For years, POET was getting ready for the Module side of the business, as data loads were pushing the hyper scalers to push more bits through their centers due to video and the growth of cloud usage.

Then came the AI onslaught. This created a race among the bigs—Microsoft, Google, Amazon, Meta, and others. They see it as an existential threat.

Photonic circuits, including modules, and, more importantly, Celestial AI's type of deployments (not here yet, but very soon) will be a massive driver of photonics. 

Nvidia has taken the brass ring so far. Still, all those great chips need photonic interconnects for better overall efficiency and to break through specific bottlenecks (memory, CPU to GPU comms, energy, heat) inherent with electronic traces and wires.

So yes, POET has pivoted again, but we should all appreciate this magnificent pivot to a highly profitable sector. The bigs cannot wait; they are in that existential moment I mentioned. 

Nvidia is enjoying the first and most significant portion of the investments to win this race. A whole ecosystem of improvements is needed, and POET is right there. 

I am not technical enough to make the comparisons to predict the winner here. I'll defer to the investors in Celestial AI, the SHINE Institute, Dr. Michal Lipson, Theresa Ende, Arista Networks, Adva, etc., etc. These institutions and folks are better positioned to know. 

My take is that POET makes a chip that is less than half the size of my pinky fingernail, and it can have the latest and greatest modulators, lasers, and all the other stuff on it so it can grow with the new, new things as they come. The makers of those things now have a predictable way to insert their products into the systems, with no need to wonder how to integrate, as long as POET becomes what we hope for. 

What could beat it? I don't know. Making it a bit smaller seems irrelevant. The waveguides are an order of magnitude better performing. They were among the first to make a TxRx module, and it was a rapid path to it. 

I believe the industry will essentially accept that POET is the way to go for many of the needed deployments and that its flexibility removes another design headache. 

We'll forgive all the previous pivots, and the market's early disconnect that has us suffering.