Slowly, the year is coming to an end. After 2025, we are stepping into 2026.
The obvious question is: what will the next year bring?
AI is no longer new. We use it on a daily basis.
Tools exist. Models exist. Subscriptions are paid.
It’s fair to say that most companies are already past the experimentation phase.
And this is where my main point starts.
The AI bill comes due in 2026 — not in the form of new models, but in questions about value, cost, and ownership.
For the last few years, we’ve been buying almost anything that had “AI” in its name.
Copilot here. Cursor there. ChatGPT everywhere.
But when you ask a very simple question – what value do we actually get from this? – the answers are vague at best.
We know how much we pay.
We don’t know what we get back.
From a business perspective, that’s not sustainable.
The AI bill comes due in 2026 – and nobody has the numbers
One of the biggest problems we currently face is measurement.
Today, AI is often embedded into our SDLC without changing how success is measured. We still look at cycle time and throughput, but we don’t know how much of the change is caused by AI and how much by everything else.
There’s no clear correlation.
No reliable signal.
So we keep paying, hoping it helps.
Relying on gut feeling and intuition.
That won’t be enough much longer.
As leaders, we will be asked very direct questions:
What is the value AI gives us? What is the benefit? Where are the numbers?
Another discussion we won’t avoid next year is the one about roles.
I don’t believe engineers will suddenly disappear. What I do believe though is that the nature of engineering work will shift.
Less focus on raw output (eg. how much code is written) and more focus on outcomes:
- what functionality is delivered,
- does it actually solve a problem,
- what is the cost to maintenance.
That shift alone changes how we should think about seniority, growth paths, and team composition.
The same applies to hiring.
I don’t expect mass layoffs of developers next year.
What I do expect is slower hiring.
If delivery speed increases but product discovery doesn’t scale at the same pace, adding more people simply doesn’t make sense. The bottleneck moves elsewhere, to the product side.
Over time, this leads to fewer people writing code in the traditional sense, and more people who supervise AI and understand systems, products, and trade-offs.
This is also why maturity matters so much.
AI doesn’t magically fix bad environments.
If processes are weak and systems are messy, AI just adds another layer of complexity.
Teams that adapt their processes will gain speed.
Teams that don’t will amplify chaos.
To benefit from AI, organizations will have to:
- rethink their processes,
- remove unnecessary steps from the SDLC,
- define clear boundaries for where AI helps and where it hurts.
Otherwise, any short-term speed-up will be paid back with interest during maintenance.
Why engineering roles will shift when the bill arrives
Next year won’t be about AI replacing engineers.
It will be about adoption and ownership.
Engineering will move further away from execution and closer to decision-making.
Leaders will be forced to answer uncomfortable questions about cost, value, and outcomes.
Teams that can’t do that will slowly fall behind.
And in the long run, that gap will only grow.
That’s what 2026 represents for me.
Not a technological breakthrough.
A reality check.
Food for thought:
https://waydev.co/2026-tech-trends-a-guide-for-engineering-leaders/