- The AI Voice Newsletter
- Posts
- Telnyx Unveils LiveKit on Telnyx
Telnyx Unveils LiveKit on Telnyx
🔊 Soundcheck
Voice AI scales better, cheaper—Telnyx owns the stack
Gnani.ai Nets $10M for Global Voice AI
Retell AI Snags VC Enterprise Tech 30 Spot
Miravoice Powers Phone Surveys with AI
Read time: 5 minutes
🔥 Hot Mic
Big moves, deep dives, and standout stories.
Telnyx now offers fully hosted LiveKit infrastructure that cuts costs, slashes latency, and delivers carrier-grade voice AI capabilities.
Telnyx just launched LiveKit on Telnyx, a fully hosted platform that lets developers deploy Voice AI agents using the familiar LiveKit framework, but with Telnyx’s own carrier-grade infrastructure powering everything. This means agents run on GPU-powered systems right alongside the telephony hardware, so audio never leaves the network and performance stays blisteringly fast.
The economics change too—during beta, session fees are gone, and speech services like STT and TTS cost up to 50% less than LiveKit Cloud equivalents. On top of that, key enterprise features—like HD voice codecs, caller ID attestation, call transfers, and global SIP capabilities—are built in from the start. Migration is zero-friction: your code stays the same, you package your agent with Docker, deploy via API, point your number, and you're live.
Key Points:
Developer code remains unchanged—uses identical LiveKit framework and APIs.
Beta waiver of session fees eliminates major LiveKit Cloud cost pressure.
STT/TTS costs slashed by roughly 50% compared to LiveKit Cloud.
Sub‑200 ms round‑trip latency via GPU inference colocated with telephony.
Full enterprise telephony: HD codecs, SIP, STIR/SHAKEN, transfers native.
Telnyx controls infrastructure stack—from carrier network to AI compute.
Takeaway: LiveKit on Telnyx represents a smart shift from proof‑of‑concept to production—by owning the stack, Telnyx delivers more reliable, lower‑cost Voice AI infrastructure with enterprise telephony built in, all while letting developers keep building with the tools they already know.
Gnani.ai raised $10 million in Series B funding to fuel global expansion, R&D, and multilingual voice AI.
Gnani.ai, a Bengaluru-based voice‑first AI startup, just closed a $10 million Series B round led by Aavishkaar Capital alongside Info Edge Ventures. The funding will be used to bolster its global reach, deepen research efforts, and scale its multilingual, enterprise‑grade AI stack.
The startup processes over 30 million voice interactions a day across 12+ languages and serves more than 200 enterprise clients, including Fortune 500 firms. Gnani.ai develops its full AI stack in‑house—spanning speech‑to‑text, text‑to‑speech, speech‑to‑speech, and agentic AI—to deliver human‑centric customer automation worldwide.
Key Points:
Raised $10 million in Series B led by Aavishkaar, with Info Edge Ventures
Handles over 30 million daily voice interactions in 12+ languages
Delivers full in‑house AI stack: STT, TTS, speech‑to‑speech models
Serving 200+ enterprises including Fortune 500 clients
Takeaway: This funding marks a milestone in bringing sovereign, multilingual voice AI to enterprise workflows globally—highlighting Gnani.ai’s technological depth and its vision for voice as a core automation layer across industries.
Startup Retell AI has been selected for Wing Venture Capital’s 2026 Enterprise Tech 30 in the early-stage category, marking it out as a voice AI standout. Its platform now handles over 50 million real-time AI phone calls each month and has hit $50 million in annual recurring revenue just a year after launching in 2024. The company positions its infrastructure as a modern alternative to legacy IVR systems, enabling quick deployment of AI voice agents across voice, chat, email and SMS. CEO Bing Wu notes enterprise pilots almost always progress into full production, driven by marked ROI, faster issue resolution and improved customer satisfaction.
Key Points:
Selected in Wing VC’s 2026 Enterprise Tech 30 early-stage list
Handles over 50 million real-time AI phone calls monthly
Reached $50 million annual recurring revenue within one year
Replaces traditional IVR with AI agents deployable in days
Takeaway: Retell AI’s rapid scaling—from 2024 launch to handling tens of millions of calls and generating significant revenue—demonstrates that enterprise-grade voice AI is moving decisively from pilot to production in a very short time.
Miravoice is reimagining phone surveys. With $6.3 million in seed funding, it’s building AI voice agents that can run lengthy, highly structured phone interviews—quickly, affordably, and without coding.
These agents follow strict scripts and branching logic, conduct multilingual, large-scale campaigns, and handle hundreds of questions per call. With this funding, Miravoice aims to scale infrastructure and extend the reach of what’s traditionally been a manual and costly process.
By automating phone-based market research and polling, Miravoice brings survey capabilities closer to the simplicity of building an online form. Teams can launch campaigns across languages and zones without engineers, cutting cost and time dramatically.
This shift opens new possibilities for accessing high-quality, representative voice data. Smaller organizations and researchers gain tools once limited to well-funded operations, with AI agents offering speed, consistency, and broad reach.
Key Points:
Raised $6.3M seed funding led by Unusual Ventures
AI voice agents conduct long, rule‑based phone surveys
No‑code setup—launch complex surveys without engineers
Reduces cost and time compared to human call centers
Supports multilingual campaigns and large-scale calls
Takeaway: Miravoice’s funding marks a turning point: AI voice agents are breaking into structured phone research, turning what once took weeks and budgets into fast, scalable, no‑code surveys—making rigorous voice‑based insights accessible to many more organizations.
🎙️ Mic Drop
What else is making noise in voice AI.
watsonx Orchestrate adds multilingual voice AI and enterprise controls via partnership with IBM and ElevenLabs. (eenewseurope.com)
Planet DDS introduces voice AI charting, enabling hands-free dental exam documentation with voice commands in Denticon. (drbicuspid.com)
Serve Robotics debuts 'Maggie,' a delivery robot featuring conversational edge AI and 5G connectivity at NVIDIA GTC 2026. (seekingalpha.com)
Voice AI agents are increasingly deployed to manage simultaneous calls, reshaping the economics and quality of customer service. (computerweekly.com)
WellSaid, a text-to-speech provider, secures venture debt to fund enterprise sales and accelerate sector penetration. (slator.com)
Sinch launches Voice Relay, connecting text-based AI agents to live phone calls for real-time conversational use cases. (simplywall.st)
Twilio receives analyst upgrade citing voice AI as a key driver of future growth and product strategy. (proactiveinvestors.com)
Twilio stock rises after Jefferies cites voice AI as a core growth engine, boosting the price target. (tikr.com)
Folk musician faces unauthorized AI-generated covers on Spotify, spotlighting copyright and ethical challenges for voice synthesis. (techbuzz.ai)
The Better Business Bureau cautions about AI-powered voice scams impersonating loved ones to extort funds. (wistv.com)
Q01 smart glasses offer voice assistant integration for athletes, enabling hands-free control and voice-based interaction. (notebookcheck.net)
Guide explores TikTok’s AI-powered text-to-speech features, setup steps, and best practices for creators. (shopify.com)