Quanta Services' AI Story Is Real — the Price Has Run Ahead of It

Quanta Services has been one of the market's loudest winners, and the pitch writes itself. AI data centers need enormous amounts of electricity, the aging U.S. grid cannot move it fast enough, and Quanta is the biggest specialty contractor in the country standing at that bottleneck. The shares are up roughly 50% so far in 2026 and about 70% over the trailing year, and that enthusiasm is now baked into a price that carries a real question for anyone thinking about joining it: what does the stock actually cost, and what would have to go right to earn that price?

The place to start is with what Quanta is, because the multiple only makes sense once you see the economics underneath it. Quanta does not own the power plants or the data centers. It builds and maintains the electric-transmission, pipeline, and network infrastructure on contract. That is a services business with thin margins — a gross margin around 15% and an EBITDA margin under 9% — where earnings ride on labor, project execution, and a steady pipeline of awarded work. Contractors rarely deserve premium multiples, because a handful of bad fixed-price jobs or a cost overrun on a large project can wipe out a year's profit.

None of that changes the fact that the demand is real and verifiable. In its second quarter, reported in late July, Quanta posted revenue of $9.56 billion, up 41% from a year earlier, with adjusted diluted earnings of $4.24 a share against a $3.31 consensus. Backlog reached a record $53.4 billion, up from $35.8 billion in June 2025, and management raised its full-year 2026 guidance across every key metric, calling for adjusted EBITDA of $4.09 billion to $4.21 billion and free cash flow of $2.0 billion to $2.5 billion. On the surface, this is a company firing on all cylinders.

What the price already assumes

Now the gap between that operating strength and the valuation becomes the whole story. At roughly $98 billion in market capitalization and a $103 billion enterprise value, Quanta trades near 34 times trailing EBITDA and roughly 74 times trailing earnings. Even against the raised 2026 guidance of about $4.1 billion in adjusted EBITDA, the enterprise value works out to around 25 times forward EBITDA.

Here is the comparison that matters. The market pays near 17 times EBITDA for MasTec, roughly 16 times for EMCOR, and about 15 times for Primoris — contractors with the same kind of work and the same exposure to the buildout. Quanta fetches roughly double that, a premium the bulls justify on faster growth and a larger share of the AI-electrification backlog. Growth does deserve a premium; the open question is whether it deserves all of it.

The cash-flow test makes the point more sharply than the multiple. Free cash flow for 2026 is guided to $2.0 billion to $2.5 billion, which against a $98 billion market cap is roughly a 2% to 2.5% free-cash-flow yield. The dividend yields a token 0.06%. In plain terms, a buyer today is collecting almost no current income and is depending on the multiple staying elevated while the company keeps compounding at rates only recently achieved. There is no cash cushion under this price — the entire return hinges on years of flawless execution.

Where the margin of safety went

Two details keep me from treating the premium as earned. First, a meaningful slice of the growth is being bought rather than grown organically: Quanta spent about $1.24 billion on four acquisitions in the second quarter and July, contributions that add an estimated $120 million to $140 million of adjusted EBITDA in 2026 — roughly 3% of the guided total. Acquisitive growth can be excellent, but it is not the same compounding as organic demand, and at 34 times EBITDA the price gives no credit for that distinction.

Second — and this is the part that decides it for a cash-flow investor — paying 34 times trailing EBITDA removes any margin of safety. My discipline is to establish durable cash flows first and then measure how far the price sits below what they are worth. Here the cash flows are genuinely strong, but the price is not below anything; it sits above what even generous peer multiples would justify, and the stock has no built-in buffer of yield or cash return to absorb a stumble. If data-center capex pauses, a few large jobs overrun, growth merely normalizes to the teens, or sentiment simply rotates, the multiple is the first thing to contract — and at 2% free-cash-flow yield, a lower multiple is not offset by income.

None of this says Quanta is a bad business or that the electrification buildout is a mirage; the backlog and the raised guidance are real evidence to the contrary, and long-term holders have been rewarded handsomely. But a wonderful company and a wonderful price are different things. At roughly twice its closest peers and near-record multiples, the market has already paid for the optimistic version of events. A newcomer buying today is underwriting near-perfect compounding with no margin of safety — the premium has become the risk, not the opportunity. For an investor who needs a cushion between price and value, Quanta at 34 times EBITDA does not offer one.