tgroenwals shared this post · 44m ago
Justin R.

Nearly half are over budget on AI.
Most fund it anyway.

That's what Futurum's 2H 2026 survey of 1,636 IT decision makers found, and it's the pattern I'd test for in any transformation portfolio.

46.9% of enterprises are running AI spend over plan.
Among them, the common response is to ask for more budget or absorb it into next year.

The issue isn't the model.
It's that nobody was asked to own the cost of running it.

Four questions to test your own programme:
1️⃣ Who owns total cost to run, not just to build?
2️⃣ What's the ceiling before someone has to say stop?
3️⃣ Which value metric proves the spend, other than usage?
4️⃣ Who signs off scale-up, and on what evidence?

If two have no name against them, the budget is a forecast, not a control.

KPMG's Q2 2026 pulse adds a pattern worth watching.

Firms with full visibility into AI operating costs report established ROI at 15%, against 3% without it. That's correlation, not proof. It still points at the same fix.

Prove measurable value, not just higher spend.

Transformation no one can see is not transformation.

🔔 Follow Justin R. for more enterprise transformation insights.
🚀 Subscribe to The Transformation Constant — my newsletter — https://lnkd.in/d_mtyWzr

20
Nelson Uzenabor A budget is a forecast, not a control is the line that stuck. If nobody owns the cost of running AI, overspend is just a matter of time.
Justin R. Author Nelson. Spot on, clear cost ownership turns AI budgeting from planning into real financial control.
Pooja Jain You cannot manage an AI portfolio with an annual IT budget spreadsheet. You need real-time AI FinOps, token telemetry, and hard architectural ceilings!!! Justin R.
Justin R. Author Absolutely, token usage and architectural limits can give teams stronger control over AI costs, Pooja.