The Future Has No Backlog

What the data-center boom can teach us about evaluating opportunities before the revenue arrives.

Weathered wooden double door in Lisbon with two cast brass fist knockers throwing hard afternoon shadows
3½ Things · September 2026 · Growth Strategy
Photo by Jigar B. Desai, Lisbon, Portugal

Data centers are having an interesting year.

Texas has ordered a verification and audit of data-center projects before they advance further through the Electric Reliability Council of Texas (ERCOT) interconnection process. New York paused discretionary state environmental permits for qualifying new hyperscale projects for up to a year. Yet CBRE reported record capacity under construction across its primary North American markets in the first half of this year.

All of that is true at the same time, which sent me looking under the hood.

This is not an economic forecast. Numbers depend on what we measure, when we measure it, and what sits underneath the average.

Numbers support more than one story. They do not support every story.

I have been in civil and architectural engineering long enough to know that the markets driving our work today will not be the same ones driving it tomorrow.

From May Edition 01

In May, I wrote Kupa-Manduka, the Frog in the Well: A Parable for AEC Growth and ended by asking how we were envisioning 2031.

This edition picks up the next question: when someone sees an opportunity with an early use case and a vision, how does a firm evaluate it before the revenue arrives?

Thing One

One industry. Several realities.

Data card titled These four numbers describe the same industry: plus 33 percent data centers 2026 construction spending forecast; minus 0.3 percent all nonresidential buildings 2026 construction spending forecast; 46.6 Architecture Billings Index for July 2026, where below 50 means billings fell; and 12 months median reported engineering firm backlog
AIA July 2026 Consensus Construction Forecast · AIA/Deltek Architecture Billings Index, July 2026 · ACEC Research Institute Q3 2026 Engineering Business Sentiment Study

The four signals above point in different directions. That is exactly the point.

AEC is not moving as one market. Architecture remains under sustained pressure. Engineering firms report resilient workloads and durable backlogs. Overall nonresidential building construction is soft. Data centers sit at the other end of the spectrum. AIA describes the outlook as K-shaped, with different parts of the built environment moving in different directions at the same time.

The architecture downturn is now the longest in the history of the index.

3½ years.

The four signals also measure different points in the cycle. ABI looks ahead, while backlog reflects work already sold. Engineering backlogs are also supported by public infrastructure that does not move with the private building cycle. That helps explain how weakness and resilience can coexist.

The point is neither that data centers are about to collapse nor simply that firms should diversify. The more interesting question is what happens inside a firm when one part of the market presents unusually convincing evidence that it deserves attention.

Thing Two

Today’s work comes with a scoreboard.

I have been on both sides of the table of an opportunity discussion. One opportunity arrives with clients, revenue, backlog, and utilization. Another arrives with an idea, conviction, and early evidence that the market may be changing. Both ask leadership for a decision. Only one comes with numbers the organization already knows how to score.

Earlier in my career, I carried ideas into the room, asking for resources with no revenue attached. Somebody said yes. Today I am often the person influencing or deciding, and the argument in front of me is the one I used to make.

The current business arrives with revenue. What comes next arrives with an idea, early evidence, and conviction that something is changing.

Revenue, margin, utilization, backlog, client concentration, and repeat work are useful measures. Mature businesses should be accountable to them.

But use those same measures too early, and a new opportunity faces a circular test: prove that you are already a business before we give you enough room to find out whether you can become one.

The standard for evaluating new opportunities should not be lower. It should be different.
Thing Three

Make the idea earn its next step.

An emerging opportunity should not get a free pass just because someone calls it “innovative,” “futuristic,” “this thing will smoke the competition,” or “game-changing.” I have no issue with being early. Just do not tell me ten other firms have already done this.

Give the new opportunity an owner with resources and P&L responsibility. Then ask three questions.
  1. What evidence would make this more credible? Look for the same problem described independently by more than one client, and for at least one of them willing to pay to solve it.
  2. What would have to be true for it to scale? Name what breaks first when one client becomes ten.
  3. What would make us stop? Write that condition down now, while nobody is invested in the answer. It is what keeps conviction from quietly turning into attachment.

Create a gated, venture-type approach that decides: if we achieve this milestone, then what?

A mature business proves that it performs. An emerging opportunity should prove that it is becoming more real.

AI can lower the cost of thinking bigger. Use it to pressure-test the early use cases, expose the missing evidence, and ask what would have to be true for the idea to scale. Then the judgment comes back to people. A written stop condition does something else, too. It means nobody has to oversell an idea to get it heard.

A new opportunity does not need mature-business proof. It needs evidence that it is becoming more real.
Lantern-lit vaulted stone corridor at the Jerónimos Monastery, narrowing toward a distant doorway
Jerónimos Monastery, Belém, Lisbon, Portugal
½ Thing — Still Working Through

Who is actually looking around the corner?

Every organization has people who believe they see something others do not.

Some are early. Some are wrong. Some are distracted. In real time, they look remarkably similar.

That is the part I have not worked out.

How do I know which person in the room is reading a weak signal correctly before everyone else can see it?

The hardest leadership skill may be recognizing good foresight before hindsight makes the answer obvious.
How do you tell the difference before the numbers arrive?
Ever Upward,
Jigar
Read on LinkedIn

Sources

(AI + MI) × HI™ is an original framework developed by Jigar B. Desai.

← All Editions Prev: Nobody Told the Concrete

Get 3½ Things Editions

Get the next edition delivered. No spam.

or
Subscribe on LinkedIn