Deepgram, which is building enterprise voice-recognition tech, raised a $47M Series B extension led by Madrona Venture Group, bringing its total funding to $86M
Context & Ripple Effects
Deepgram has been on a steady funding cadence in enterprise speech recognition: a $12M Series A led by Wing VC in 2020, then a $25M Series B led by Tiger Global in early 2021 for its custom speech models. This $47M extension, led by Madrona Venture Group rather than either prior lead, lifts total funding to $86M and signals that investors kept backing the category even as 2022 capital markets tightened.
The bet aged well: four years later Deepgram closed a $130M Series C at a $1.3B valuation, making this extension the bridge round between its early custom-model phase and its scale-up.
First-order effects
- Deepgram gains an extended runway to keep building enterprise voice-recognition tech without raising a full new round during a hostile market, with Madrona replacing Tiger Global and Wing VC as lead.
Second-order effects
- Rivals selling conversational AI into the same enterprise buyers — Kore.ai, which later raised a $150M round co-led by Nvidia and FTV Capital, and avatar-maker Deepbrain AI — face a better-capitalized competitor pushing voice infrastructure deeper into corporate stacks.
Third-order effects
- If the pattern holds, enterprise voice AI consolidates around a few heavily funded platform vendors, forcing smaller speech-tool builders to differentiate on vertical specialization or exit.
The trend: Enterprise voice-recognition startups are compounding through successive mega-rounds, turning speech APIs from niche tooling into core enterprise AI infrastructure.