Resemble AI
Category: Voice / Speech AI
A dual-platform generative voice AI and multimodal deepfake detection company that enables enterprises to create ultra-realistic synthetic voices while simultaneously protecting against AI-generated audio, video, and image threats. Resemble AI was founded in 2019. The company is led by Zohaib Ahmed. Based in Mountain View, United States. Team size: 30-40. Total funding raised: $25.0M. Latest round: Strategic Round ($13.0M, Dec 2025). Key investors include Javelin Venture Partners, Comcast Ventures, Craft Ventures, Google AI Futures Fund, Sony Innovation Fund, Okta Ventures.
AMW Analysis
Resemble AI operates a dual-platform business that both generates synthetic voices and detects deepfake audio, video, and image threats. The company’s value proposition centers on a unified system that claims 98.1% detection accuracy across 50+ languages and offers on-premises deployment. In December 2025, Resemble AI raised a $13 million strategic round from Sony Innovation Fund, Okta Ventures, and Google’s AI Futures Fund, bringing total funding to $25 million. The funding round was framed around combating deepfake fraud, with the company citing $1.56 billion in fraud losses in 2025 and projections of $40 billion by 2027. Around the same time, Resemble AI released Chatterbox Turbo, an open-source text-to-speech model that clones voices from five seconds of audio with sub-150ms latency. The model is MIT-licensed and includes built-in watermarking. The company’s detection model, DETECT-3B Omni, is reported to achieve 98% accuracy across 38 to 40+ languages. The news flow shows Resemble AI positioning itself at the intersection of voice generation and deepfake defense, with recent product releases and funding emphasizing the detection side of the business.
AMW analysis, generated from 4 tracked news signals.
- Founded
- 2019
- Headquarters
- Mountain View, United States
- Team size
- 30-40
- Total funding
- $25.0M
Value proposition
Provides the only unified platform that enables enterprises to create production-quality AI voices while simultaneously defending against deepfake threats, offering 98.1% detection accuracy across 50+ languages (ranked #1 on Podonos Audio DFD Bench) with on-premises deployment options for maximum data security.
Products and solutions
Chatterbox Turbo - #1 ranked open-source voice AI; preferred 2:1 over ElevenLabs in blind eval (22.5k+ GitHub stars, MIT), DramaBox - Expressive open-source TTS; first verifiable and directable speech engine with PerTh watermarking by default, DETECT-3B Omni - Multimodal deepfake detection for audio, video, images; #1 on DFBench & Speech DF Arena (98.1% audio accuracy), PerTh Watermark / PerTh Multimodal - Neural watermarking for audio, video, image, text (EU AI Act compliant), Resemble Intelligence - Multimodal explainable AI powered by Google Gemini 3 for transparent detection, Resemble Meetings - Real-time deepfake detection bot for video calls, Resemble Identity - Real-time biometric voice verification, Resemble Enhance - AI-powered audio enhancement and noise removal, Deepfake Detector Chrome Extension - Browser-based detection, On-Premises Solutions - Air-gapped deployment for enterprise and government
Unique value
Only platform offering dual capabilities in generative voice AI creation and deepfake detection. Developed DETECT-3B Omni, the industry's strongest multimodal detection model ranked #1 on public benchmarks with 98.1% audio detection accuracy across 50+ languages, battle-tested against 160+ generative AI models. Created Chatterbox, the #1 ranked open-source voice AI model that outperforms ElevenLabs in blind evaluations, with 22.5k+ GitHub stars.
Target customer
Enterprise organizations across media & entertainment, government agencies, financial services, and customer service sectors requiring both AI voice synthesis capabilities and protection against synthetic media threats
Industries served
Media & Entertainment, Government & Defense, Financial Services, Healthcare, Customer Service & Contact Centers, Cybersecurity, Enterprise Software
Technology advantage
Proprietary PerTh (Perceptual and Threshold) Neural Speech Watermarker technology that embeds imperceptible, tamper-resistant watermarks surviving compression, resampling, and editing with nearly 100% detection accuracy. Offers zero-shot voice cloning from just 5 seconds of audio. Provides on-premises deployment enabling air-gapped operation for government and enterprise security requirements. CEO testified before U.S. Senate Judiciary Committee on AI voice risks in July 2024, establishing thought leadership in AI safety.
