AI data centers may wait until 2029 for power.
Equipment suppliers can earn before then.
That's why GE Vernova $GEV is on my watchlist.
Wärtsilä's new 282 MW U.S. data-center power order has delivery planned for 2028 and full operation expected in 2029.
This isn't a GEV contract. It illustrates the investment cycle: suppliers can turn the AI power shortage into orders, deliveries and earnings before the data center starts operating.
GEV already has substantial demand behind that thesis:
$176B total backlog
53 GW gas-equipment backlog
63 GW additional slot reservations
>$5B YTD Electrification data-center orders
Important: reservations are not firm orders.
But strong demand doesn't automatically make $GEV cheap.
At roughly $981, GEV was about 20.9–22.7× FY2026 guided free cash flow. Customer advances can also support cash flow before delivery, so I don't treat all of that as recurring cash economics.
My decision: WATCH.
What I want to see:
Reservations → firm orders → profitable deliveries → durable cash flow.
Project delays, reservation attrition, weaker margins or poor recurring cash conversion would weaken the thesis.
AI's power shortage can be real, the supplier can benefit, and the stock can still be a bad investment if you overpay.
Equipment suppliers can earn before then.
That's why GE Vernova $GEV is on my watchlist.
Wärtsilä's new 282 MW U.S. data-center power order has delivery planned for 2028 and full operation expected in 2029.
This isn't a GEV contract. It illustrates the investment cycle: suppliers can turn the AI power shortage into orders, deliveries and earnings before the data center starts operating.
GEV already has substantial demand behind that thesis:
$176B total backlog
53 GW gas-equipment backlog
63 GW additional slot reservations
>$5B YTD Electrification data-center orders
Important: reservations are not firm orders.
But strong demand doesn't automatically make $GEV cheap.
At roughly $981, GEV was about 20.9–22.7× FY2026 guided free cash flow. Customer advances can also support cash flow before delivery, so I don't treat all of that as recurring cash economics.
My decision: WATCH.
What I want to see:
Reservations → firm orders → profitable deliveries → durable cash flow.
Project delays, reservation attrition, weaker margins or poor recurring cash conversion would weaken the thesis.
AI's power shortage can be real, the supplier can benefit, and the stock can still be a bad investment if you overpay.