IBM just warned Q2 revenue and earnings will miss estimates.
The stock dropped about 25% after the warning.
Large enterprise deals are not closing fast enough.
Meanwhile AI spending keeps pulling budgets away from traditional software.
Legacy businesses like COBOL still bring in money.
New AI coding tools keep eating into that advantage.
IBM keeps pushing hybrid cloud, enterprise AI, and quantum.
The enterprise AI backlog keeps growing.
Investors want revenue now, not backlog later.
The power side looks more interesting.
AI data centers keep running into grid limits, long connection queues, high utility costs, and power shortages.
An off-grid campus solves several problems at once.
Rolls Royce gas turbines already provide reliable baseload power without depending on overloaded grids.
Cheap natural gas access becomes part of the business model, not just a utility bill.
Build the data center.
Generate the electricity.
Lease around half the capacity to hyperscalers that cannot find enough power.
High utilization turns a cost center into recurring revenue.
That is exactly the kind of predictable cash flow Wall Street usually rewards.
Scale it to gigawatt levels and the monthly revenue potential reaches into the billions.
The upfront costs would be enormous.
Land.
Gas turbines.
Cooling.
Transmission.
Construction.
Competition from pure AI infrastructure players would be fierce.
IBM still has one advantage.
Enterprise customers already trust the brand.
The AI race may end up rewarding whoever controls power and capacity before whoever builds the next chatbot.