GE Bought a Foundry, Not an Earnings Boost

GE Aerospace announced its biggest deal on Tuesday — $11.75 billion for a castings company called Consolidated Precision Products. By most financial measures, the deal barely registers. It adds roughly 1.5% to GE's 2028 EPS forecast of $10.25. The purchase multiple is steep: about 18 times projected 2027 EBITDA if you include expected synergies, closer to 26 times without them. If you're looking for a cheap acquisition that boosts earnings, this is not it.

GE doesn't want the numbers from CPP. It wants the furnaces.

CPP manufactures one of the most technically difficult components in industrial manufacturing: turbine blades and vanes. These are not ordinary metal parts. High-pressure turbine blades operate in gas streams hotter than their own melting point. They're cast from nickel-based superalloys as a single, uninterrupted crystal to eliminate the grain boundaries where stress fractures start. The process requires vacuum furnaces, ceramic molds built through repeated slurry dips, and controlled withdrawal rates measured in microns per second. One deviation and the whole blade is scrap.

That difficulty is the entire point. Western production of these components is concentrated in four companies: Howmet Aerospace, Berkshire Hathaway-owned Precision Castparts, Doncasters' DPC Holdings, and CPP. If you need blades or vanes and you're not one of these four, you're waiting — sometimes more than a year — before you can get them. OEM qualification takes years, and once a supplier is approved, it rarely gets replaced.

GE Aerospace is the world's largest jet engine maker, powering Boeing 787s and 777s and, through its CFM joint venture with Safran, every Boeing 737 and many Airbus A320neos. Its engine backlogs stretch into the 2030s. The company is also facing surging demand for spare parts as airlines extend the life of existing aircraft, plus defense orders that aren't slowing. GE's CEO Larry Culp said the acquisition is about securing "mission-critical casting capacity" to serve all three markets at once. The underlying constraint is simple: GE can design engines faster than it can get the physical parts to build them.

Now the bottleneck is getting even tighter, because a second demand curve has appeared. AI data centers are driving electricity demand so sharply that utilities are ordering gas turbines in volumes that foundries didn't plan for. The same blades and vanes that power commercial jets also run industrial gas turbines. Elon Musk called turbine blades and vanes in February 2026 "the real constraint on the AI infrastructure build-out." He's not wrong. While the market has spent years focusing on chip shortages, the limiting factor for powering data centers has been the physical foundry capacity to cast the turbines themselves.

The competitive response has been rapid and aggressive. SpaceX announced plans in September to build its own foundry in Bastrop, Texas, to cast gas turbine blades in-house — a vertical integration play that could accelerate turbine availability by 18 months. Howmet Aerospace's stock dropped roughly 11% following the announcement, as investors realized their largest customers can vertically integrate if the bottleneck gets expensive enough. The broader pattern is clear: OEMs who don't control their own supply are either waiting longer or paying more, and the suppliers who held pricing power are now facing the threat that their own customers will build the furnaces themselves.

This is the war game Jefferies referenced. The firm noted that valuation multiples for Howmet and DPC have already compressed as investors price in the risk of OEMs bringing casting in-house. The "Blades & Vanes War" isn't about one deal — it's about control over a capacity-constrained chokepoint where the only qualified players are four companies, and everyone else is trying to become the fifth.

What GE is actually buying from CPP becomes clearer against the financial backdrop. GE reported $9.8 billion in operating cash flow over the last twelve months, with free cash flow growing 53% year over year. It holds $9.3 billion in cash. The deal requires $7 billion in cash plus new debt financing for the remaining $4.75 billion. GE's total debt sits around $10.9 billion. The math works — but it's not the math that drives the call.

CPP brings roughly $2 billion in sales and approximately $450 million in pre-synergy EBITDA at 22.5% margins. GE expects $200 million in Year 3 synergies. None of that changes GE's earnings trajectory in any meaningful way. What changes is that GE gains direct control over one of four global sources of a component that has been physically limiting its ability to deliver on backlogs that extend past 2030. In a market where the foundry, not the design, is the constraint, buying a foundry is the move.

CPP also has a long track record with GE. GE has been CPP's customer for over 15 years. This isn't a company GE needs to qualify or restructure from scratch. The integration risk is lower than most acquisitions of this size.

The deal closes in the second half of 2027, subject to regulatory approvals. GE's shares traded roughly flat on the announcement — a muted reaction to an $11.75 billion acquisition that, on its financial surface, doesn't seem to move the needle. That reaction may be exactly the disconnect the deal is built on.

GE is priced at about 43 times forward earnings and trades at roughly 32 times trailing EV/EBITDA. These are premium multiples for a company that's already delivered strong execution — 21.7% revenue growth, 23% ROIC, and a track record of beating earnings estimates across multiple quarters. The stock has gained about 9% year-to-date and sits within striking distance of its 52-week high. It is not a cheap stock. But the CPP deal signals that GE's management is thinking about the next layer of constraint, not the next quarter's EPS. If the foundry bottleneck tightens further, or if competitors like Howmet lose customers to vertical integrators, the premium GE pays today for casting capacity could look like a bargain. If the bottleneck eases and demand softens, the $11.75 billion price tag — and the debt it adds — starts looking like overpayment for a constraint that no longer exists.

The central question for investors is whether this is a bottleneck that persists or one that opens. The evidence tilts one way: lead times exceed a year, qualification cycles run years, and the four qualified foundries are already booked. SpaceX building a foundry doesn't add industry capacity — it redirects it. GE buying one doesn't either. In a four-player market where every move is a zero-sum reshuffle of who controls which customer, the math of the deal matters less than the logic of the position.

GE didn't buy CPP to make a little more money. It bought CPP because the company that controls the furnaces controls the output.