Infinera is a great example of a company that's using photonic integrated circuit compound semiconductor technology, so, yes, as they further their research they'll have great results especially in the long haul where compound semiconductors excel, namely InP(Indium Phosphide).

POET is trying to bridge the gap that is III-V & Si(silicon). POET uses GaAs(compound), Gallium Arsenide is usually found in short to medium length, POET is going for the very-short to short-reach market as first point of entry. What makes POET different you ask? It's the natural performance of GaAs coupled with the monolithical integration of the POET process, but most importantly is the cost of goods.

Take Infinera for example, the speeds and efficiency they achieve is great for Long-Haul, DCI, Metro, but the technical specs & economies don't work at very short reach-medium reach. They connect service providers to service providers, ocean to ocean and data centres to other data centres, they don't connect the servers within the data centres.

POET is talking about up to 500m cables.

http://www.poet-technologies.com/docs/POET-Transceiver-White-Paper.pdf

Infinera is talking about cables ranging from 500KM to 9000 kilometres <----

http://www.advaoptical.com/en/innovation/100g-transport/100g-metro.aspx

https://www.infinera.com/telstra-validates-advanced-coherent-toolkit-9000-km-trans-pacific-cable/
...

So, POET has the same advantages as Infinera in terms of using a Compoud-Semi Conductor, it's a different medium with different capabilities than Infinera's but it's still taking advantage of the natural properties of Compoud-Semi Conductor properties, and the intergration level is where you get the costs and performance improvements.

That being said, note that POET always said they could do InP as well, and the Denselight aquisition points to furthering InP technologies. But they are definitely not there yet..