SiMa.ai, a software-centric, embedded edge machine learning system-on-chip (MLSoC) company, has announced MLSoC Modalix, which it claims is the industry’s first multi-modal edge AI product family. The MLSoC Modalix supports CNNs, Transformers, LLMs, LMMs, and Gen AI at the edge.

Krishna Rangasayee, Founder and CEO at SiMa.ai, said, “Given the amount of change the AI industry goes through, being software-centric is key. Over the last year, we have doubled the number of customers, so we’re growing quite well. Multimodality has become big and we will see that shift in the edge.”

“Large language models and generative AI have also really taken off and they’re happening not only at the cloud but also at the edge in certain applications. We have introduced an extension to our ONE platform for edge AI with a product line called MLSoC Modalix.”

MLSoC Modalix is supposed to enable developers to push the envelope on performance with a small power envelope and physical footprint across applications. They will be able to run end-to-end GenAI application pipelines on just a single chip while delivering measurable contributions to the business’s top and bottom lines.  

Rangasayee said that while the company’s primary competition ends up being Nvidia. For the edge market, both performance and power matter.

“A key metric our customers care about is performance per watt, as in how much to compute for a given power budget. We are consistently 10x of Nvidia. Doing something 10x of the market leader is not easy and as a startup, we are doing this quite well globally.”

He added that the company has customers in India and the rest of the world.  

SiMa.ai MLSoC Modalix is the second generation of the commercially deployed first generation MLSoC. MLSoC Modalix is offered in 25 (Modalix 25 or “M25”), 50 (Modalix 50 or “M50”), 100 (Modalix 100 or “M100”) and 200 (Modalix 200 or “M200”) TOPS configurations, in multiple form factors, and is purpose-built to provide deployment of GenAI for the embedded edge ML market.

Fully software compatible with first-generation MLSoC, the MLSoC Modalix product family was designed to enable the capability to run DNNs, as well as advanced transformer models, including LLMs, LMMs and Generative AI. Samples of MLSoC Modalix will be available to customers in Q4 of 2024.  

The founder also added that SiMa.AI is scaling globally and doubling its customer footprint. The company also established its R&D center in Bangalore in 2020 and currently has around 70 people.

“We are engaged with some significant large market leaders and smaller companies in India. We are in multiple markets like robotics, automotive, aerospace defense, smart vision systems, medical and industrial automation. Given our footprint in India, we’re excited about supporting the customers locally from our India center,” he said.

SiMa.ai was founded in 2018, has raised $270M, and is backed by Fidelity Management & Research Company, Maverick Capital, Point72, MSD Partners, VentureTech Alliance, and more.