Deepgram was founded in 2015 by Scott Stephenson, Noah Shutty, and Adam Sypniewski in San Francisco. The founders, who met while working on dark matter detection at the University of Michigan, applied their signal processing expertise to speech recognition.
Deepgram raised over $86 million, including a $47 million Series B led by Tiger Global in 2022. The company has also received funding from Madrona, Wing VC, and Y Combinator.
Unlike ASR providers that fine-tune open-source models, Deepgram trains its own end-to-end deep learning models from scratch. Their Nova-2 model delivers industry-leading accuracy across multiple languages and domains, with significantly lower latency than alternatives — critical for real-time applications like live captioning and voice agents.
The company offers both speech-to-text and text-to-speech APIs, along with audio intelligence features like topic detection, sentiment analysis, and intent recognition. Deepgram’s APIs support streaming and batch processing, with pre-built integrations for popular platforms.
Deepgram serves enterprise customers including NASA, Spotify, and several Fortune 500 companies. The platform processes billions of minutes of audio, with particular strength in noisy environments, multi-speaker scenarios, and industry-specific vocabularies. The company employs around 100 people and competes in the rapidly growing voice AI market.
