What if edge AI could run multi-modal applications, GenAI, and real-time interactions while using under 10 watts? A new platform is making it possible.
SiMa.ai, a company specializing in embedded edge machine learning system-on-chip (MLSoC) solutions, has launched the Modalix 50 TOPs device and introduced an Early Access Program for MLSoC Modalix. Modalix is a platform designed to integrate multi-modal applications, including Transformers, LLMs, LMMs, and GenAI, alongside computer vision algorithms at the edge. It offers power efficiency and cost-effectiveness while delivering performance per watt that exceeds alternatives by more than 10X.
The MLSoC and its architecture enable real-time human-machine interactions while processing multiple sensor inputs. This makes it suitable for applications that require decision-making and interactions in the physical world. Modalix builds on SiMa.ai’s first-generation MLSoC, supporting on-device AI and GenAI use cases while operating at under 10 watts. It brings GenAI and conversational AI capabilities to applications such as humanoid robots, LMM-based infotainment systems, factories, automated mobile robots, retail and vision systems, medical technology, aerospace and defence, agriculture, construction, and industrial drones.
The SiMa.ai ONE Platform for Edge AI simplifies the adoption of Modalix. This development framework includes the Palette software suite, allowing customers to build and deploy AI solutions at the edge. The ONE Platform supports multi-modal and GenAI models, ensuring an upgrade path while optimizing Total Cost of Ownership (TCO).
“Since announcing MLSoC Modalix last fall, customer interest has exceeded our expectations,” said Krishna Rangasayee, Founder and CEO at SiMa.ai. “With the sampling of Modalix, we’re now ready to put this breakthrough technology directly into customers’ hands, enabling them to experience firsthand how Modalix can transform their multi-modal and GenAI applications with unmatched industry-leading performance and efficiency. I am proud of the SiMa.ai team and their dedication and focus in delivering our second-generation product to customers in record time.”
For more information, click here.
