What is the demand picture for the trades?
Unusually strong at the top and unusually specific about where it lands. US data centre power demand is projected to climb from 31 gigawatts in 2025 to 41 gigawatts in 2026 and 66 gigawatts in 2027, with year-over-year capacity additions of 13.6 gigawatts in 2026 and 36.3 gigawatts in 2027 against realized additions of 6.4 gigawatts in 2024 and 8.5 gigawatts in 2025 (Goldman Sachs Research, 20 May 2026).
That is a step-up in annual capacity additions of more than four times between 2025 and 2027, and the share of total US peak summer power demand attributable to data centres is projected to rise from 4.1% in 2025 to 5.3% in 2026 and 8.5% in 2027 (same source). For anyone doing electrical, mechanical, civil, site or interconnection work, that is the demand curve underneath the next two years.
The same research carries its own discount and it matters more to a contractor than to an investor. Historically only about 72% of scheduled data centres go live on time, and the authors adjust to roughly 60% on time over the next year and about 50% over two years, while still projecting 11.5 gigawatts in the final three quarters of 2026 alone (same source). Schedules slip; the work does not disappear, but it arrives later than the announcement implies.
The wider frame is consistent. Engineering and construction leads value growth across industrials and services at plus 64%, while overall industrials and services values are forecast up 9% and volumes down 7%, with engineering and construction among the sub-sectors whose volumes are forecast to decline in 2026 (PwC, Global M&A trends in industrials and services 2026 mid-year outlook, 23 June 2026, on LSEG data to 31 May 2026). Fewer, larger transactions is the pattern.
What does a buyer look for in a backlog?
Margin, contract form and concentration, in that order, and not size. A large backlog at thin margin on fixed-price terms concentrated with one general contractor is worth less than a smaller backlog at good margin on cost-plus terms spread across ten customers, and buyers price the difference precisely.
Margin quality is the first test because it is the one that can be verified against completed jobs. A buyer will compare bid margin against as-built margin across a multi-year population, look at the size and frequency of change orders, and examine how disputes have been resolved. A contractor whose bid margin and realized margin track closely has demonstrated estimating discipline, which is the operating capability being purchased.
Contract form determines who carries risk. Fixed price transfers cost, schedule and escalation risk to the contractor; cost-plus and time-and-materials do not. A backlog of fixed-price work signed before a period of material cost increases is a liability wearing the clothes of an asset, and that is exactly the situation a buyer is looking for.
Concentration in the trades is usually programme concentration rather than customer concentration, and owners under-report it for that reason. Three separate contracts with three separate general contractors, all on the same data centre campus or the same public programme, is one exposure rather than three. The right disclosure names the ultimate end client and the programme, not the counterparty on the paper.
| Item | What raises the price | What lowers it |
|---|---|---|
| Margin quality | Bid margin and as-built margin tracking closely across several years, with small and infrequent change orders | A persistent gap between bid and as-built, or margin that depends on disputed change orders |
| Contract form | Cost-plus and time-and-materials work, or fixed-price work with escalation protection | Fixed-price work signed before a period of cost increases, with the contractor carrying escalation |
| Concentration | Work spread across end clients and programmes as well as across contracting counterparties | Multiple contracts that trace back to one campus, one programme or one general contractor |
| Bonding | A surety introduced early, with a written view on post-closing limits | Capacity resting on the departing owner's personal indemnity, tested for the first time after signing |
| Workforce | Documented crew retention, training throughput and a second layer of estimating and project management | Heavy sustained overtime covering a staffing gap, with estimating and supervision concentrated in two or three people |
“Owners lead with the backlog number because it is the biggest number they have, and the buyer has stopped listening by the second slide. What they want is bid margin against as-built margin across four years, contract form by value, and the end client behind each job. Produce that unprompted and you are a different kind of seller from the one who produces a total.”
Why does bonding capacity matter so much?
Because it determines what the business can bid for after closing, and a change of control can put it at risk. Surety capacity rests on the balance sheet, the working capital position, the track record and, in many cases, personal indemnities from the owner. Every one of those changes in a sale.
The practical failure mode is specific and common. A buyer acquires a contractor, the personal indemnity behind the surety line falls away with the departing owner, the surety re-underwrites the programme against the new ownership, and the single-job and aggregate limits come back smaller than they were. The business that was bought cannot bid for the work it was bought to win.
