Funding and Deals

  • Wispr raises $280M at $2B valuation and previews its 2B-parameter Canto speech model, revenue growing 150% quarterly. (TechCrunch, Reuters, SiliconANGLE)

  • Poland takes a stake in ElevenLabs, highlighting the country's role in supplying AI talent. (TechCrunch)

  • Murf AI launches Falcon 2, a $0.01-a-minute voice model aimed at OpenAI and ElevenLabs. (Moneycontrol, Bloomberg)

  • HappyRobot is live at 8 of the 10 largest US freight brokers and is reportedly raising toward a $1.2B valuation. (Forkast, The Stack)

  • HeyBreez raises $2.5M seed led by Lunara Partners for enterprise voice AI infrastructure. (waya.media)

  • Krisp joins the 8x8 partner ecosystem (noise cancellation, accent conversion, multilingual voice). (Telecom Reseller)

  • didlogic joins Vapi's Partner Directory for phone numbers and SIP connectivity. (Nat Law Review)

  • TTGI signs OEM partnership with GetVocal AI across its TaaS platform. (Newsfile)

Models and Product Launches

  • Cartesia ships Sonic-3.6, a streaming TTS model with sub-90ms latency, ranking first on both Artificial Analysis speech arenas. (MarkTechPost)

  • Adobe Firefly makes three generative audio tools generally available (music, speech, sound effects). (Adobe, SiliconANGLE)

  • ElevenLabs launches a hosted MCP connector in Claude, letting users configure and modify production voice agents in natural language. (ElevenLabs, The New Stack)

  • Superwhisper releases the S1 model family, including S1-Voice (cloud STT) and S1-mini (open-weights normalizer). (MarkTechPost)

  • Breeze Blue launches Breeze TTS 2, a real-time flagship model for interactive media. (EIN News)

  • Blue Machines AI launches Floe, an 11-language model that uses context to detect language switching. (Analytics India)

  • Grok Voice Think Fast 2.0 claims the top spot on a voice AI benchmark. (Basenor)

  • OpenWhispr launches open-source local STT for dictation without the cloud. (Trend Hunter)

  • ReVoiceLab launches an AI dubbing tool for localizing video and audio. (Trend Hunter)

  • Choicer Voicer launches an AI voice generator that produces audio in six seconds. (Digital Journal)

  • LALAL.AI introduces its Lynx algorithm for cleaning up vocal rips and samples. (Attack Magazine)

  • Microsoft adds a Customer Intent Agent for voice in Dynamics 365. (Microsoft Learn)

Enterprise and Contact Center

  • Taco Bell now runs voice AI in over 890 US restaurants, with employee-intervention rates ranging 3% to 33% in independent testing. (The Next Web)

  • Seavoice launches a multilingual voice agent for Southeast Asian contact centers with mid-call code-switching. (FinancialContent)

  • Five9 integrates Regal's autonomous voice agents through its AI Agent Connect program. (Webull, Simply Wall St)

  • New American Funding partners with Kastle to deploy end-to-end borrower voice agents. (HousingWire)

  • NewtekOne integrates Glia's banking voice AI into its client service platform. (Quiver)

  • AudioCodes Voca CIC becomes one of the first solutions certified under Microsoft's Teams Voice Agent program. (PR Newswire)

  • HeyGuest launches a voice AI ordering platform targeting the 43% of hospitality calls that go unanswered. (Hospitality and Catering News)

  • Wave Sales launches voice agents for home solar sales. (pv magazine)

  • Click Voice AI launches a 24/7 receptionist for home service contractors. (FinancialContent)

  • Orvera AI is recognized on the CMP Prism 2026 for voicebot and conversational IVR. (Business Wire)

  • RingCentral uses agentic voice AI to auto-document conversations. (USA Today)

  • Zendesk's Jon Aniano on how voice AI is shifting customer-service economics. (Computer Weekly)

  • SoundHound AI accelerates enterprise voice expansion. (Mshale)

  • AI receptionist market projected to grow at 27.8% CAGR. (Market.us)

  • Seven best AI phone agents for recruitment teams in 2026. (Onrec)

  • Why voice AI is becoming enterprise infrastructure. (The AI Journal)

Healthcare

  • Summa Health partners with Hippocratic AI in its first wave of tech deployments. (Forbes)

  • Hippocratic AI unveils Agentic Orchestrators for healthcare voice agents. (Macau Business)

  • Cloud conversational voice AI targets healthcare access via EHR-integrated scheduling and triage. (HIT Consultant)

  • AI identifies aspiration risk from a two-second post-swallow voice recording. (Korea Biomedical Review)

  • Connected-speech analysis screens stroke language impairment at 90% accuracy. (medRxiv)

  • SAVER: speech analysis to predict psychotherapy treatment outcomes. (PsychArchives)

  • Working paper argues therapy bots offer a counterfeit therapeutic alliance. (Zenodo)

  • Text-audio intent recognition framework for atypical speech in elderly care. (Sensors)

  • Munson Army Health Center adopts ambient listening. (DVIDS)

  • AI medical scribing market projected to reach $9.67B by 2035. (GlobeNewswire)

  • AI agents for clinical documentation market. (Precedence Research)

