AEP's AI Power Trade Is Real — But Someone Left the Balance Sheet Running the AC on Full Blast

# AEP's AI Power Trade Is Real — But Someone Left the Balance Sheet Running the AC on Full Blast

Here's the thing about American Electric Power right now: it's simultaneously the poster child for the AI power boom and a cautionary tale about what happens when you finance that boom on a maxed-out credit card. Both things are true at once, and the stock's recent whiplash proves it.

Let's rewind to July 30. AEP put up Q2 2026 numbers that read like a Rorschach test depending on which line you stared at. GAAP earnings cratered to $713 million, or $1.31 a share, down from $1.226 billion ($2.29/share) a year ago — a 43% haircut. Strip out the noise and operating EPS came in at $1.36, which sounds better until you realize that still missed consensus by nearly 9%. Meanwhile revenue beat by about 3.5%. So: top line strong, bottom line soft, and yet management had the nerve to walk out and raise full-year guidance to $6.25–$6.55 a share, up from $6.15–$6.45.

That's not a contradiction — that's a utility telling you it knows exactly why the quarter looked messy and doesn't care, because the real story is the demand curve. Data center load grew 15% in the quarter. Contracted large-load demand through 2030 is now sitting at 69 gigawatts, up from a 56 GW forecast that itself only got set a couple quarters ago. This is the same table everyone else in the power sector is trying to get a seat at — the AI-and-cloud buildout turning utilities into de facto infrastructure plays for hyperscalers — and AEP just moved its chair closer.

So why isn't this an easy buy?

Because growth this size doesn't come free. AEP is running a capital program north of $70 billion — reports have pegged it as high as $78 billion — to build out the generation and transmission needed to actually serve that 69 GW pipeline. That's an enormous amount of steel, permitting, and labor to execute flawlessly, and it's being funded while free cash flow is running negative (-$126 million as of June) and debt-to-equity sits at a genuinely alarming 160.7x. That's not "growth company reinvesting aggressively" leverage — that's a company that needs the capital markets to stay friendly and the regulators to stay cooperative, for years, without a hiccup.

And the market has started to notice the tension. Mizuho trimmed its target to $135 from $141 — not a call that the growth story is fake, but a signal that the multiple had gotten ahead of itself. Meanwhile Goldman is still at $142 with a Buy, and the broader analyst tape reads Moderate Buy. Technically, the stock has broken below its 20-day average, is sitting well off its highs, and chart-watchers see a path down toward $128–130 before it finds real footing, with resistance stacked up at $135–140. Translation: this is a stock in a tug-of-war between a genuinely exciting demand story and a balance sheet that's earned some skepticism.

Where does that leave you?

Not chasing it here, and not abandoning it either. The dividend — a 2.97% yield with a manageable 56% payout ratio — gives you a reason to stay patient rather than a reason to load up. If shares slide toward the $125–128 zone, that's the level where the risk/reward starts looking like it favors the buyer instead of the storyteller. Above $130 with signs of stabilization, this becomes a name to accumulate toward the $140 area as the capex program starts converting into rate base instead of just cash burn.

The bear case isn't "AI demand for power is fake." It's "AEP is trying to fund a $70-billion-plus buildout on a stretched balance sheet while GAAP earnings are going backwards, and the stock still wants credit for the bull case before any of that de-risks." Until free cash flow turns or debt metrics improve, this is a Hold — a real story, an unreal amount of leverage, and a price that hasn't fully priced in the difference.