RTX, and the Interceptor Nobody Can Build Fast Enough

RTX is the house's first defense name, and we are not buying it as a war trade. We are buying a production-capacity problem, and the one prime that owns the hardest part of it. The binding constraint in air and missile defense is not demand, which is overwhelming and visible to everyone. It is the physical ability to build interceptors fast enough to replace the ones being fired, and that capacity takes years and dollars to add. RTX owns the stack that sits inside that constraint: the Patriot system, the GEM-T interceptor, the LTAMDS radar, the Standard Missile family, AMRAAM, and the Coyote counter-drone effector. We went long on 16 October 2026 and we intend to hold for three to five years.

The number the market stares at is the backlog: a record $289 billion at the end of the second quarter, up 22% year over year. The number the market mostly misses is that a large and growing share of the defense half is munitions and air-defense effectors, the one category where the order does not convert into revenue by signing a contract. It converts when a second production line opens, when a solid rocket motor supplier adds a shift, when a component that had one domestic source gets a second. RTX spent the first half of 2026 doing exactly that, and told investors it had more than doubled year-over-year output across its critical munitions. That is the engine of this thesis. Backlog is the promise; capacity is the delivery; and capacity is what compounds over our horizon.

We are honest about what this is not. It is not cheap. RTX closed at $193.92 the day we opened, which is above the average sell-side target of about $228 only modestly and well above where several of the fundamental models say fair value sits. It is not a pure defense play: roughly 60% of the backlog and nearly half of revenue is commercial aerospace, which carries its own cycle and its own unresolved problem in the Pratt & Whitney geared turbofan. And it is not a catalyst trade. We report into earnings on 20 October, four days after we opened, and we are telling you plainly that we are not positioned for that print. We are positioned for the conversion of a multi-year order book that the current multiple only partly credits.

RTX Corporation is three businesses under one roof, and the distinction matters because only part of the company is the thesis. In the second quarter of 2026 the three segments were close to the same size: Collins Aerospace at $8.21 billion of sales, Pratt & Whitney at $8.89 billion, and Raytheon at $8.27 billion, against total company sales of $24.7 billion, up 14% and up 16% organically.

Collins makes avionics, interiors, actuation and the connective systems of an aircraft. Pratt makes engines, commercial and military, and is the home of the geared turbofan. Both are aerospace businesses, levered to air travel, new aircraft build rates and the aftermarket. Raytheon is the defense business: integrated air and missile defense, effectors, radars, and the land and naval systems that use them. When we say RTX owns the bottleneck, we mean Raytheon specifically. The air-and-missile-defense franchise is what we are underwriting, and the other two segments are, for our purposes, a well-run aerospace company attached to it that funds the dividend and carries a known risk we discuss in section 10.

This structure is why RTX is not Lockheed Martin or Northrop Grumman, which are closer to pure defense. RTX is a hybrid, and the hybrid is both the reason the stock is less of a one-way bet on conflict and the reason a reader has to separate the parts to see the thesis. We are not asking you to like all three businesses equally. We are asking you to recognise that the defense segment sits on top of a constraint the other primes do not own in the same way.

At the end of the second quarter RTX reported a company backlog of $289 billion, split $170 billion commercial and $119 billion defense. The headline is the size. The structure underneath it is what we care about.

First, the backlog grew 22% year over year, which is faster than revenue, so the book-to-bill is comfortably above one and the order book is lengthening rather than draining. Second, the defense portion is increasingly weighted toward effectors and air defense, the categories pulled by live conflict and allied rearmament rather than by long-cycle platform programs. Raytheon booked over $5 billion of GEM-T interceptors in the first half, mostly international, and took its first domestic GEM-T production order in more than thirty years. AMRAAM awards were $1.8 billion, AIM-9X $1.1 billion, and the Coyote counter-drone effector saw output more than double on Army and Navy demand.

Third, and this is the part the $289 billion figure understates, management pointed to five framework agreements that are not yet in backlog, which it expects to drive significant demand over the next decade if the budgets and multi-year funding come through. In other words, the record backlog excludes a layer of probable future orders that sits in the exact categories where capacity is the constraint. For a three-to-five-year thesis, the backlog is the floor, not the ceiling, and the thing that converts it is capacity, which brings us to the constraint itself.

Here is the house question applied to defense: where is the structural constraint, and who owns it? In air and missile defense the constraint is not the radar and it is not the launcher. It is the interceptor, and specifically the rate at which interceptors can be manufactured.

