Some may recall the discussion here about the problems Facebook had with their initial 100G deployment. The lasers were dead-on-arrival with very high failure rates. That was attributed to the lack of burnin and I expect it was exacerbated by the need to maximize the output of the lasers to achieve the necessary signal strength for the receive circuit.

POET is not only building and testing at wafer scale but they also burn in at wafer scale. So they are not restricted in those time consuming steps that are otherwise required. They can maintain high throughput without taking shortcuts. When this attribute is combined with a waveguide stack that does not require an isolator which creates a restriction and increased losses along with the low loss waveguides we end up with significantly lower laser power requirements. 

Remember also that traditional silicon photonics  companies need an efficient remote light source…so POET has something for everybody...enough said. Have a great weekend!