A technology leader today can usually point to somewhere the cost equation is starting to move. A task that takes less time than it did a year ago. A report that took a week and now takes a day. An engineering assumption that no longer looks fixed.
What is much harder to answer is what those gains mean for the total economics of running and evolving a large technology environment over the next few years.
And that question has a deadline. Budgets, investment plans, and vendor commitments are being made now, before the full impact is clear.
Last issue we suggested watching what leadership intends to do with productivity gains. What we keep running into since is one level harder. Even when leaders can see a gain clearly, translating it into a future financial model is still out of reach. The budget calendar does not wait for the answers.
Some of the gains are already visible
An engineering executive we spoke with recently walked through how he plans offshore delivery. For years his working math was roughly one onshore person to three offshore. As the tools improve, he believes that could eventually move closer to one to two. That is his directional planning assumption, not a market fact, and he would be the first to say so.
He also said something that mattered more than the ratio: some of what looks like labor savings, he expects, will shift into token and tooling costs. The line item shrinks in one place and grows in another. Separately, he is thinking through where stronger senior talent fits into that mix.
The point is not whether the ratio holds. It is that he has to start putting assumptions like these into future planning before he can know whether they will.
The economics do not move evenly
Large established companies already operate complex technology environments: architecture, infrastructure, data, security, integrations, legacy systems, vendor commitments, production support, and a portfolio of continued investment. Faster code generation is one changing input inside that much larger economic system. The other pieces do not get cheaper because one input got faster.
A CIO we spoke with expected a program to retire an aging system to produce a meaningful savings number. Timing changed, and the expected benefit was cut roughly in half. He then had to find the difference somewhere else inside an eight-figure technology budget, while his organization continued funding an ERP replacement, platform investment, new technology products, and the engineering attached to revenue work.
That was not an AI story. It was a savings assumption meeting a real budget. When one number moved, everything around it had to move too. That interdependence is the environment every new efficiency lands in.
Where does the benefit go?
Here is what makes those two conversations more interesting together: they are two leaders inside the same company, describing different kinds of efficiency at the same time.
One is modeling what better tools could do to engineering ratios, with part of the savings expected to shift toward tooling. The other is managing savings from platform changes while continuing to invest in systems and products elsewhere. Same technology budget, different gains, different destinations.
That is the broader pattern we keep seeing: the money does not necessarily leave technology. The benefit may move rather than disappear. A lower total technology budget is one place the gain could ultimately show up. It is not the only one.
The budget is due first
Companies cannot wait for these economics to settle.
Budgets for next year are being built right now. Vendor commitments are being signed. Decisions about which capabilities to fund, which systems to replace, and which teams to grow have dates attached. The engineering executive above said plainly that he has clear visibility into some of his resource needs about six months out, and that next year’s budget still has to be built and approved.
So the assumptions go into the spreadsheet anyway. A ratio that might not hold. A savings number that might erode. A cost that might move somewhere new. The planning cannot pause until the answers arrive.
We are not going to predict whether technology budgets ultimately go up or down. The gains are real. How they change the full economics is still being worked out, and next year’s budgets are being built now.
We publish what we’re seeing. Send us a note with what you’re seeing.
