MIT optical computing spinout Lightmatter released information in November of 2020 about its Passage chip, which allows users to attach third-party hardware from companies like Nvidia and AMD to the modular optical chip and take advantage of its speed and energy efficiency. Lightmatter, which was founded in 2017 but started filing for patents in 2019, has 16 published patents and applications in optical computing including one for optical matrix processors that was granted in just eight months.
Despite active innovation as indicated by recent patent filings, Lightmatter has not announced any new fundraising since its $33M Series A round in early 2019. The majority of that funding came from Alphabet GV, Google's venture capital arm. The speed and compute power of optical computing is particularly useful for implementing machine learning/artificial intelligence software such as Google's popular, open-source TensorFlow package. Other notable startups in the field include Lightelligence and Ayar Labs, which recently raised $35M in a Series B roundfor a total of $60M in funding. A look at the patent portfolios of each company does indicate differentiation. Ayar Labs focuses on optical elements such as lenses and lasers. Lightelligence focuses specifically on artificial neural networks, while Lightmatter has developed matrix-based processors. These startups have been more active in recent patenting than established hardware companies like IBM or Qualcomm, which suggests they are strong contenders for acquisition or licensing. You can view more startups, as well as larger corporations, working on optical computing on the Magic Number® Patent Radian®.