Security, Deepfakes and Scams

  • Scammers now clone voices from as little as three seconds of audio for emergency-fraud calls. (The420, Indian Express)

  • Families and orgs revive verbal passwords as a deepfake defense. (The Tech Buzz)

  • Scammers harvest audio from Instagram to clone children's voices. (WFMY)

  • Kolkata senior loses roughly 96,000 rupees to a voice-cloning scam. (Times of India)

  • South Korea's anti-phishing platform prevents 66.4 billion won in losses over nine months. (Korea JoongAng Daily)

  • GCash iGnite Summit addresses AI voice-cloning fraud in digital finance. (Streamline)

  • Fact check: several common warning signs for spotting cloned voices are obsolete. (Notebookcheck)

  • Resemble AI releases DETECT-World for AI-generated audio, video, and images. (Resemble)

  • Corsound AI launches a voice-verification solution on the Microsoft Marketplace. (Microsoft Marketplace)

  • New ELAD-SVDSR dataset advances synthetic-voice detection and biometric security. (arXiv)

  • Gated-fusion framework combines linguistic, emotional, and rhythmic features for fake-speech detection. (Applied Sciences)

  • "Reputation arbitrage" working papers describe how cloned voices exploit borrowed credibility, plus a proposed action-authentication defense. (Zenodo trust paper, Zenodo defense model)

Legal and Regulation

  • Apple, Meta, and seven other tech giants face lawsuits over harvesting recordings to train AI. (Reuters, American Bazaar)

  • Delhi High Court bars unauthorized AI voice cloning of actress Khushi Kapoor. (ANI)

  • Japan issues guidelines on unauthorized AI imitations of voice actors. (JAPAN Forward)

  • Indonesia's regulatory framework struggles with voice cloning and deepfakes. (Governance Jurnal)

  • AI music boom raises new copyright questions over artists' voices. (Radio and Music)

  • Attorney warns radio contracts have an AI voice-clone problem. (Radio Ink)

Consumer, Assistants and Gadgets

  • Google adds voice discussion of research reports to Gemini Live. (TechCrunch)

  • Amazon unlocks Alexa Plus free for all US Fire TV owners. (About Amazon)

  • Sonos preps new hardware and AI voice controls to challenge Alexa and Google. (The Tech Buzz)

  • Microsoft retires its Mico companion from Copilot Voice, moving it to the classroom. (Notebookcheck)

  • One year in, Google and Amazon's rebuilt generative smart-home agents still fall short. (Forkast)

  • Reddit tests turning text posts into AI-narrated video and audio. (Mashable, The Next Web)

  • Wispr Flow dictation earns praise across Windows, macOS, iOS, and Android. (ZDNET)

  • A NYT columnist criticizes Wispr Flow's performance and usability. (NYT)

  • On-device dictation apps outpace Apple's built-in option on Mac. (Digital Trends)

  • Logitech warns poor audio quality drives costly AI dictation errors. (TelcoNews AU)

Research and Academia

  • Survey of state-of-the-art generative audio models (synthesis, conversion, cloning). (AI Review)

  • Personality-matched embodied agents improve social perception and rapport in VR. (Virtual Reality)

  • Physics-based speech synthesis using vocal-tract waveguide models, no neural inference. (Zenodo)

  • Project Chimera improves neural audio codec fidelity via curriculum sampling. (Scientific Reports)

  • Framework predicts speaker age, gender, and party from text speeches. (Iran J Computer Science)

  • Head orientation and high-frequency cues improve speech-in-speech recognition. (bioRxiv)

  • No overall credibility gap between native and non-native accented TTS for US listeners. (Frontiers in Communication)

  • AI classifies poetic meter in spoken Arabic poems despite scarce labels. (PeerJ CS)

  • KAIST builds a neuromorphic chip using semiconductor noise for motion and speech. (Dong-a Science)

  • Design thinking plus AI language tools improves EFL learners' oral skills. (SSAHO)

  • Reliance on speech-recognition tech predicts interpreting learning outcomes. (JELTAL)

Global and Languages

  • IISc's SraVaani pushes speech recognition beyond India's 22 mainstream languages. (Moneycontrol)

  • New open speech model supports 65 Indian languages. (BW Education)

  • MWire Labs releases Lemka for six Northeast Indian languages. (Highland Post)

  • Voice AI finds traction in West Africa, centered on phone conversations. (Forbes Africa)

  • Chosun Ilbo adds AI summaries and natural voice playback. (Chosun)

  • A Tokyo snack bar uses speech recognition to make conversation visible for hard-of-hearing customers. (Japan News)

Media, Entertainment and Funky

  • Spotify AI DJ voice actor Xavier "X" Jernigan revealed. (ABC News)

  • Survival game Rerouted uses proximity voice chat so in-game wildlife reacts to your mic. (Tech Times)

  • Voice Heist, a comedy heist game where you command an AI robot by voice, hits PC in 2027. (COGconnected)

  • The industry recreating the voices of deceased relatives for ongoing conversations. (Democrata)

  • Mechanics VoiceOver ships an AI voice localization for Assault on Dark Athena after failing to fund traditional voice acting. (IXBT Games)

  • James May tests Tesla's Grok-powered assistant "Ara," calling the multiple personas insincere. (Benzinga)