The diner sits two blocks from the plant gate, and Rosa has run it for nineteen years — long enough to read the local economy in breakfast orders. When the big facility down the road is hiring, the 6 a.m. crowd doubles and the lunch rush needs a second cook. When orders slow, the booths empty out and the regulars start ordering coffee instead of the full plate. Rosa doesn’t follow earnings calls. She follows eggs. (She’s a composite, but every factory town has a Rosa, and every Rosa knows exactly what I’m about to describe.)
This week, the news from the industrial giants gave every Rosa in America reason to pay attention — because the biggest bets in business right now aren’t on ideas. They’re on bottlenecks: the physical choke points where the modern economy either flows or jams.
Start with the largest of them. GE Aerospace is buying Consolidated Precision Products — CPP — for nearly $12 billion, its biggest deal since the GE breakup, according to The Daily Upside’s September 9 edition. CPP makes the castings that go into jet engines; it has supplied GE Aerospace for more than fifteen years and will keep what GE calls a “significant book of third-party business” with Honeywell, RTX, and Lockheed Martin. CPP expects roughly $2 billion in revenue in fiscal 2027 — about 60 percent commercial aerospace, about 20 percent defense — which means GE is paying around 26 times projected 2027 earnings. That is a full price by any standard. You don’t pay 26 times earnings for a casting supplier unless you believe the castings are the whole game.
And the numbers say they might be. GE’s backlog stands at $210 billion. Commercial engine orders were up roughly 50 percent in the first half of the year. And in the second quarter, spare-parts delinquencies and delays rose 20 percent, according to CFO Rahul Ghai — the unglamorous metric that tells you demand is outrunning the supply chain’s ability to deliver. CEO Larry Culp put the logic plainly: the casting capacity is “needed to support the strong simultaneous demand across commercial engines, aftermarket and defense.” Translation: the bottleneck isn’t the order book. The bottleneck is the metal. So GE is buying the bottleneck. The deal is expected to close in the second half of 2027 — and notably, it follows Elon Musk suggesting that SpaceX may develop in-house casting capacity of its own, as its AI unit faces a power-turbine shortage. When the rocket company and the engine giant are both scrambling for the same foundry capacity, you are looking at a genuine choke point.
Now pan across to defense, where the bottleneck trade is even more explicit. European Commission President Ursula von der Leyen announced that EU members have approved €3.2 billion — roughly $3.7 billion — of the €90 billion Ukraine Support Loan to buy Patriot interceptors, and it is Lockheed Martin that makes the PAC-3 MSE interceptors those euros will purchase. That followed the United States awarding Lockheed a seven-year contract in late July for up to $53.86 billion in missile interceptors, bringing the total multi-year contract value to $58.62 billion — a commitment sized to let interceptor production capacity roughly triple by the end of the decade. Sweden, meanwhile, is buying HIMARS rocket systems in a deal worth about $728.8 million, according to its defense minister Pal Jonson.
Wall Street has noticed. UBS upgraded Lockheed Martin to a buy and raised its price target to $674 from $581. The company’s second-quarter sales rose 11 percent year over year to $20 billion, and it lifted its full-year revenue forecast to as much as $81.75 billion. The stock is up about 8 percent this year — modest against the market’s wilder stories, but built on something those stories often lack: contracted, government-backed, multi-year demand for things that cannot be downloaded.
And then there is the bottleneck behind all the other bottlenecks: the machines that make the chips. ASML has inked deals for its advanced high-NA EUV lithography systems with Samsung, which agreed to adopt the technology by 2028, and TSMC, which agreed to adopt it by 2030. Here is the staggering fact, again from The Daily Upside’s September 9 edition: ASML holds a 100 percent commercial monopoly on EUV lithography. Every advanced chip on earth — the AI accelerators, the phone processors, the guidance systems — passes through machines that exactly one company knows how to build. ASML’s shares are up 126 percent in the past year and closed at $1,764.85 on the news, up 2.91 percent on the day; the average analyst target price of $2,414 implies another 40 percent of upside, per Zacks. Separately, ASML, Intel, Samsung, and TSMC agreed to double the standard photomask size from 6 to 12 inches — the kind of unglamorous industry coordination that quietly unlocks the next decade of chipmaking.
Step back, and a pattern snaps into focus. GE isn’t buying growth; it’s buying castings. Governments aren’t buying weapons platforms so much as the interceptors that make the platforms meaningful — the consumable, the replaceable, the thing you run out of. And the entire semiconductor industry is, functionally, a tenant of a single Dutch company. The investment thesis of 2026 isn’t “who has the best idea.” It’s “who controls the thing everyone else runs out of.”
For Rosa at the diner, this is neither abstract nor academic. Bottleneck economics is what decides whether the plant down the road expands or consolidates, whether the second shift comes back, whether the breakfast crowd doubles. A $12 billion bet on casting capacity is, translated into her language, a bet that the orders keep coming faster than the metal — and that somebody’s town gets the foundry. Concentration like this cuts both ways for communities: the work is well-paid and durable, but when one supplier or one contract is the whole story, a town’s fortunes can turn on a single boardroom decision in a city far away.
There is a broader caution worth voicing, too. Bottlenecks are wonderful assets until they aren’t — until capacity catches up, until a substitute appears, until the cycle turns and the choke point becomes the glut. Anyone paying 26 times earnings for a casting supplier is making a decade-long bet that demand stays ahead of supply. History suggests humility about such bets, even well-reasoned ones.
But there is also something genuinely hopeful here, and it belongs to the Rosas. In an economy increasingly made of tokens, prompts, and valuations unmoored from revenue, the bottleneck trade is a bet on the physical world — on metal poured into molds, on interceptors that must actually be manufactured, on machines so precise they print circuits with light. You cannot fake a casting. You cannot hallucinate an interceptor. The industrial build-up now underway is, among other things, a reminder that the real economy still runs on real things, made by real people, in real towns.
It’s not enough to just chase the next breakthrough — we must listen to what the bottlenecks are telling us about where demand truly lives, learn to value the unglamorous capacity the world can’t do without, and contribute our attention to the communities where that capacity actually gets built.
Rosa will keep reading the economy in breakfast orders. The rest of us would do well to learn her method: watch what people actually need, follow it to its source, and never forget the hands at the other end of the supply chain.
