SiMa.ai, a Silicon Valley startup specializing in embedded machine learning (ML) system-on-chip (SoC) platforms, has announced a $70 million extension funding round. The company intends to use this funding to bring its second-generation chipset, designed for multimodal generative AI processing, to market.
The market for AI-supporting chips is projected to more than double by 2027, reaching $119.4 billion globally, according to Gartner. While few players have entered the dedicated semiconductor space for AI applications, there is growing demand for AI processing at the edge, closer to data gathering sources. SiMa.ai aims to capitalize on this trend by offering its edge AI SoC to various sectors, including industrial manufacturing, retail, aerospace, defense, agriculture, and healthcare.
SiMa.ai’s first ML SoC targeted the market segment between 5W and 25W of energy usage and combined proprietary chipsets with no-code software called Palette. This combination has already been adopted by over 50 companies globally.
While the first-generation chipset focused on classic computer vision, SiMa.ai plans to introduce its second-generation ML SoC in the first quarter of 2025. This new chipset will emphasize multimodal GenAI capability and will feature architectural tunings over its predecessor while retaining fundamental concepts.
The GenAI SoC will adapt to any framework, network, model, and sensor, and will support various modalities, including audio, speech, text, and image. It aims to serve as a single-edge platform for all AI applications, covering computer vision, transformers, and multimodal GenAI.
SiMa.ai’s approach is to evolve around a chosen vector while maintaining architectural flexibility and nimbleness through a software-centric design. This approach allows the company to adapt to changing market demands effectively.