Mike Shepard

First of all, you know, the

The 164 gigawatts deviation to 160 gigawatts is just 2.4% drop.

And the way power companies work, they try to double the projected average.

And this is not a significant drop in terms of power production.

However, the fact that the days of the industry is actually growing, not by any ways diminishing,

is because of the fact that data centers are not just bottlenecked by power, but also by water, by connectivity, by human capital, by policy, taxation, and so many other reasons.

That's why, for example, workforce, there's simply not enough people out there to run data centers.

But the most important reasons that this report could be misleading is two reasons.

First, chip maturity, because chips are evolving so fast.

That it became so difficult for day center owners and operators to build day centers fast enough that before they become obsolete, it usually takes 24 months to build a day center.

And chips were changing on a monthly basis, practically.

But in 2026, we're predicting that a lot of maturity and clarity will be provided.

Yeah, but one of the main reasons that I don't find this report too relevant is because it's totally ignoring on-site power generation.

Because of the slow development of the power companies and them meeting the demand needs of the days of the operators, a lot of days of companies are reverting to on-site power generation.

If you want to grab a gas turbine, you can't find anything in the market today.

Everything is backlogged.

So a lot of people are reverting to on-site power generation, whether it's gas, solar, other means, hydro, and including SMRs.

This report totally dismisses the fact that a lot of data center industries rely on on-site power generation and not on the grid.

And that's why you really can't depend on the report.

Yeah, I mean, the maturity process, governance, everything.