Giga Raises $61M to Expand Voice AI

November 11, 2025

🔊 Soundcheck

  • Giga nets $61M to supercharge enterprise voice AI

  • AI receptionist startup Beside banks $32M

  • CAMB.AI & Broadcom: Voice AI on Chip

  • AppTek’s Human‑Emotion TTS for AI Dubbing

Read time: 4 minutes

🔥 Hot Mic

Big moves, deep dives, and standout stories.

Giga scores $61 million Series A to scale voice‑based AI agents for enterprise customer support.

Giga, a San Francisco voice‑AI startup founded by IIT Kharagpur alumni Varun Vummadi and Esha Manideep, just closed a $61 million Series A round led by Redpoint Ventures, with backing from Y Combinator and Nexus Venture Partners. The fresh capital will be used to boost enterprise deployments and grow its engineering team. Giga’s platform powers real‑time, emotionally intelligent voice agents that understand context, act in under a second, and support complex workflows. Its deployment at DoorDash has already reduced escalations, sped up resolution, and unlocked smoother operations across languages and regions.

Key Points:

  • Raised $61M Series A led by Redpoint Ventures

  • Backers include Y Combinator and Nexus Venture Partners

  • Voice-AI agents deploy in under two weeks

  • DoorDash deployment led to fewer escalations, faster resolutions

Takeaway: This funding highlights a turning point for enterprise voice AI—Giga delivers real‑time, emotionally aware customer support at scale, proving that useful, fast‑to‑deploy voice agents can transform how companies serve customers.

Beside raised $32 million to scale its AI receptionist that handles calls, bookings, and follow‑ups for small businesses.

Beside just emerged from stealth with $32 million in funding to grow its AI receptionist that manages calls, schedules, and customer follow‑ups for small businesses. The startup connects to existing business phone lines, remembers customer details, and automates appointment booking while creating a searchable “second brain” of calls and conversations. Founded by Maxime Germain, Beside has already reached $4 million in ARR with over 20,000 paying customers, and aims to expand product development, hiring, and move into European markets.

Key Points:

  • Raised $32 million in Series A funding led by EQT Ventures and Index Ventures

  • AI receptionist handles millions of calls per month for SMBs

  • Reached $4 million ARR with 20,000+ paying customers pre-launch

  • Features voice‑clone, searchable transcripts, call and appointment automation

Takeaway: Beside is making automated voice reception accessible to small businesses — turning their phones into memory‑rich, always‑on assistants that book, follow up, and free owners to focus on what matters most.

CAMB.AI’s MARS speech model now runs on Broadcom NPUs enabling ultra‑low‑latency, private, on‑device voice AI.

CAMB.AI and Broadcom have teamed up to embed CAMB.AI’s MARS voice model directly into Broadcom’s neural processing unit, letting text transform into natural speech right on the device. That means real‑time dubbing, captioning, and accessibility features without needing cloud access. It minimizes latency, boosts privacy, and cuts operational costs.

The collaboration opens doors to powerful local voice AI for devices already powered by Broadcom chips—like set‑top boxes and broadband gateways—potentially delivering multilingual and accessible content across more than 500 million deployed devices. The next phase will explore on‑device real‑time translation for over 150 languages, bringing true edge AI to everyday products.

Key Points:

  • MARS speech model embedded into Broadcom NPU chipsets

  • Text‑to‑speech, dubbing and captions run fully on‑device

  • Avoids cloud, cutting latency, enhancing privacy and lowering cost

  • Future plans include real‑time translation in 150+ languages on‑chip

Takeaway: Bringing CAMB.AI’s voice AI into Broadcom’s hardware marks a shift: AI-powered communication moves from the cloud into users’ devices, offering responsiveness, privacy, and global access in one swipe.

AppTek unveils a new multilingual TTS for AI dubbing with granular emotional, prosodic and ethical control.

AppTek has introduced a next‑generation expressive text‑to‑speech model designed specifically for AI dubbing. The model captures nuanced prosody, emotion, and voice characteristics across languages, giving users minute control over pace, tone, pronunciation, dialect and emotional inflection.

Executives in blind tests described the output as voices that “talked to my soul,” highlighting a leap in authenticity. Built from ethically sourced training data, this TTS offers customization down to sub‑word timing, pronunciation lexicons, and even non‑verbal cues like breaths and laughter.

This production‑ready solution blends technical precision and artistic expression with enterprise needs in mind, all while maintaining ethical standards in data use and voice rights. It’s a powerful new tool for anyone looking to localize content with emotional fidelity.

Key Points:

  • Multilingual expressive TTS trained on ethically sourced data

  • Granular control over word‑level emphasis and sub‑word timing

  • Authentic non‑verbal cues—laughter, breaths, gasps, and more

  • Blind tests praised speech quality as emotionally unmatched

Takeaway: AppTek’s new TTS rethinks AI dubbing by merging studio‑quality speech, emotion, customization, and ethical sourcing into a tool that delivers voices that feel deeply human—empowering creators with control and integrity.

🎙️ Mic Drop

What else is making noise in voice AI.

Popai Health’s voice-driven intelligence aims to boost patient engagement and care intervention in the healthcare sector. (streetinsider.com)

Prosper AI will use fresh funding to grow its AI agent platform supporting healthcare scheduling and claims management. (mobihealthnews.com)

Guide details installing open-source Dia-1.6B, providing devs with a speech alternative to ElevenLabs—especially for privacy-focused use cases. (vocal.media)

A Berlin judge fined a YouTuber for using AI-generated voice, setting legal precedent for voice clone consent in commercial projects. (ppc.land)

GM will debut eyes-off driving and new conversational AI, accelerating in-car voice assistant capabilities for the automotive industry. (roboticsandautomationnews.com)

Perfect Corp. adds a voice conversational agent, showcasing new use cases for voice-driven beauty and fashion experience. (businesswire.com)

Stream Ring brings a finger-worn device to market, placing always-on voice assistance on users’ hands—expanding wearable applications. (gadgetsandwearables.com)

Pandit.ai launches specialized multilingual voice AI for religious ceremonies, demonstrating verticalization of voice technologies. (business-standard.com)

Market research reports substantial growth for the voice recognition sector, led by major global tech companies. (openpr.com)

Freya lands seed funding to expand AI automation for business conversations, reflecting increased interest in conversational operations. (ain.ua)

WIZ.AI will expand domain-specific conversational LLM voice agents across APAC after securing 'tens of millions' Series B funding. (antaranews.com)

VCAT AI debuts conversational interface for visual content creation, blending voice/text dialogue with multimodal AI pipelines. (manilatimes.net)

Google faces user complaints during Gemini voice assistant rollout, highlighting ongoing implementation challenges for large-scale voice platforms. (thehansindia.com)

Australian mobile banks implement conversational AI for improved user experience and proactive customer service features. (cdotrends.com)