SoundHound AI Delivers Voice Assistants at Scale with NVIDIA

SoundHound AI, a provider of voice artificial intelligence, has expanded its collaboration with NVIDIA, integrating NVIDIA AI Enterprise to accelerate inference and retrieval capabilities.

By integrating SoundHound's voice AI technology with NVIDIA NIM and NeMo microservices, part of the NVIDIA AI Enterprise software platform, SoundHound is driving low-latency AI processing, real-time retrieval-augmented generation (RAG), and scalable model optimization. This expanded collaboration reinforces SoundHound's voice AI solutions across industries, including automotive, restaurant, and customer service. The integration reduces inference latency, improves response accuracy, and optimizes AI deployments for both cloud and edge environments.

Key NVIDIA AI Enterprise innovations coming to SoundHound's advanced voice platform include the following:

• Optimized AI Inference using NVIDIA NIM Microservices, enabling containerized and high-speed inference for large language models, significantly reducing response latency.
• Smarter Retrieval with NVIDIA NeMo Retriever Microservices for RAG, enhancing context-aware retrieval-augmented generation by surfacing highly relevant responses.
• Scalable Model Fine-Tuning with NVIDIA NeMo to optimize LLMs across multi-GPU environments at scale.

SoundHound's integration of NVIDIA AI Enterprise software is already playing a role in automotive AI, including in Lucid's voice AI system. Currently, select modules within the Lucid Assistant leverage SoundHound and NVIDIA AI to enhance specific functionalities such as faster response times and improved user interactions. SoundHound is working to expand NVIDIA AI software integrations across additional modules, bringing even greater efficiency and intelligence to automotive voice experiences.

"The AI industry has made massive strides in model training over the last two years, but the real challenge now is deploying these models at scale efficiently," said Keyvan Mohajer, CEO and co-founder of SoundHound AI, in a statement. "Our collaboration with NVIDIA on NIM and NeMo microservices is solving this problem, helping us optimize inference so businesses and consumers can benefit from faster, more accurate, and highly scalable voice AI."

"The AI industry is at an exciting turning point, where the focus is shifting from breakthrough model training to deploying these advancements efficiently at scale," said Rishi Dhall, vice president of automotive at NVIDIA, in a statement. "With NVIDIA AI Enterprise, SoundHound is unlocking new possibilities for voice AI, delivering faster, more accurate, and highly scalable interactions that provide more accurate and engaging user in-vehicle experiences."