We empower Artificial Intelligence and deep learning technology companies to be more agile, innovative, and efficient.
Our capacity can be enlarged to accommodate rapid growth in cloud and edge computing.
Revolutionary architecture accelerates Artificial
Intelligence and deep learning technology for mobile edge
Our technology was designed to enable Artificial Intelligence and neural computing from a ground-breaking approach, unlike traditional market solutions.
We unify neural computing convolutional networks and recurrent networks on one system, allowing imaging and voice processing in parallel.
We transform the complexities of Artificial Intelligence and neural computing hardware into a simple platform, improving performance through efficiency.
⇣ Transforming the complexities of Artificial Intelligence hardware into a simple platform, to make Artificial Intelligence Productizations possible. ⇣
Powered by APiM (AI Processing in Memory) architecture, Lightspeeur® has 28K parallel computing cores, requires no external memory for AI interface, and can eliminate huge data movement.
With superior energy efficiency (9.3TOPS/w) level, Lightspeeur® accelerates AI edge computing and datacenter machine learning up to several orders of magnitude faster than the competition