Six Names, Three Very Different Businesses
Lumping these together as "construction stocks" hides what matters. They do fundamentally different work, and that's the whole game:
- MTZ — MasTec. Infrastructure across communications, power delivery, pipelines and clean energy. Direct data-center and grid exposure, boosted by its electrical-contractor acquisition. A self-perform builder with real skin in the projects.
- MYRG — MYR Group. Specialty electrical contractor: high-voltage transmission & distribution for utilities, plus commercial/industrial electrical work that increasingly means data centers. The purest "build the grid and the electrical guts" name of the six.
- ACM — AECOM. The odd one out — an asset-light design and consulting firm (transportation, water, environment), not a self-perform builder. It draws the plans; it doesn't pour the concrete. Different risk profile entirely.
- PRIM — Primoris. Utility and energy infrastructure — power delivery, pipelines, gas generation and (its problem child) renewables/solar EPC. A data-center electrical bolt-on is growing but not yet the story.
- FLR — Fluor. Diversified global E&C — infrastructure, mining, energy, and government/nuclear (a NuScale stake, Centrus fuel work). Broad, cyclical, and only incidentally an AI play.
- AGX — Argan. Builds gas-fired (and some renewable) power plants. The most direct "AI needs electricity, someone has to build the power plant" bet in the group.
Read that and the sort begins on its own: MTZ and MYRG are levered to the grid/data-center theme; AGX is levered to the power-generation piece of it; ACM, PRIM and FLR are diversified E&C names whose quarters had far more to do with their own project books than with anything happening in AI.
Where the Prices Are
The drawdowns here are more violent than the marquee names — and, crucially, more uneven. As of the August 24 close:
| Ticker | Price | Off 52wk high | vs 200-day | RSI(14) |
|---|---|---|---|---|
| MTZ | $255.36 | −42.2% | −15.6% | 35.7 |
| MYRG | $305.27 | −39.4% | −4.6% | 29.5 |
| ACM | $64.99 | −52.0% | −25.2% | 42.7 |
| PRIM | $75.14 | −63.4% | −40.4% | 36.7 |
| FLR | $51.28 | −11.5% | +8.1% | 48.4 |
| AGX | $468.98 | −41.8% | −7.8% | 31.4 |
Two things stand out. First, MYRG at RSI 29.5 is the only genuinely oversold name on the board. Second — and this is the tell — Fluor barely dropped (−11.5%, still above its 200-day) precisely because it was never priced as an AI story. The names that ran the hardest on the data-center narrative (PRIM, ACM, MTZ, AGX) are the ones that fell the hardest. The market is un-pricing a theme, and it's doing it most aggressively where it was priced most aggressively.
What the Quarter Actually Said
Here's the split that matters. Three of these delivered; three tripped over their own feet.
| Rev (YoY) | Adj EPS | Backlog | Fwd P/E | |
|---|---|---|---|---|
| MTZ | $4.37B (+23%) | $2.22 (+49%) | $21.4B (+30%) | ~27x |
| MYRG | $1.08B (+20%) | $3.17 (+86%) | $3.16B (+20%) | ~28x |
| ACM | $3.59B (−14%) | −$0.50* | $27.8B (+13%) | ~18x / ~12x† |
| PRIM | $1.69B (−11%) | −$0.27 | $13.9B (rec.) | ~32x |
| FLR | $4.33B (+9%) | $0.91 (beat) | $26.9B (−5%) | ~19x |
| AGX | $291M (rec.) | $3.24 (beat) | ~$2.9B (rec.) | ~45–67x |
*ACM's loss is entirely a one-time $337M charge on a legacy 2019 construction-management contract; ex-charge, its guidance is intact at ~$5.90–6.10. †~18x on reported guidance, ~12x on ex-charge earnings. AGX reports on an offset fiscal calendar; figures are its most recent quarter (ended April 2026), with no formal guidance issued. "rec." = record.
The clean operators: MTZ posted a record $21.4B backlog (+30%), raised full-year EPS guidance to ~$9.30 (+42%), and expanded margins — its lone blemish was missing EPS by a single penny. MYRG beat by 21% with EPS up 86% and margins expanding. Both are executing.
The wounded: AECOM's headline loss is a legacy-contract charge, not operational rot — ex-charge, it's a steady ~18x business trading at ~12x. Fluor beat on EPS but its backlog is shrinking (−5%) and it's still absorbing losses on the Gordie Howe bridge. Primoris is the worst: gross margin collapsed from 12.3% to 4.9% on renewables cost overruns, dragging the whole company to a loss despite a record backlog.
The Real Risk Here Isn't AI Demand — It's Execution
This is the lesson the quarter hammered home, and it's the one worth internalizing. In fixed-price construction, the demand side has rarely been better — every one of these companies is sitting on a record or near-record backlog. But three of the six just demonstrated how a contractor actually loses money: not because the work dried up, but because a job went sideways. AECOM's $337M charge, Primoris's renewables overruns, Fluor's bridge losses — none of those are demand problems. They're execution problems, the kind that live in change orders, underbid scopes, and fixed-price contracts that don't survive contact with the field.