How they differentiate
Only platform offering dual capabilities in generative voice AI creation AND deepfake detection. Industry-leading DETECT-3B Omni multimodal detection with 99.8% accuracy across 38+ languages. Chatterbox open-source model with 22.5k+ GitHub stars and 5-second voice cloning. On-premises deployment with air-gapped operation for government/enterprise security requirements.
Main competitors
ElevenLabs, Murf AI, Play.ht
Key partnerships
Netflix - Recreated Andy Warhol's voice for Emmy-nominated 'The Andy Warhol Diaries' documentary series, Universal Pictures - Generative voice AI for entertainment applications, Paramount Pictures - Voice technology for media production, Carahsoft - Strategic government distribution partner for public sector, Google Cloud - Partnership for generative AI model acceleration (Gemini 3 integration in Resemble Intelligence), Sony Innovation Fund - Strategic investor and technology partner, Okta Ventures - Strategic investor focusing on identity security, The World Bank Group - Enterprise voice solutions, Wa'ed Ventures (Aramco) - Strategic investment for Middle East expansion (Mar 2026), KDDI Open Innovation Fund - Strategic investment and Japan market access, Fortune 500 companies across multiple sectors (undisclosed)
Notable customers
Netflix, Universal Pictures, Paramount Pictures, The World Bank Group, Fortune 500 companies, Government agencies
Major milestones
Raised $13M strategic round in Dec 2025 led by Google AI Futures Fund and Sony Innovation Fund, Series A $8M led by Javelin Venture Partners and Comcast Ventures in Jul 2023, Recreated Andy Warhol's voice for Emmy-nominated Netflix documentary 'The Andy Warhol Diaries', Launched DETECT-3B Omni, industry's strongest multimodal deepfake detection with 99.8% accuracy, Released Chatterbox, #1 ranked open-source voice AI model with 22.5k+ GitHub stars, CEO testified before U.S. Senate Judiciary Committee on AI voice risks (Jul 2024), Partnered with Carahsoft for government distribution and Google Cloud for AI acceleration
Growth metrics
Raised $13M strategic round in December 2025 to combat rising AI-generated threats as deepfake cyberattacks surge. Total funding reached $25M across multiple rounds since 2019 founding.
Market positioning
Enterprise-focused voice AI platform with strategic emphasis on security and deepfake threat protection. Positioned as the trusted solution for Fortune 500 companies and government agencies requiring both creative voice synthesis capabilities and robust protection against synthetic media threats.
Geographic focus
North America (San Francisco Bay Area headquarters), with global enterprise reach through partnerships and government contracts
Patents and IP
Proprietary PerTh Neural Speech Watermarker technology (open-sourced on GitHub). While specific patent numbers are not publicly disclosed, the company maintains trade secrets around their detection algorithms and watermarking techniques. Chatterbox TTS model released under MIT license.
About Zohaib Ahmed
Co-Founder and CEO of Resemble AI. Previously served as Senior Software Engineer at Magic Leap (2017-2019), focusing on user interface and interactions for Lumin OS. Worked as Software Engineer at Hipmunk (2015-2017), a travel search platform acquired by SAP. Completed multiple software engineering internships at BlackBerry (2012-2014). Computer Engineering graduate from University of Waterloo (2010-2015).
Latest news about Resemble AI
- Resemble AI raised $13M from Sony Innovation Fund, Okta Ventures, and Google's AI Futures Fund to combat the deepfake crisis as fraud losses hit $1.56B in 2025. With projections reaching $40B by 2027
- Resemble AI released Chatterbox Turbo, an open-source TTS model that clones voices from 5 seconds of audio with sub-150ms latency. This 350M parameter model is MIT-licensed, commoditizing high-fidelit
- Resemble AI just raised $13M from Google, Sony, and Okta to combat deepfake fraud that caused $1.56B in losses in 2025 alone. Their 3B-parameter Detect-3B Omni model achieves 98% accuracy across 40+ l
- Resemble AI has secured $13M in funding, totaling $25M, to combat the escalating threat of deepfakes and synthetic media. Their DETECT-3B Omni platform achieves 98% accuracy across 40+ languages for m
More Voice / Speech AI companies
Official website: https://www.resemble.ai/
