From a short NewsFactor article lacking in real detail, it seems that Intel has been beaten at its own game of silicon photonics by a start-up company named Luxtera. While Intel has only managed a tenth of the 10 Gbps speeds claimed by Luxtera, they have also said that products based on the on-chip-fibre-optic technology would not start arriving until the end of the decade. Luxtera says they have a complete, working product already.

Intel has declared its intention to build similar devices, but has so far only managed a tenth of the speeds Luxtera is claiming. Intel had also predicted that the first products would not be available until the end of the decade.

"Until now, all of the progress in silicon photonics has been in the form of research. Luxtera is the first company to develop a complete, working product that meets a market need," said Linley Gwennap, principal analyst at The Linley Group, a semiconductor technology analyst firm.

Another example of David vs. Goliath, if it pans out. I've read my share of ambitious announcements by start-up companies in the past from which I didn't hear anything later. However, it's interesting to note that Luxtera is making use of current CMOS technology for their silicon photonics implementation, meaning that production ought not be a problem, as most chip foundries use a CMOS process. They could definitely turn out to be a thorn in Intel's side.