GE Aerospace's $12 Billion Castings Purchase: Owning the Engine's Hidden Chokepoint

In late 2026 the world's biggest maker of jet engines did something it had not done since becoming a standalone company: GE Aerospace agreed to pay $11.75 billion for Consolidated Precision Products (CPP), a supposed specialist in the least glamorous part of an engine — castings. The stock barely budged, which is exactly why most retail readers can see the headline and miss what is happening.

The deal is GE buying the narrowest node in its own supply chain. Understand that node, and the "$12 billion" question — is it worth it? — stops being a mystery.

The part that holds the whole engine hostage

Stick with the unglamorous details, because this is the entire story. Inside a jet engine, the parts that do the hardest work are the turbine blades and vanes — the rows of airfoils that sit directly in the hot gas stream and spin the shaft. They have to hold their shape at temperatures that would melt ordinary steel, so they are cast from expensive superalloys, grown as a single crystal, laced with internal cooling channels that only a casting can form.

Investment casting is the process that shapes them: build a wax pattern, coat it in ceramic, melt away the wax, pour in the alloy. It sounds like a craft, and in a way it is. Complexity makes failure common — porosity, inclusions, a solidification mistake — and every defective blade is money burned, because the raw alloy and the ceramic tooling are already spent. There is no quick route around a short yield. The foundries that can do this reliably, at scale, and with the paperwork to prove each blade to a regulator are extremely few. By one prominent Wall Street estimate, Howmet Aerospace holds more than half of the industrial gas-turbine blade casting market, and it is one of only a handful of shops in the world.

That is the textbook shape of a chokepoint, not a mere capacity squeeze. A bottleneck evaporates when someone adds a line. A chokepoint stays because qualification is slow, failure is expensive, and the customer has no practical route around it — you do not casually certify a new foundry for a flight-safety blade.

Why GE, of all companies, is paying up

GE should be the last buyer that needs to be told this. Castings have been an engine-programs supplier to GE for more than fifteen years, feeding the LEAP, GEnx, T700 and F110 programs. GE knows the constraint from the inside: it has been raising engine output while the foundry layer behind it barely grows. The numbers make the tension concrete. GE projects airfoil demand will climb more than 30% by 2030 across commercial engines, aftermarket spares and defense — three demand streams hitting the same thin layer of foundry capacity at once. One analyst at Jefferies described the market as in "war games mode."

What GE is paying for is not just output but ownership of the node. CPP, founded in 1991 in Cleveland and now owned by private-equity firms Warburg Pincus and Berkshire Partners, is one of the world's largest aerospace casting manufacturers, with about 6,600 employees across more than twenty facilities, and roughly 60% of its revenue from commercial aerospace. GE will fold it in rather than hope a supplier with other customers and other allegiances prioritizes GE's ramp. That is the same logic driving Elon Musk's SpaceX, which has said it plans to cast its own gas-turbine blades for data-center power — still a planned move, with Howmet remaining a current major supplier — but even a giant's ambition to own the foundry underscores that the node is real, not a figment of a supply-chain diagram.

The price and the tension

Now the honest part: GE is paying a lot, early, at the top of the cycle.

The $11.75 billion works out to roughly 26 times CPP's projected 2027 EBITDA before integration benefits — roughly 18 times once it harvests the ~$200 million of annual synergies it expects. GE says the deal adds to adjusted earnings per share and free cash flow in the first full year, and it is funding it with about $7 billion of cash plus new debt. To size it against the parent: GE generates around $8.4 billion of trailing free cash flow, so a single check of $11.75 billion is roughly a year and a half of free cash flow going out the door for one foundry business. GE's shares trade at a high multiple — a P/E in the high 30s — and its dividend barely yields half a percent, so none of the enthusiasm is cheap.

This is where the "worth it" question splits in two. The structural case is sound: a chokepoint that cannot be quickly substituted is exactly the kind of dependency worth owning, not renting from a concentrated supplier base. Buying capacity at premium multiples is rational if capacity is the binding constraint on a demand wave that is genuinely long-dated — and defense backlogs and multi-year commercial engine programs are as long-dated as industrial demand gets.

The cyclical case is the risk. The price reflects 2027 expectations — a point in time when both the castings "war games" scarcity and the co-investment by others (SpaceX, Pratt's suppliers, Howmet's own expansions) could relieve the very scarcity GE is paying up for. Pellucid, unglamorous, and defensible as CPP is, no one pays 26 times EBITDA for a business without already pricing in flawless execution. If the constraint migrates — if capacity arrives faster than the 30% airfoil growth, or a key program's ramp disappoints — GE owns an expensive foundry, plus the debt used to buy it, at cycle top.

None of this makes the deal wrong. Control of the gating step in the engine supply chain is a genuinely structural asset, and GE's 15-year relationship with CPP lowers the integration risk that kills most deals like this. But "the bottleneck is real" and "worth 26 times EBITDA" are two different claims, and the market's shrug on announcement day — a stock roughly flat even as rival Howmet slid, then down over the following session — is the market pricing exactly the gap between them. The node is confirmed. Whether GE converts ownership into enough extra output and cash to justify the multiple is the part only the next few quarters of engine deliveries will answer.