Deepgram announced this morning its acquisition of Daily, the company behind the open-source Pipecat framework, widely considered the most widely used orchestration system in the world for production voice agents. The value of the deal was not disclosed, but the transaction marks a milestone in the consolidation of the conversational AI infrastructure market.
Pipecat started as an internal tool at Daily, created to handle the technical challenges of keeping real-time voice conversations running — latency, interruption detection, switching between STT (speech-to-text), LLM, and TTS (text-to-speech). As voice agents evolved, the framework gained a massive developer community and became the de facto orchestration layer for teams that need to ship a voice bot into production, whether for customer service, medical scheduling, or personal assistants.
Scott Stephenson, CEO and co-founder of Deepgram, described the acquisition as a strategic step toward a complete voice AI platform. "Too many teams treat voice AI as a feature to add late in the roadmap. In reality, it's a distinct infrastructure discipline, with its own demands around latency, accuracy, and scale," Stephenson said. Deepgram, which now operates for more than 5,000 organizations globally and has processed more than 200,000 years of audio across its platform of 650,000 developers, will absorb the Daily team and use the company's global network of 75 points of presence as the backbone for voice transport infrastructure.
The point that matters most for the open-source community: Pipecat will remain open source under the BSD V2 license, with a promise of "forever." Kwindla "Kwin" Hultman Kramer, co-founder of Daily, will now take on the role of Head of Open Source at Deepgram. "Scott was super committed to Pipecat being open-source and vendor-neutral forever," Kramer told TechCrunch. "I think there's both a personal commitment and a commercial driver there. If you sit in the middle of an ecosystem that's growing very fast, and you have many parts of that ecosystem that are genuinely the best from your company, then everything being as open as possible is like a commercial advantage."
Beyond the acquisition, Deepgram launched the Real-Time Context Standard (RCS), an open protocol co-developed with the Pipecat team. RCS is designed to carry additional context alongside the audio stream — tone, volume, emotion signals — and instructions for the LLM on how to interpret those signals when generating a response. It's an ambitious attempt to standardize how voice agents perceive and react to human expressiveness in real time.
To understand the depth of this acquisition, it's useful to look at Daily's history. The company was founded with the goal of simplifying real-time audio and video infrastructure, tackling problems like optimized packet routing, compute cluster deployment across dozens of regions, and building client libraries that enabled embedding video calls into web applications. When the wave of voice agents began in 2024, Daily pivoted part of its focus to the voice AI stack, launching Pipecat and Pipecat Cloud. Pipecat Cloud reached general availability in mid-2025, with more than 1,000 teams already running in production.
Deepgram was already a unicorn after a $130 million Series C round in 2026, and the acquisition of Daily — which had raised more than $62 million from investors including Tiger Global, Highfield Capital, and TenOneTen — signals a phase of consolidation in the sector. Companies like ElevenLabs, Gradium, Fish Audio, LiveKit, and Vapi are also collecting significant rounds this year, competing for slices of the voice infrastructure market. With open-source frameworks like Pipecat at the forefront, the question is who will control the orchestration layer of the next cycle of conversational agents — and the answer probably isn't another isolated startup.
Sources: TechCrunch, Deepgram, Unite.AI
✓ Independent sources cross-checked and verified before publishing
