Deepgram and Fortanix announced a partnership that will allow enterprises to run voice AI in their own environment while ensuring their data is securely protected. With the partnership, Deepgram can leverage Fortanix Confidential AI and NVIDIA Confidential Computing to add advanced security to self-hosted environments.

“AI labs and model owners face enormous pressure to commercialize and distribute their AI models widely, but doing so can create security risks that threaten the very intellectual property they invested huge amounts of time and money to create,” said Anand Kashyap, CEO and co-founder, Fortanix. “Fortanix Confidential AI changes that by providing model owners a secure, proven path to unlock new markets and enterprise customers without exposing their most valuable IP.”

Deepgram said that the AI model weights remain protected during active processing in their own environments, aiming to allow organizations to build highly-secure real-time voice applications without sacrificing performance.

The company said that the on-premises solution runs Deepgram’s voice AI models with Fortanix Confidential AI on NVIDIA Confidential Computing-enabled GPUs. NVIDIA GPUs with Confidential Computing aims to enable AI workloads to process sensitive data inside a trusted execution environment. Deepgram said that the partnership serves to deliver in-use data protection that has not been seen previously on the enterprise level.

The Deepgram, Fortanix, and NVIDIA solution serves to provide voice AI applications that handle sensitive customer and patient interactions, transcription layers across enterprises that capture calls, meetings and internal conversations for analytics and search, as well as allows voice-enabled IT, operations and service desk applications to run inside an organization’s secure perimeter.

“AI labs and model owners face enormous pressure to commercialize and distribute their AI models widely, but doing so can create security risks that threaten the very intellectual property they invested huge amounts of time and money to create,” said Anand Kashyap, CEO and co-founder, Fortanix. “Fortanix Confidential AI changes that by providing model owners a secure, proven path to unlock new markets and enterprise customers without exposing their most valuable IP.”

Fortanix said that its Confidential AI protects data and AI model weights while they’re running. It builds on NVIDIA GPUs with Confidential Computing to create Trusted Execution Environments that isolate the AI workload from the underlying infrastructure and OS. This aims to allow regulated organizations to unlock AI innovation with security while meeting HIPAA, GDPR and national-data residency requirements.

“Real-time voice AI helps industries scale efficiency and safety — empowering healthcare providers to update records more quickly, letting technicians repair equipment without putting tools down to check manuals — but sensitive data needs to be secure throughout processing,” said Justin Boitano, vice president, Enterprise AI Platforms, NVIDIA. “NVIDIA Confidential Computing, combined with Fortanix and Deepgram, provides a powerful platform for regulated industries to deploy production-grade voice AI applications with sovereignty and trust.”