GE Vernova Inc. GEV

Key data

The business

GE Vernova builds the hardware that makes and moves electricity: gas and steam turbines (Power), onshore and offshore wind (Wind), and the grid gear, transformers and switchgear, that carries power to the wall (Electrification). It spun out of General Electric in April 2024. Power and Electrification are the profit engines. Wind, especially offshore, is the wound. Customers are utilities, independent power producers and, increasingly, the hyperscalers building data centers, who reserve turbine slots years ahead. The moat is the installed fleet: about a quarter of the world's electricity runs through GE machines, and every one needs parts and servicing for decades, a services book that pays through any cycle. Buyers put money down before delivery, so growth funds itself. Contract liabilities and deposits sat near $26 billion in June, cash the company holds and invests while it builds. What customers actually wait for is a slot on the sold-out H-class turbine line.

The numbers

Two of the last five quarters carried enormous net income on tiny operations.

The $13.39 and $17.44 prints did not come from selling turbines. Both were one-time tax benefits, deferred tax assets the company could finally recognize as it turned durably profitable. Strip them and the real story is the two clean quarters at each end: revenue up 22% year over year to a record $11.1 billion, gross margin widening to 21.3% from 20.3%, operating income nearly doubling to $653 million. That margin step is the one the business needed to show, and it showed, answering the open question from a week ago about whether the march toward higher margins was on schedule. But the clean quarter still disappointed. Q2 diluted EPS of $2.47 landed well under the $3.17 analysts modeled, the first real operating miss since the spin.

Here is the trap in the multiple. Trailing EPS of $34.94 makes the stock look like 27 times earnings, cheaper than either year since it turned profitable. About $30 of that $34.94 is the two tax gains. On the clean Q2 run rate, roughly $2.47 a quarter, under $10 a year, you pay more than 90 times what the business earns today. Consensus does not reach $34.76, where trailing already sits, until 2028, and $47 by 2029. So the bet is that real operating earnings more than triple in three to four years. The $176 billion order book, equipment-heavy and rising, is what makes that arithmetic plausible rather than fanciful.

The lever is operating margin. Q2 added $2.0 billion of revenue and $275 million of operating profit, a 14% incremental margin, held down by Wind losses and the Prolec transformer integration. The whole case needs that number closer to 25% as the backlog converts.

Management

Scott Strazik has run GE Vernova since the spin on a plain framework: fund the fleet, buy back stock, pay a small dividend. Buybacks are real, about $3.3 billion in 2025 and $1.3 billion more in Q1 2026, paid from deposit-rich cash flow rather than borrowing, with net debt negative. Insiders only sold this year, three officers for a combined $11.5 million, the CTO, HR chief and finance chief, plan status not disclosed. The sums are routine against their holdings and tell you little. The pay plan rewards adjusted EBITDA and cash, which aligns with the margin march but can flatter a year that leans on price over volume.

How it fails or surprises you

Data-center demand outruns the slots (right tail). Hyperscalers are reserving gas capacity years out, and the $176 billion backlog counts only firm orders. If reservations convert to signed equipment orders faster than the sold-out H-class line can show today, 2027 to 2028 revenue and pricing both step up. The tell: equipment RPO climbing past $88 billion and Power segment margins widening. The tape pays for growth, not yet for acceleration.

Offshore Wind keeps bleeding. The company still flags cost and schedule pressure across its existing backlog, and Vineyard Wind's timeline moved when Interior paused, then unpaused, offshore leases over the winter. Each troubled project is a fixed-price contract that eats cash. A single charge of a few hundred million would swamp a clean quarter's $653 million of operating income. Watch Wind segment losses and any new contract charge.

The earnings are not what the multiple says. Two-thirds of trailing EPS is one-time tax recognition, and the one clean quarter missed by 22%. If real operating EPS sits closer to $10 than the $34 on the tape, the stock is 90-plus times earnings, not 27, and any slip in backlog conversion re-rates it hard. The print that settles it: four consecutive clean quarters of rising operating EPS.

Closing thoughts

This is mostly a priced market. The $176 billion backlog is disclosed, the gas-for-data-centers thesis is consensus, and analysts already model EPS to $47 by 2029. The edge is not in knowing the order book exists. It is in judging whether operating margin climbs the way the backlog implies, and whether Wind stops offsetting it. On the other side of your position sits a holder who accepts a sub-1% real earnings yield today for that compounding. The left tail is a Wind charge or a gas-order pause. The right tail is the supercycle running hotter than the sold-out line can yet show. The downside is real money because you pay a supercycle price, and the upside needs the margin to follow the revenue.

The bet is still that the world keeps buying GE Vernova's turbines and grid gear to power data centers, and that the $176 billion already booked gets built at a widening margin. What breaks it is the two numbers that do not yet agree: revenue up 22% against an incremental operating margin of only 14%, with Wind the drag. When those converge the stock is cheap. If they do not, the tax-flattered 27 times is a mirage sitting on 90 times reality. Four clean quarters of rising operating EPS proves it true or false.