The right sequence is to bring the surety into the process early rather than to treat consent as a closing condition. That means an introduction to the buyer before exclusivity, an understanding of what balance sheet the surety will require post-closing, and a written view on what limits will be available. It is uncomfortable, and it is considerably less uncomfortable than discovering the answer afterwards.
We should be clear that no publisher issues a series on surety capacity in middle-market construction transactions, so we cannot tell you how often this goes wrong or by how much. What we can say is that it is the diligence item most likely to change the shape of a transaction in this sector, and the one most often left until last.
What about the workforce?
It is the constraint on converting demand into revenue, and buyers now diligence it as seriously as they diligence the backlog. A contractor with a bidding pipeline it cannot staff is not a growth business, it is a business with a queue.
The measurable version of this question is retention and productive capacity: crew retention by trade and by tenure, the age profile of licensed and certified personnel, apprenticeship or training throughput, overtime as a share of hours, and the use of subcontracted labour to cover shortfalls. A business running persistent heavy overtime is signalling either strong demand or a staffing gap, and the buyer will want to know which.
Supervisory depth is where the key-man question lives in this sector. Estimating, project management and field supervision are the roles that carry the margin, and where those sit with two or three people the risk is concentrated regardless of how many field staff the business employs. Documented estimating methodology and a second layer of project management convert that into transferable capability.
The labour question also determines which buyer wins. A strategic acquiring for capacity is buying the crews and the licences, and will pay for them. A financial buyer is underwriting whether the business can grow, and will discount a staffing constraint that limits it. That divergence is a reason to run a process containing both.
How should an owner in the trades think about timing?
By separating the demand cycle from the negotiating cycle, because they are not currently aligned. The demand data above is the strongest in years and extends into 2027. The transaction market is running fewer, larger deals with a thinner sponsor bid.
The relevant near-term constraint is that the sponsor buyer is weaker than it was. US private equity deal value fell 38% to $177 billion in Q2 2026, the lowest in two and a half years, while corporates raised a five-year high of $53.6 billion in leveraged loan activity (Sikich, Q2 2026 Credit Market Update, 13 July 2026). In this sector that is close to neutral, because strategic acquirers buying capacity have historically been the strongest bidders for specialty trade businesses anyway.
The medium-term constraint is the schedule risk in the demand itself. If roughly half of scheduled capacity arrives on time over a two-year horizon, then a business whose forward story depends on announced projects rather than signed contracts is carrying a timing risk the buyer will not accept on the seller's terms. Signed work prices; announced work does not.
The preparation that helps most is unglamorous. A multi-year bid-to-as-built margin analysis, a contract-form summary of the backlog by value, an end-client map behind the customer list, a written surety position and a workforce retention record. Every one of those is a document rather than a claim, and in a sector where buyers have been disappointed before, documents are what get paid for.
As of August 2026
Sources: Goldman Sachs Research, US Data Center Power Demand Projected to Double by 2027, 20 May 2026, for US data centre power demand rising from 31 gigawatts in 2025 to 41 gigawatts in 2026 and 66 gigawatts in 2027, for year-over-year capacity additions of 13.6 gigawatts in 2026 and 36.3 gigawatts in 2027 against realized additions of 6.4 gigawatts in 2024 and 8.5 gigawatts in 2025, for data centres rising from 4.1% to 5.3% to 8.5% of total US peak summer power demand across 2025 to 2027, and for the authors' own adjustment that historically about 72% of scheduled facilities go live on time, reduced to roughly 60% over one year and about 50% over two, while still projecting 11.5 gigawatts in the final three quarters of 2026; PwC, Global M&A trends in industrials and services 2026 mid-year outlook, 23 June 2026, on LSEG data to 31 May 2026, for engineering and construction leading value growth at plus 64%, for overall industrials and services values up 9% and volumes down 7%, and for engineering and construction volumes forecast to decline in 2026; Sikich, Q2 2026 Credit Market Update, 13 July 2026, for US private equity deal value falling 38% to $177 billion, the lowest in two and a half years, and corporates raising a five-year high of $53.6 billion in leveraged loan activity. No publisher issues a series on surety capacity in middle-market construction transactions, and none is quoted here. The backlog, bonding and workforce diligence framework is drawn from our own mandate practice. Companion articles on this site cover customer concentration, the working capital peg and how buyers underwrite backlog-driven businesses generally.