Deepgram Acquires Daily to Expand Voice AI Infrastructure
· GlobeWire Daily Editorial Desk · Source: techcrunch.com

Image: Ivan Mehta · Photo source · Resized and letterboxed to 16:9; no AI generation
Voice AI Consolidation Continues
Voice AI infrastructure company Deepgram has announced the acquisition of Daily, the creator of the open-source voice AI framework Pipecat. While financial terms of the deal were not publicly disclosed, leadership from both organizations confirmed that Pipecat will remain an open-source project.
Pipecat is designed to manage real-time voice and video streams while reducing latency across speech-to-text (STT), large language models (LLMs), and text-to-speech (TTS) components. Daily previously raised over $62 million in funding, while Deepgram reached unicorn status earlier this year following a $130 million Series C funding round. Both companies originally participated in the same Y Combinator batch in 2016.
Leadership Perspectives and Open-Source Commitment
Daily co-founder Kwindla Kramer stated that the companies evaluated various acquisition offers before agreeing to terms with Deepgram. Kramer highlighted Deepgram CEO Scott Stephenson's commitment to keeping Pipecat vendor-neutral and open source as a primary deciding factor.
- Vendor Neutrality: Stephenson emphasized that developers should be able to integrate models from various partners rather than being restricted to a single ecosystem.
- Ecosystem Growth: Leadership noted that maintaining an open architecture provides commercial advantages while fostering broader industry adoption.
Introduction of the Real-Time Context Standard
Alongside the acquisition announcement, Deepgram introduced the Real-Time Context Standard (RCS), an open protocol co-developed with the Pipecat team. The protocol is designed to handle metadata such as audio context, instructions for language models, and tonal guidance for agent responses.
Practical Implications and Verification
- Developer Flexibility: The integration aims to give developers tools to construct end-to-end speech systems using mixed model stacks.
- Verification: Interested parties can review official announcements and technical documentation directly through Deepgram and Daily's official corporate channels or press releases.
AI-assisted reporting; source claims are attributed and have not been independently verified.
More to read
Related coverage
Related reporting available at publication.
123NET Awarded $86 Million in Federal BEAD Funding to Expand Michigan Fiber Infrastructure
NIIT Learning Systems Acquires Bluewater Learning to Scale AI-Driven Talent Transformation
Google Launches Connected Apps Platform to Expand AI Agent Integration
Tech Stocks Slip as AI Infrastructure Spending Surges and PC Demand Softens
Data Security Council Launches India-Specific Guide on Agentic AI Liability
US Banks Block Third-Party AI Agents as Open Banking Rule Stalls