CAMB.AI and Broadcom have teamed up to embed CAMB.AI’s generative voice model, MARS, directly into Broadcom’s neural processing unit (NPU) chipsets. In the next phase, the collaboration aims to bring CAMB.AI’s real-time translation model to Broadcom’s on-device NPU, enabling seamless translation across more than 150 languages.
The collaboration allows text-to-speech and localisation capabilities to operate directly on consumer devices, eliminating the need for constant cloud connectivity.
This integration allows real-time translation, dubbing, captioning, and audio descriptions to run locally on devices, minimising latency, enhancing privacy, and lowering operational costs for both users and content creators.
Akshat Prakash, Co-founder and CTO of CAMB.AI, stated, “Our partnership with Broadcom allows us to bring this technology to consumers worldwide faster, more securely, and seamlessly integrated into everyday devices like never before.
Powered by Broadcom’s SoC-integrated NPU, CAMB.AI’s text-to-speech model transforms written text into lifelike speech across multiple languages. This on-device approach enhances accessibility for visually impaired users, streamlines communication in e-learning and customer support, and reduces dependence on external servers.
Rich Nelson, Senior Vice President and General Manager of Broadcom’s Broadband Video Group, remarked, “We’re ushering in the next generation of user experiences that combine advanced intelligence with a strong focus on privacy.”
In the next stage, the partnership will focus on bringing CAMB.AI’s real-time translation model to Broadcom’s on-device NPU, enabling seamless translation in over 150 languages. With Broadcom’s chips already embedded in more than 500 million devices worldwide from set-top boxes to broadband gateways this advancement has the potential to make multilingual and accessible content available to households across the globe.
CAMB.AI, recognised for its multilingual AI localisation expertise with partners like IMAX, Comcast NBCUniversal, and Major League Soccer, has developed its MARS and BOLI models, which are currently accessible on AWS Bedrock and Google Vertex AI.
The collaboration with Broadcom represents a significant step toward integrating localisation and accessibility directly at the hardware level, making AI-driven communication more seamless and closer to end users than ever before.
