Modulate Raises $25M to Sell AI That Listens to Voice Calls, Not Just Transcripts
Modulate, a Boston company that builds AI models to analyze speech directly rather than working from text transcripts, has raised $25M in a round led by Future Ventures. Hyperplane and Lakestar also took part, bringing the company’s total funding to $60M.
The Company and Its Founders
The company was founded by Mike Pappas and Carter Huffman, who first met at MIT when Mike solved a physics problem that Carter was working on in a hallway. Their shared interest in how computers affect people’s ability to connect led them to become friends.
Modulate initially gained recognition in the gaming world. Since 2023, its ToxMod system has helped moderate voice chat in Call of Duty, and Lakestar led a $30 million Series A funding round in 2022.
Modulate's Technology and Applications
Modulate’s main product, Velma, can analyze audio signals, picking up on emotion, tone, intent, emphasis, and whether the voice is synthetic. Velma employs what Modulate calls an Ensemble Listening Model, which combines more than 100 smaller, specialized audio models for each task.
- Efficiency and Accuracy: This method is up to 1,000 times more efficient than using a single large model and Velma is twice as accurate as general-purpose large language models at identifying actual problems while generating seven times fewer false alarms.
- Applications: Modulate’s technology is used for fraud prevention, customer service, and monitoring AI voice agents. It also offers voice masking to protect high-risk staff and detects child grooming in voice conversations.
Market and Future Plans
Demand for Modulate’s technology is spreading beyond gaming into AI agents, security, and customer experience. The funding will be used for research, engineering, and expanding access to developers through SDKs, APIs, and partner integrations.
"We have gained a significant technical lead in audio-native AI," said Steve Jurvetson, co-founder of Future Ventures.
Modulate’s public benchmark results are impressive, with its transcription models taking first place on Hugging Face’s Open ASR Leaderboard in July and its deepfake detector ranking first on Hugging Face’s Speech Deepfake Arena since March.