An AI-power investment theme

Who turns AI demand into profit?

Google’s power deals give AEP, Bloom Energy and Constellation different ways to make money. Following the theme means checking what each company earns after its costs.

This is the work TickerTrac helps you keep up with: connect your investment thesis to new evidence, inspect the source, and preserve why your view changes.

Compare the three company cases ↓

The example thesis

Long-term AI power demand can drive profitable growth for power producers and equipment suppliers, if selling prices cover delivery, operating costs and the capital required.

TickerTrac editorial example · Prepared from the linked public sources. This is a selected, dated reading, not an automated app assessment, live monitoring feed or investment recommendation.

The demand evidence is stronger

Google has made 20-year agreements to buy power from two separate projects. That supports the demand premise. The profit case still needs testing.

AEP · Announced October 14

702 MW

A planned Wyoming plant using Bloom fuel cells.

Full operation expected by end-2028.

Read AEP’s announcement

Constellation · Announced October 6

890 MW

Added output from upgrades to existing nuclear plants.

First addition expected by 2028.

Read Constellation’s announcement

MW means megawatts of generating capacity. These are announced plans and targets, not power already delivered.

Bloom’s sale is part of AEP’s cost

Both can benefit, but the same project reaches their profits differently. Constellation has a separate balance between upgrade costs and added power sales.

American Electric Power (AEP)

Sell electricity

A new plant using Bloom fuel cells would sell power to Google. AEP plans about $4 billion of investment in the project.

The profit test

Power payments need to cover equipment, construction, financing, fuel and running costs.

Next checkpoint and source

Watch the plant’s cost and delivery schedule, then the return earned on the capital invested.

The $4 billion is planned project investment, not a disclosed Google contract price or Bloom revenue. Full operation is expected by the end of 2028; it has not been delivered yet.

Bloom Energy (BE)

Supply the equipment

Bloom supplies the fuel-cell technology for AEP’s plant. Its equipment revenue forms part of AEP’s project cost.

The profit test

Equipment sales need to leave a margin after manufacturing and delivery costs.

Next checkpoint and source

Watch whether orders become profitable deliveries. A bigger project does not establish Bloom’s profit on it.

This historical agreement establishes the supplier relationship. It is not a new October 2026 order or a disclosure of the current project’s equipment price or margin.

Constellation (CEG)

Sell added nuclear output

A separate Google agreement supports upgrades at 11 existing nuclear units, adding 890 MW of generating capacity.

The profit test

Extra power sales need to justify upgrade spending and operating costs.

Next checkpoint and source

Watch upgrade costs, completion dates and added output. The first addition is expected by 2028.

These upgrades are separate from AEP’s plant and from Constellation’s agreement for 2,700 MW of existing output. The announcements do not establish which company earns the best return.

What would challenge the thesis?

Delays, cost overruns or weak equipment margins could absorb the gains from rising demand. The announcements do not disclose enough detail to rank the companies’ returns. A rising share price alone does not settle the question.

Keep checking your own investment idea

Start with one company or a broader change you’re following. TickerTrac connects your saved thesis to evidence and its sources, so you can see what supports or challenges it over time.

Start free with your own thesis

No card required. Free includes one monitored company and one Investment theme. This example stays editorial; your own thesis begins when you save it.