Modulate Raises $25M for Voice AI as DetectifAI Targets On-Device Deepfake Detection

Modulate was founded in 2017 by Mike Pappas and Carter Huffman, who met as MIT physics undergraduates. It started with voice modulation for gaming and later focused on voice-based moderation. The company now sells transcription, emotional analysis, deepfake and AI music detection, and policy enforcement for voice agents in regulated industries. Its Velma platform runs more than 100 small specialized models under an Ensemble Listening Model architecture, which Modulate says is up to 1,000 times more efficient than using one large model for the same audio.

Modulate said more than 10 million hours of audio pass through its models each month, and lifetime usage recently passed 600 million hours. Two products have taken first place on Hugging Face leaderboards this year: its transcription model topped the Open ASR Leaderboard in July, and Velma Deepfake Detect, launched in March, sits at the top of the Speech Deepfake Arena. The company said the deepfake model is 98.9% accurate on public benchmark data. Developers can buy batch transcription through its API for three cents an hour.

Customers include health care institutions screening callers who impersonate staff with deepfake voices, call centers that want alerts about possible scams, and companies running voice AI agents. Those customers use the models to assess call quality and compliance, and Modulate said its technology is also used to monitor cyberattacks through voice calls. Huffman said voice is becoming a primary interface for AI, creating problems that cannot be solved from a transcript alone. He added that developers shouldn't have to rebuild the audio intelligence layer every time they create a new voice experience. The company has 40 to 45 employees and plans to add about 10 people, and it is working on more on-premises and on-device deployment options for privacy.

Steve Jurvetson, co-founder of Future Ventures, said Modulate has gained a significant technical lead in audio-native AI and that demand is spreading into AI agents and security. TechCrunch, citing PitchBook, reported that Modulate had raised $41 million before the new round at a $170 million valuation. SiliconANGLE reported that total backing including the new round is $60 million. The new funding is intended partly for software development kits, application programming interfaces and industry-specific models, as well as hiring in research, engineering and developer relations.

The round comes as deepfake voice scams draw more attention. TechCrunch reported that Tarini Padmanabhuni's grandfather paid a ransom about two years ago after a call used an AI-generated imitation of his brother's voice to claim the brother had been kidnapped. She later founded San Francisco-based DetectifAI, which is developing deepfake voice detection designed to run on a smartphone rather than in the cloud.

According to the FBI, Americans lost close to $900 million to AI-driven scams last year, up 24% from 2024. People 60 and older lost twice as much as those aged 50 to 59. DetectifAI argues that cloud-based detection products cannot be built directly into phones, leaving targets with little defense. The startup says it designs compact AI models from the start that fit inside a phone's operating system and can give an instant verdict on whether a voice is AI-generated during calls, voice messages and other audio, without the audio leaving the device.

DetectifAI is selling first to phone manufacturers, licensing its software development kit so detection can ship as a built-in operating system feature. Padmanabhuni said the first phone maker to ship the technology will gain an edge, and she expects deepfake detection to become a standard specification. She said a secondary revenue stream will come from licensing the technology to businesses and fraud-prevention firms. The startup already handles more than 100,000 calls a month for financial institutions in India, according to TechCrunch. Those calls are made by AI voice agents handling debt collections and loan document follow-ups, with deepfake detection and speaker verification on every call. She declined to name customers.

DetectifAI competes with Reality Defender, Pindrop, Resemble AI, Microsoft Azure AI Content Safety and Nuance, which Microsoft also owns. It has raised a small seed amount from investors Josh Constine and Manohar Kamath, according to TechCrunch. The company is one of the startups selected by TechCrunch's editorial team to compete in Startup Battlefield at TechCrunch Disrupt, scheduled for Oct. 13 to 15 in downtown San Francisco.