AI capability is real. The economics still have to work.

We’re spending money on AI faster than we’ve figured out how to make money from it. Both things can be true at once.

By David Morris

3 MIN READ

I believe deeply in AI.

I also think we’re spending money on it faster than we’ve figured out how to make money from it.

Both are true for me, even though this conversation usually demands that you pick a side.

$30T¹

Global data center spending Reuters recently reported could be exceeded by 2050.

$4.2T¹

New revenue Bain estimates the companies building AI infrastructure will need over the next five years to fund it.

Those numbers made me pause.

What I keep seeing

I’ve spent much of this year trying to put AI to work inside actual companies. What I keep seeing is that most businesses don’t need a smarter model right now.

They need to decide:

Which work is worth changing

Who owns it

What good looks like

Whether the change made the business better

If the models stopped improving tomorrow, we would still have more capability than most organizations know how to use well.

That should be good news.

But it also means companies can no longer blame the model when nothing changes.

Two speeds

The market is building infrastructure at one speed. Companies are turning that capability into revenue, margin and cash at another.

At MPS, we keep coming back to a fairly boring sequence:

Pick a real business problem.
Redesign the workflow.
Name the owner.
Measure what changed.

The boring part is usually where the value is.

Real technology, unproven economics

I’m not ready to throw the baby out with the bathwater because some valuations are stretched. The technology is real.

But a real technology can still be surrounded by bad economics.

AI may eventually justify every dollar being invested. That doesn’t mean it will happen on the timetable the money is assuming.

THE QUESTION

If AI never improved from here, could you point to one place in your business where it is already producing an outcome worth paying for?

If you can — or if you can’t yet — tell us where you land.

SOURCES

Reuters, “AI’s race to transform the world before the money runs out,” October 3, 2026 — including the $30 trillion data center spending projection and Bain’s estimate of more than $4.2 trillion in new revenue needed over five years. Figures are reported estimates, not forecasts by MPS.
reuters.com/business/retail-consumer/ais-race-transform-world-before-money-runs-out-2026-10-03

AUTHOR

David Morris

Find the outcome worth paying for.

Twenty questions, about three minutes, and a 0–100 readiness score with the three things to fix first.

Take the AI Readiness Assessment → See the C-Suite Agents →