summary
- AI dictation tool Wispr raised a $280 million Series B led by Menlo Ventures at a $2 billion valuation, bringing total funding to $361 million
- The company also unveiled a new model called Canto, which it says cuts transcription error rates from 30% to under 10%
- As competition among dictation apps intensifies, Wispr is expanding into meeting note-takers, ring-style wearable integration, and a new interface research lab
- 시리즈B 투자액
- 2억8000만달러
- 밸류에이션
- 20억달러
- 누적 투자액
- 3억6100만달러
- 리드 투자자
- Menlo Ventures
- 새 모델
- 칸토(Canto), 오류율 30%→10% 미만
- 인터페이스랩 리드
- Ariya Rastrow (초기 아마존 알렉사 개발 참여)
- 직전 라운드 시점
- 10개월 이내
Error Rate Cut From 30% to Under 10%
Over the past few weeks, users of Wispr's dictation app "Wispr Flow" had complained that recognition quality had declined. Wispr responded to the issue with a new model called Canto. The company said adopting Canto cut its error rate from 30% to under 10%. In numerical terms, that's a reduction of errors to less than a third of the previous level.
| Category | Error Rate | Ratio |
|---|---|---|
| Previous model | 30% | 100 |
| After Canto | Under 10% | 33 |
Alongside the model launch, Wispr also announced funding news. The company raised $280 million in a Series B led by Menlo Ventures, which valued the company at $2 billion. This is the second major round in the past 10 months, bringing cumulative funding to $361 million.
Who Put Money In
Many existing investors returned to increase their stakes in this round, and new investors also joined.
| Category | Investors |
|---|---|
| Lead investor | Menlo Ventures |
| Existing investors (increased stake) | Notable Capital, NEA, Neo Ventures, 8VC, MVP Ventures |
| New investors | Acrew, Forerunner, Goodwater, Peak XV, Together Fund, PLUS Capital |
Behind Wispr's rising valuation is intensifying competition in the dictation market. Apps like Willow, Monologue, Aqua, and Superwhisper are targeting the same market, and on top of that, developers keep releasing free or low-cost dictation tools one after another. It's becoming increasingly difficult to differentiate on dictation alone.
Beyond Dictation, Into Meeting Notes
Since last November, Wispr has released an Android app and built out sales teams in regions including India and the UK. Alongside this regional expansion, the new funding will go toward its recently launched meeting note-taker product. This product extracts meeting summaries and to-do lists, competing with existing meeting-notes tools like Granola, Fireflies, and Read AI. However, features that would integrate with other work tools to automatically draft documents or emails remain an area for future expansion.
There's also movement on the hardware side. Wispr is partnering with wearable makers such as Oasis ring to let users dictate to devices without speaking aloud. As noted in a recent article on voice-ring startup Sandbar, the AI wearables market has been so littered with failed products that it's sometimes called a "graveyard." Wispr's choice to partner with existing wearable makers rather than build its own hardware appears to reflect this harsh market reality.
A New Lab Led by a Former Alexa Developer
Last month, Wispr announced the formation of a new organization called "Wispr Interface Labs." It is led by Ariya Rastrow, who was involved in early development of Amazon Alexa. The company said the lab will explore new interfaces for human-computer interaction. Having started as a single dictation app, the company appears to be broadening its scope beyond voice input to human-device interfaces as a whole.
Editor's Take
What stands out in this round isn't the size of the investment itself, but what Wispr says it plans to do with the money. Dictation apps have become a low-barrier-to-entry market. With free tools and cheap prosumer apps flooding in, it's hard to sustain a $2 billion company on dictation alone. So Wispr's simultaneous push into meeting note-takers, wearable integration, and an interface lab is better read as diversification for survival rather than defensive expansion.
Anyone who has actually used a dictation tool will find this pattern familiar. In the early days, all that mattered was accuracy, but at some point users start asking "what more can I do with this transcribed text?" Wispr is now facing the same question that Granola and Fireflies confronted first in the meeting-notes market. Cutting the error rate from 30% to under 10% is just getting the basics right — the real competition is likely to be decided by how smoothly that text flows into drafted documents or emails.
For teams in Korea considering adopting meeting-notes or dictation tools, it's more practical right now to examine the "post-summary workflow" rather than error-rate figures. Before settling on a tool, it's worth testing whether the output automatically connects to other work tools or requires manual rework — that upfront check saves time later. The next thing to watch in this market is whether Wispr's meeting note-taker adds integration features with other collaboration tools in the coming weeks, and whether competitors move at a similar pace to expand into interfaces.
