Vcselrevolution: IMEC professor Van Campenhout explains their challenges clearly in this article.

Interesting read, vcselrevolution! I especially like this paragraph that demonstrates why silicon photonics solutions are costly:

• But if we want to reach our cost target, additional efforts are needed. In particular, the assembly and packaging of the technology – in particular, the interconnection and alignment of the different building blocks – raise the cost.

What I don't like so much is the progress they made in the monolithic arena:

• Or, on a longer term, monolithically integrated lasers grown directly on the substrates. This way, we do not longer need a laser-to-wafer assembly. In 2015, we took an important step forward. In collaboration with Ghent University, we demonstrated the first arrays of indium phosphide lasers monolithically grown on 300mm substrates.

It somehow escaped my attention that Imec announced this achievement last October. According to the wording, I think this was a just a first success that will still need much work and time to become production-ready. This is supported by van Campenhout speaking about "a longer term" regarding monolithically integrated lasers.

To put this into perspective, it was back in April 2011 when POET Technologies (Opel Solar at those times) announced their first laser. It took half a decade and a couple of milestones to develop the full POET device suite and to get where we are today. So Imec will also need some time. Like Suresh said, the window of opportunity is open today and for longer than just a couple of months. However, competition never sleeps and there will be more advances in the silicon world.

The question is which developments will come, how disruptive™ they are, and when they will come. Regarding the Imec laser, I'd really like to know how the POET team compares it to their own lasers.