That reframes how to value the group. In a sector where backlog is abundant, the premium belongs to the operators who convert it without detonating margins. That's MTZ and MYRG. The discount belongs to the ones proving they can't — and the trick is telling a one-time stumble (AECOM) apart from a recurring one (Primoris's serial renewables trouble).
The Verdict — Who I'd Back
Forward-looking, current-conditions ranking:
My pick: MasTec (MTZ). It's the best combination on the board — a record $21.4B backlog up 30%, guidance raised 42% on EPS, expanding margins, and genuine diversification across communications, power delivery, pipelines and clean energy, with direct data-center and grid exposure layered on top. It's a self-perform builder that just proved it can grow the book and convert it cleanly. At ~27x forward with that growth and that visibility, cut 42% from its high, it's the one I'd anchor to.
Best oversold bounce: MYR Group (MYRG). The cleanest quarter of the six — EPS up 86%, a 21% beat, expanding margins — and the only oversold chart (RSI 29.5). It's a pure electrical-contractor play on the grid and data-center build. The knocks: it maintained rather than raised guidance, its biggest transmission awards don't start burning until 2027+, and free cash flow was negative on timing this quarter. Smaller and less diversified than MTZ, but if you want the highest-quality operator at the most washed-out price, this is it. (I wrote it up standalone earlier this month — the thesis has only gotten cheaper.)
The contrarian value: AECOM (ACM). Cut in half, and the headline loss is a single legacy charge masking an asset-light design franchise that — ex-charge — trades around 12x with a record $27.8B backlog. It's the cheapest name here and the lowest execution risk going forward (it draws plans; it doesn't carry fixed-price build risk the way the others do). The catch: it's the least AI-levered of the six, so it's a value-and-mean-reversion bet, not a data-center bet. Fine — just know which bet you're making.
The high-torque speculative: Argan (AGX). The purest expression of "AI needs power, someone builds the plant." Record backlog, a clean beat, gas-turbine EPC demand tied directly to data-center load growth. But it issues no formal guidance, trades at a rich and noisy 45–67x, and its next print isn't until September. Highest torque to the power-generation theme, least visibility to underwrite it. A trade, sized like one.
Pass for now: Primoris (PRIM) and Fluor (FLR). PRIM has a real data-center bolt-on and a record backlog, but until it stops losing money on renewables execution, the backlog is a promise it hasn't shown it can keep — down 63%, and deservedly cautious. FLR held up best, but that's the point: a shrinking backlog and only incidental AI exposure mean there's little here to catch the theme's upside. Cheapest-looking, least catalyzed.
The Headwinds That Could Break This
- Fixed-price execution risk — the sector's signature killer. Half this group just proved it. Any of these names can turn a record backlog into a charge with one bad job. It's why the premium belongs to clean converters (MTZ, MYRG) and why "cheap" names (PRIM) can stay cheap.
- AI-capex digestion. The same demand fear pressuring the whole complex. The direct plays (MTZ, MYRG, AGX) have the most to lose if hyperscaler and utility capex plans slow; the diversified names (ACM, FLR) are insulated but also don't benefit as much.
- Skilled-labor scarcity. Every one of these is bidding for the same electricians, pipefitters and operators. Wage inflation is the quiet margin tax on backlog conversion.
- Backlog timing. MYRG's and MTZ's biggest awards largely burn in 2027+. The demand is booked; the earnings are a year or more out — and a year is a long time for the market's mood to change.
- Integration risk. MTZ (Superior) and PRIM (PayneCrest) are both digesting electrical-contractor acquisitions. Contractors lose money on integration as readily as on demand.
Bottom Line
This tier of the build is where the difference between a good contractor and a bad one shows up on the income statement, and this quarter showed it plainly. I'd own MasTec (MTZ) as the clean, diversified, guidance-raising core; keep MYR Group (MYRG) as the highest-quality operator at the most oversold price; treat AECOM (ACM) as a cheap mean-reversion bet on a one-time charge rolling off; and size Argan (AGX) as a speculative, high-torque play on data-center power. I'd wait on Primoris (PRIM) until it proves it can build renewables without losing money, and I'd leave Fluor (FLR) to investors who want cheap diversified E&C rather than AI exposure. In a field flush with backlog, the edge isn't demand — it's the operators who don't blow up the job.
Prices and technicals pulled live from Yahoo Finance as of the August 24, 2026 close (market closed). Q2 2026 fundamentals from company releases and BusinessWire/StockTitan/24-7 Wall St./Zacks coverage; AECOM figures are fiscal Q3 and Argan fiscal Q1 (offset calendars). Forward P/E figures are approximate. Levels and multiples are point-in-time and will move.