LiveKit has raised $100M in a Series C funding round, reaching a $1B valuation as it accelerates development of infrastructure for real-time, voice-first applications.
The round was led by Index Ventures, with participation from Salesforce Ventures, Hanabi Capital, and existing investors Altimeter and Redpoint Ventures.
Voice AI moves from experiment to production
Voice is emerging as a primary interface for interacting with software, moving beyond experiments into real-world deployments across industries. Enterprises are increasingly adopting voice agents to automate workflows, handle customer interactions, and unlock new service models.
While many voice AI use cases are still early, some are already operating at scale, supporting customer service, sales, and operational tasks for global brands. LiveKit expects 2026 to mark a broad rollout of voice AI across thousands of production use cases worldwide.
Advertisement
Why voice applications need a new stack
Voice-driven applications differ fundamentally from traditional web software. Unlike stateless web requests, voice interactions are real-time and stateful, often lasting minutes or hours while continuously maintaining conversational context.
This shift breaks conventional assumptions around how applications are built, tested, deployed, and monitored. Voice-native software requires infrastructure designed specifically for low-latency communication, persistent state, and continuous orchestration across multiple AI models.
LiveKit is building a full-stack platform designed from the ground up for these requirements.
Building voice agents at scale
LiveKit provides SDKs across platforms to support voice application frontends, while its agent framework enables backend orchestration, model integration, and real-time conversational dynamics such as turn-taking and interruptions.
For teams that prefer low-code workflows, LiveKit also offers tools to prototype, configure, and share voice agents without starting from scratch in code.
Testing, deployment, and real-time operations
Because AI systems are non-deterministic, testing voice agents requires statistical evaluation rather than simple pass/fail assertions. LiveKit supports unit testing, tracing, and simulation workflows to help teams evaluate agent behavior at scale.
On the deployment side, voice agents face unpredictable session lengths and traffic spikes, requiring different approaches to capacity planning and failover. LiveKit has introduced serverless deployment options and built a global real-time network optimized for voice and video routing, handling billions of calls annually.
The platform also integrates inference routing, abstracting complexity across multiple model providers and colocating models with deployed agents to reduce latency and improve reliability.
Observability for human-AI conversations
LiveKit has developed observability tools tailored specifically for voice agents, enabling teams to replay sessions, inspect transcripts, analyze latency, and understand how users and agents interact in real time.
This feedback loop allows developers to continuously refine agent behavior, improving reliability, accuracy, and user experience with each iteration.
Building the foundation for voice-first software
LiveKit positions voice as one of the largest paradigm shifts in computing, particularly in environments where voice is already dominant, such as phone calls, vehicles, and smart devices.
As models improve and new form factors emerge, voice-native applications are expected to move from novelty to default. LiveKit aims to provide the development and runtime layer that sits between foundation models and end-user applications, making voice AI as scalable and accessible as web development.
The new funding will be used to accelerate product development, expand infrastructure, and support the growing community building voice-first applications worldwide.
