Intel today announced its intent to purchase Omnitek, a U.K.-based video and vision field programmable gate array (FPGA) provider, as the chipmaker doubles down on its FPGA processor business. Terms of the deal were not disclosed.
Omnitek was founded in 1998 in Basingstoke, England, and all 40 of its employees are expected to join Intel when the acquisition closes.
Over the course of two decades the company has developed more than 220 FPGA internet protocol cores and the accompanying software. Its technology adds high-performance vision and artificial intelligence (AI) capabilities on top of its FPGAs. Use cases for these include video conferencing, projection and display, and medical vision systems.
According to Intel, Omnitek’s technology will complement its own FPGA business. The chipmaker intends to leverage Omnitek’s services for video, vision, and AI inferencing applications on its own FPGAs.
Intel has been investing heavily in its FPGA business group, which is a part of its Programmable Solutions Group that was formed in 2015 following its $16.7 billion acquisition of Altera. According to reports, Altera and Omnitek have partnered previously so the integration of the groups’ technologies should be fairly straightforward.
The chipmaker claims that the opportunity for silicon is around $300 billion and programmable solutions comprise about $8 million of that opportunity.
Just this month, Intel rolled out new FPGAs, deemed the Intel Agilex FPGA family that will serve compute-intense workloads like AI across the network, cloud, and edge computing. The Agilex products combine the FPGA fabric that is built on Intel’s 10-nanometer (nm) process technology and a heterogeneous 3D system in a package technology for integration.
This is not the first acquisition that the company has made to boost its programmable group — in July 2018 it bought programmable chip company eASIC, which builds structured ASICs that act as an intermediary technology between ASICs and FPGAs. At that time, Intel said the demand for FPGAs, and the reason for its focus there, was driven by their programmability which makes them ideal for cloud, IoT, and 5G networks.