U.S. startup Wispr has completed a $280 million funding round at a $2 billion valuation as interest in its AI-powered dictation platform continues to grow. The sum is well above the size of a typical Series B round, in line with the outsized checks that AI application startups have commanded from investors over the past two years.

The San Francisco-based company confirmed the Series B investment in a post on its website on Monday. The round follows a $30 million Series A raise last May, taking Wispr from an early-stage round to a $2 billion valuation in roughly a year.

Wispr’s core product is the dictation app Wispr Flow, which allows users to interact with mobile and desktop environments using speech instead of text. The company said the app stands out from similar tools because of its popularity. Flow enters a market in which Apple, Google and Microsoft already build dictation into their operating systems, and in which open-source speech models such as OpenAI’s Whisper have made high-quality transcription widely available to developers.

In the post, co-founder and CEO Tanay Kothari said that just over a year ago there was widespread skepticism about voice technology. He said: “That skepticism was natural. People had been trying voice tools for 15 years and getting the same result every time, so the reasonable assumption was that this one would disappoint them too.” That skepticism has a long history: commercial dictation software such as Dragon NaturallySpeaking has been sold since the 1990s, and voice assistants such as Siri and Alexa later normalized speaking to devices without replacing the keyboard as the main way people write.

According to the company, users now say Wispr Flow saves them significant time.

Wispr said the product goes beyond basic transcription by formatting users’ thoughts, learning their vocabulary and producing text that mirrors the way they speak. The company also said its global reach has helped drive adoption, with about 60% of the app’s output in non-English languages.

The company said it is planning further improvements, including a new proprietary speech model called Canto. Wispr said Canto is being developed to handle voices in difficult conditions, including background noise, music playing and strong local accents. Noisy environments and heavy accents have long been among the hardest problems in speech recognition, and how well Canto performs against them is one of the main things to watch as development progresses.

While Flow focuses on speech directed at a computer, Wispr has also recently released Notetaker, a tool designed for meetings with other people. The move places Wispr in the crowded field of AI meeting transcription and note-taking tools that has grown up around video calls.

Wispr said it will use the new funding to support further product development and continued accuracy improvements.

Menlo Ventures, a longtime investor, led the Series B round. Returning backers included Notable Capital, NEA, Neo Ventures, 8VC and MPV Ventures.

New investors in the round include Acrew Capital, Forerunner Ventures, the Together Fund and Goodwater Capital. Other investors include several celebrities and sports figures, among them Dallas Cowboys quarterback Dak Prescott, Pittsburgh Steelers wide receiver DK Metcalf and Baywatch gymnast Livvy Dunne. Their participation extends a pattern of athletes and entertainers taking equity stakes in AI startups that has become a recurring feature of recent funding rounds.