Wispr has raised $280 million in Series B funding at a $2 billion valuation, as the AI voice company looks to expand beyond dictation and into meetings, enterprise workflows and new forms of human-computer interaction.
Menlo Ventures led the round, with returning investors including Notable Capital, NEA, Neo Ventures, 8VC and MVP Ventures. New backers include Acrew, Activate, Forerunner, Goodwater, Peak XV, Together Fund and PLUS Capital.
The funding brings Wispr’s total capital raised to $361 million.
The company is best known for Wispr Flow, a cross-platform voice dictation product that converts spoken input into formatted text across applications. Alongside the funding news, Wispr announced Canto, its first proprietary speech model, which it says has been built to improve transcription accuracy in real-world conditions, including noisy environments, heavy accents and multilingual speech.
Posting on X, Deedy Das, Principal, Menlo Ventures said: “Winning on the app layer is difficult: software is easier to replicate than ever before (what’s the moat? Can’t I just vibecode this?) so you have to maniacally focus on product quality, growth, retention and monetization while scaling your infrastructure, spinning up enterprise sales and growing a world class team. It is much easier said than done.”
“Products that seem simple on the surface like voice dictation or video creation hide a ton of this complexity under the hood.”
Canto Targets Real-World Voice Accuracy
Wispr said Canto can reduce word error rates from more than 30 percent to between 5 percent and 10 percent in its most difficult test conditions. It also claimed the model could reduce the number of everyday dictations requiring edits by 30 percent to 35 percent.
The company did not disclose its benchmarking methodology, including the languages, audio conditions or comparison models behind those figures.
Wispr is placing accuracy at the centre of its growth plans as it seeks to make voice a more viable input method for professional communication. The company said users have written more than 60 billion words through Flow to date, and that the product is used by people at “almost all” Fortune 500 companies and at more than 10,000 enterprises.
Those figures do not clarify how many organisations are paying customers or have formally deployed the platform across their workforce.
The company’s focus on real-world conditions is significant for workplace use cases, where voice input is often affected by open offices, travel, inconsistent microphones, background noise and multilingual teams. Wispr also says Canto is designed to better handle code-switching, including users moving between languages within the same sentence.
Wispr Moves Into Meeting Intelligence
The new capital will also support Wispr’s push into the growing market for AI meeting tools. Earlier this month, the company launched Wispr Flow Notetaker, which produces meeting summaries and action items.
Meeting intelligence is a competitive category, with established providers including Fireflies, Read AI and Granola, while Zoom, Microsoft Teams and Google Meet continue to embed their own AI transcription, summarisation and workflow features.
Wispr has positioned its product as an alternative that can operate across meeting environments, rather than being tied to one collaboration platform. However, the company has yet to detail what integrations it will offer with enterprise systems such as Microsoft 365, Google Workspace, Slack and CRM platforms, or whether its meeting product will be able to take workflow actions beyond generating summaries.
Microsoft Teams is introducing controls that allow administrators to automatically block detected external AI bots from meetings, potentially creating a barrier for third-party meeting notetakers.
For UC and collaboration teams, those questions will be central. Meeting transcription tools increasingly need to demonstrate more than accurate notes: they must fit into established communication and business workflows while meeting expectations around governance, security, consent and data handling.
Enterprise Governance Will Be the Next Test
Wispr has not outlined in its funding announcement whether meeting recordings, transcripts or other customer data are used for model training, nor the retention, data residency and opt-out controls available to enterprise customers.
The company is also investing in its Advanced Interfaces Lab, led by Chief Scientist Ariya Rastrow, who previously helped develop Amazon Alexa. Wispr describes the lab’s goal as building systems that can understand intent, retain the context of a user’s work and turn voice input into outcomes rather than simply a block of text.
That ambition points to a wider shift in the market. AI voice tools are moving from transcription products towards assistants that can capture conversations, identify tasks and potentially initiate follow-up work across business applications.
Wispr’s $280 million raise gives it substantial resources to pursue that opportunity. Its next challenge will be proving that its voice technology can meet enterprise standards for accuracy, integration and governance, while differentiating from both specialist meeting AI vendors and the increasingly capable assistants built into the collaboration platforms organisations already use.
