Alternatives, mostly. When the mechanical share of engineering hours compresses, the freed time does not buy speed first. It buys comparison: design options that die unevaluated today get evaluated. A Swedish rail project just showed what a single such comparison is worth.
The question
Every infrastructure design is a chain of forks. Elevated or at grade. Concrete or steel. This corridor or that one. In principle, each fork deserves a priced comparison. In practice, most forks get one fully worked branch, because evaluating an alternative costs weeks of document work: load cases, quantities, cost estimates, environmental consequences, all produced and checked by hand.
So alternatives rarely lose on merit. They are simply never examined. What gets built is often the first solution that was affordable to describe in full.
Product developers call this point-based design: commit to one solution early, then defend it through every later stage. When ground conditions, traffic forecasts or budgets shift, there is no evaluated fallback to reach for, and the project absorbs the change at whatever it costs.
A live example
Roughly two billion kronor, in the most recent public case. Ostlänken, the new double-track railway between Järna and Linköping, was planned to run on an elevated structure through central Norrköping. In early 2026, a new analysis showed that the bridges would require considerably more ground reinforcement than assumed: a cost increase of just over 1.9 billion kronor against the financing agreement from 2022.
The standard ending to that story is absorption: the increase gets booked, the plan gets revised, the project rolls on. This time, Trafikverket and Norrköping municipality went back and re-evaluated an alternative late in the process. In March 2026, they announced a new planning direction with the tracks at ground level, cutting the estimate by roughly two billion kronor. Planning will take somewhat longer; total implementation time is not expected to grow, because a ground-level alignment is simpler to build.
Note what made this news. Not the cost increase, cost increases are routine. The re-evaluation. A late, serious look at a design alternative is rare enough to warrant a press release. There is no clearer measure of what evaluation costs today.
The precedent
Toyota, for decades. In a 1995 study in MIT Sloan Management Review, Allen Ward, Jeffrey Liker and colleagues described what they called the second Toyota paradox: Toyota kept sets of design alternatives alive far longer than Western carmakers, delayed committing to one, and still developed better cars faster. Exploring more options did not slow the company down. It removed the expensive rework that follows from locking in a design that later proves wrong.
Infrastructure research has arrived at the same remedy. Jonas Eliasson, director at Trafikverket and visiting professor at Linköping University, concludes in a 2025 study of cost overruns that projects and designs should be kept in competition longer:
To solve this, more projects need to be explored than will eventually be carried out. Projects which turn out to have higher costs or lower benefits than anticipated are then cancelled, until only the best projects remain.
The obstacle was never the idea. It was the price of holding options open: every alternative kept alive is another set of documents to produce, update and review, and the fee covers one. We have written before about what happens when that price drops: one more iteration per project, at the same fee.
The answer
First, a precision about what the hours do not buy: less scrutiny. The point is not to shorten deliberation or hurry decisions through review. Every review stage stays. What changes is what each stage can afford to look at. The claim is more evaluation per krona, not faster approvals.
Taken seriously, the answers get concrete. The bridge project that prices one alignment today prices three. The environmental assessment that reacts to a finished design gets to test the alternatives too. Errors get caught while they are still lines in a document, not change orders on a site. And the Norrköping manoeuvre stops being a press release and becomes a standing item: cheap enough to do at every design review, in every project.
Europe is heading into a decade of very large infrastructure spending; Sweden alone fixed a national plan of 1,171 billion kronor for 2026–2037 this spring. The question is not whether the money will be spent. It is how many evaluated alternatives each krona passes on its way into concrete. Two billion kronor in Norrköping is what one late comparison turned out to be worth. The hours coming back are what makes the next thousand comparisons affordable.
Sources
Yesper is the AI civil engineer for construction and infrastructure. AFRY, COWI, NRC Group and other Nordic firms use it to halve the time on a study, rerun calculations in minutes, and catch errors that would otherwise slip through. Get in touch if you'd like to see what it can do for you.
Book demo
