CEG

Constellation was preferred because it had pulled back, is more established, provides relatively pure exposure to the nuclear thesis, and has stronger underlying metrics than more speculative alternatives.

The operational distinction that matters is availability. A reactor runs continuously, so it can serve loads that cannot tolerate interruption, and that is a different product from intermittent generation regardless of what either costs per megawatt-hour.

The demand argument is that AI and data-center load is growing faster than new generation can be built. Constellation's existing fleet is already operating, which places it on the favorable side of that imbalance. Capacity that exists today is worth more than capacity that requires a decade of permitting.

The value of that position depends on how much output is contracted and at what price. Long-term agreements convert a commodity generator into something closer to a contracted infrastructure business, but they also cap the upside if power prices rise faster than contracts reset.

A generator earns a spread. Revenue moves with power prices while much of the cost base is fixed, so the leverage works in both directions and quarterly results are lumpy. A weak quarter often reflects power prices and outage scheduling rather than any change in the asset base.

The real risks are an unplanned extended outage at a major unit, and power prices falling while the fixed cost of running a nuclear fleet does not.