Newo Grabs $25M Series A for Voice AI Infrastructure

🔊 Soundcheck

• Newo raises $25M to power always-on AI reception

• Simple AI Raises $14M for Voice Agents

• Kani‑TTS‑2: Lightweight Voice Cloning TTS

• Voice AI Powers Safer Clinical Automation

Read time: 5 minutes

🔥 Hot Mic

Big moves, deep dives, and standout stories.

Newo Grabs $25M Series A for Voice AI Infrastructure

Newo raised $25M Series A to expand its always-on AI voice agent infrastructure for SMBs.

Newo, a San Francisco startup, secured a $25 million Series A led by Ratmir Timashev, boosting its total funding to about $32 million. The capital will be used to accelerate product development, grow partner distribution, and enhance AI voice capabilities for structured business workflows. Newo’s AI agents operate as always-on front desks, handling calls, SMS, chat, and WhatsApp across real-world environments.

Built around a ‘Zero‑Hallucination Architecture,’ Newo uses multiple agents to verify voice and text responses in parallel. That design ensures sub-second response times, human-like voice quality, and operational reliability. The platform has already generated over 15,000 agents through more than 200 certified partners, helping SMBs recover missed bookings and boost revenue.

Key Points:

• Series A of $25M led by Ratmir Timashev

• Total funding now approximately $32M

• AI handles calls, SMS, chat, WhatsApp 24/7

• Zero‑Hallucination Architecture ensures high accuracy

Takeaway: Newo is transforming voice with infrastructure-grade AI, enabling small businesses to recapture lost demand without adding staff, and positioning itself as a critical SMB operations layer.

Simple AI Raises $14M for Voice Agents

Simple AI raised $14 million in seed funding to enhance fast, personalized voice AI agents for sales calls.

Simple AI just closed a $14 million seed round led by First Harmonic, with backing from Y Combinator, Massive Tech Ventures, and True Ventures. The San Francisco–based team plans to use the funding to build custom generative voice models, analyze customer interactions, and deliver personalized, actionable analytics. Their voice agents automate inbound calls using clients’ full product catalogs, keeping system latency under 850 ms to enable smooth, natural-sounding conversations.

The startup targets voice-driven B2C workflows, from customer support to order placement, and offers tools to experiment with agent speed, accent, and personality. Early clients span industries from steak delivery and home insurance to self-storage. Simple AI claims its agents upsell up to 30% more often than human reps, positioning the startup as a unique solution to contact center challenges like seasonal spikes and inconsistent performance.

Key Points:

• Raised $14M seed round for voice AI agent development

• Led by First Harmonic, with YC, Massive Tech Ventures, True Ventures

• Supports fast, personalized inbound call automation under 850ms latency

• Agents ingest full product catalogs and deliver transcripts and insights

• Clients include Omaha Steaks, home insurance, self‑storage sectors

• Agents reportedly upsell up to 30% more than human reps

• Platform allows switching agent speed, accent, gender

• Founded by ex‑YC staffers Catheryn Li and Zach Kamran

Takeaway: Simple AI’s platform brings a fresh spin to voice commerce by integrating ultra‑low‑latency voice agents with customizable behavior and real‑time analytics—potentially redefining how businesses automate inbound calls in sales and support.

Kani‑TTS‑2: Lightweight Voice Cloning TTS

Kani‑TTS‑2 ushers in a lean, high‑performance era for open‑source TTS. Built on an “audio‑as‑language” approach, it uses LiquidAI’s LFM2 backbone and NVIDIA’s NanoCodec to convert speech into discrete tokens, enabling natural prosody without heavy compute. It was trained on a massive 10,000‑hour dataset in just six hours using eight H100 GPUs, demonstrating impressive training efficiency. Despite its compact size, it supports true zero‑shot voice cloning from a short audio clip, runs with a 0.2 real‑time factor (10 seconds of speech in about 2 seconds), and requires only ~3 GB VRAM, making studio‑quality voice synthesis feasible on everyday consumer GPUs.

Key Points:

• Uses LiquidAI’s LFM2 backbone with audio‑as‑language paradigm

• Neural codec (NanoCodec) decodes tokens into 22 kHz waveforms

• Trained on 10,000 h of speech in just 6 hours on 8× H100 GPUs

• Zero‑shot voice cloning via speaker embeddings from short clips

• Produces 10 seconds of audio in ~2 seconds (RTF 0.2)

• Runs on consumer GPUs with only ~3 GB VRAM (e.g., RTX 3060)

• Open‑source under Apache 2.0 license for commercial use

Takeaway: Kani‑TTS‑2 marks a significant step toward democratizing high‑quality, customized speech generation: its combination of compact architecture, fast training, real‑time synthesis, and zero‑shot voice cloning makes advanced TTS accessible to anyone with modest hardware and invites ethical, transparent innovation.

Voice AI Powers Safer Clinical Automation

Speechmatics teams with Edvak EHR to bring enterprise-grade, audit-ready voice AI into real-time clinical workflows.

Speechmatics has partnered with Edvak EHR, an AI-native electronic health records platform, to embed highly accurate Voice AI into live clinical workflows. This integration turns speech into structured, audit-ready documentation, powering tasks, referrals, care coordination, and coding in real time. The focus on accuracy ensures clinical meaning remains intact, even when terminology is complex or audio conditions are imperfect. This collaboration marks a shift from Voice AI as a transcription tool toward becoming healthcare infrastructure built for scale and compliance.

Key Points:

• Speechmatics’ English Medical Model achieves 93% real-time accuracy and 96% keyword recall.

• Edvak EHR uses Darwin AI for real-time conversion of speech into structured workflows.

• The system supports on-premises, private cloud, and SaaS deployment with HIPAA-aligned compliance.

• Voice AI now underpins documentation and action triggers directly within the EHR.

Takeaway: This partnership demonstrates how voice AI can move beyond dictation to become dependable, compliance‑ready clinical infrastructure, seamlessly transforming conversations into actionable care workflows while preserving meaning.

🎙️ Mic Drop

What else is making noise in voice AI.

Previously undisclosed funding rounds underline a surge of VC interest in voice AI customer service startups. (newcomer.co)

Newo’s latest funding will drive autonomous voice AI for small business front desk and call-handling automation. (ventureburn.com)

Simple AI secures major backing to accelerate its custom voice AI agents for sales, with analytics and generative capabilities. (businesswire.com)

Raven-1 adds real-time multimodal perception to conversational AI, enabling more natural agent-customer interactions. (financialcontent.com)

Secai will ramp up deployment of certified voice AI automations for the North American healthcare sector. (tipranks.com)

Compliance concerns grow as voice AI agents trigger legal scrutiny under consumer protection laws. (nationalmortgagenews.com)

Microsoft is sued for alleged unauthorized voice data collection in Teams, spotlighting data privacy risks in enterprise voice platforms. (uctoday.com)

Voice interfaces are emerging as a vital commerce infrastructure layer, reshaping digital transactions and customer engagement. (pymnts.com)

A radio host sues Google, alleging its AI cloned his voice without consent; legal implications mount for voice cloning tech. (newser.com)

Brands increasingly rely on AI voice and chat agents for instant, scalable customer support and operational efficiencies. (bbntimes.com)

Gnani.ai debuts a 5B-parameter Indian voice-to-voice AI model, expanding multilingual conversational AI capabilities. (digit.in)

Sentai’s voice AI companion is designed to provide respectful, non-intrusive support for seniors—offering a differentiated smart speaker experience. (techradar.com)