According to TechCrunch, voice AI company Wispr has secured $25 million in new funding led by Notable Capital with participation from Steven Bartlett’s Flight Fund, just months after raising $30 million in June. The company’s dictation app, Wispr Flow, has grown 40% month-over-month since June and now counts 270 Fortune 500 companies as users, with 125 signed as enterprise customers. After three months of usage, average users write more than 50% of their characters through the app, and the company maintains 70% retention over 12 months. Notable Capital’s Hans Tung, who backed companies like Airbnb, Slack, and Anthropic, is joining Wispr’s board as an observer. With this round, Wispr has raised $81 million total and plans to expand internationally and build its own voice models.
The VC feeding frenzy
Here’s the thing that caught my attention: Wispr wasn’t even planning to raise more money. They had just closed a $30 million round in June and were sitting on what CEO Tanay Kothari called a “really long runway” with a lean team. But when Notable Capital’s Hans Tung and Steven Bartlett came knocking, they couldn’t say no. That tells you something about the current AI gold rush – when top-tier investors want in, you make room for them even when you don’t need the cash.
What’s particularly interesting is how Notable Capital approached this. They didn’t just show up with a term sheet – they’d done deep research, interviewed competitors, and built a strong investment case. In today’s market, that level of preparation from VCs is becoming more common for hot AI companies. They’re essentially doing their due diligence before even getting the meeting.
More than just dictation
Wispr isn’t trying to be just another dictation tool. They’re aiming for what Tung calls a “voice-led operating system” that can automate workflows. Think beyond transcribing your thoughts – they want to handle tasks like replying to emails automatically. That’s a much bigger vision than what competitors like Aqua, Monologue, or Typeless are targeting.
Their current error rate of around 10% is already beating OpenAI’s Whisper (27%) and Apple’s native transcription (47%), which is impressive. But the real challenge will be reducing that even further while expanding to new platforms. They’re working on an Android app with beta coming by year-end and stable launch in Q1 2025, which should significantly expand their reach.
The user adoption challenge
Even with all this growth, Wispr hit a familiar wall: when non-technical users discovered the app, they’d install it, try the dictation feature within the app itself, and then drop off. The problem? They didn’t realize they could use dictation across all their other apps. Basically, the value proposition wasn’t immediately obvious to regular people.
So Wispr had to create a new onboarding flow to guide users toward using dictation in the apps they actually use daily. It’s a classic case of building something powerful but needing to educate users about how to integrate it into their workflow. This is where many productivity tools fail – the technology might be great, but if people don’t understand how to use it in their daily lives, adoption stalls.
Where this is headed
With this new funding, Wispr plans to hire top machine learning talent that might otherwise go to OpenAI or Anthropic. They’re also building their own voice models for personalized Automatic Speech Recognition and testing a closed API with select enterprises. The goal is to open that API to more developers next year.
But here’s the billion-dollar question: can voice really become the primary interface for computing? We’ve been hearing about voice-controlled everything for years, but outside of simple commands to smart speakers, it hasn’t really taken off for complex tasks. Wispr’s traction suggests maybe we’re finally reaching that tipping point where voice becomes genuinely useful for real work. The fact that they’re reaching Fortune 500 companies and seeing users actually relying on their app for the majority of their writing is pretty compelling evidence.