A Patriot battery can be delivered in a couple of years. A GEM-T or a Standard Missile cannot be surged in the same way, because the supply chain behind it runs through a small number of solid-rocket-motor lines, specialised energetics, seekers and precision components, several of which had a single qualified domestic source. You cannot buy your way out of that constraint in a quarter. Adding a second source for a rocket motor is a multi-year qualification program. This is why, through 2024 and 2025, allied governments discovered they could sign for interceptors far faster than anyone could build them, and why the United States began funding capacity expansion directly.

RTX spent the first half of 2026 attacking this constraint in the only way it can be attacked, which is slowly and with capital. Management said it had more than doubled year-over-year output across critical munitions, is investing an additional $100 million domestically to speed up LTAMDS and component production, and is adding second and third sources in constrained areas, including European AMRAAM component suppliers and co-production of Patriot, Stinger and NASAMS with allied partners. The honest read is that even this is not enough: management itself said the next step up in rates will require "another significant step up." That sentence is the thesis in RTX's own words. The demand is a given. The scarce thing is the capacity, and the capacity is being built now, which is what a multi-year holder is paid for.

Owning a bottleneck is only valuable if the position is defensible. In interceptors it is, for reasons that are structural rather than commercial.

RTX is the prime on the Patriot system and the sole manufacturer of the GEM-T interceptor, which it is now co-producing with European partners to raise rates rather than because the design is open. It makes the LTAMDS radar that is replacing the legacy Patriot radar, the Standard Missile family for naval air defense, AMRAAM for air-to-air, and Coyote for the counter-drone fight that Ukraine and the Red Sea have made a permanent line item. These are not interchangeable products. They are qualified into allied force structures, certified through years of testing, and tied to launchers and fire-control that are themselves RTX or RTX-integrated. A customer that standardises on Patriot does not re-compete the interceptor the following year.

The barrier, in other words, is the same thing that makes the bottleneck a bottleneck: qualification takes years, so an incumbent who is already qualified and already expanding capacity captures the demand that newcomers physically cannot serve in time. We looked for the counter-case, which would be a credible second source taking share, and in the interceptor categories that matter we did not find one at scale. What competition exists is at the margins and in co-production that RTX itself controls. That is what owning a constraint looks like, and it is the reason we are willing to pay a full multiple for it.

The catalyst is not a single contract. It is a replacement cycle that has only started, and the recent award flow shows its shape. We are deliberately listing dollars that are already signed or announced, not projected.

Recent air-defense and effector awards

Values are contract ceilings as announced, not revenue recognised. Multi-year awards convert across the life of the contract.

The language on the SM-3 award is the tell: it is explicitly to refill missile-defense inventories. That is the resupply cycle stated in a contract. Every Patriot engagement over Ukraine and the Middle East, every interceptor fired in the Red Sea, is a round that has to be rebuilt, and the rebuild runs through RTX. Management was careful on the Q2 call not to call this conflict-driven or use the word resupply; it said only that global events had reinforced the need for its systems and pointed to Middle East demand and rising NATO budgets. We will be less coy than the company's investor-relations language, while noting we are drawing the inference, not quoting them: a decade of under-stocked Western interceptor inventories is being rebuilt at the same time as new demand is being signed, and that is a multi-year revenue stream, not a spike.

The risk embedded in this section is the mirror image of the thesis, and we state it here rather than burying it: if the conflicts that drive the urgency ease, the pace of new orders slows. We do not think the installed under-capacity reverses on a ceasefire, because inventories still have to be rebuilt and allied force structures have already been re-planned around higher stockpiles, but a reader who believes peace is near should size this name accordingly.

RTX raised its full-year 2026 outlook alongside the second-quarter print, which is the behaviour of a company whose order book is converting faster than it planned. Guidance now calls for adjusted sales of $95 to $96 billion (from $92.5 to $93.5 billion), adjusted EPS of $7.10 to $7.25 (from $6.70 to $6.90), and free cash flow of $8.50 to $8.75 billion. Second-quarter free cash flow was $2.9 billion and adjusted EPS was $1.89, up 21%.

The balance sheet is investment-grade and carries real leverage but not dangerous leverage: roughly $37 billion of long-term debt against about $8 billion of cash at mid-year, so net debt near $29 billion, against a business generating more than $8 billion of free cash flow a year. The dividend is $2.92 a share annualised, a yield of about 1.5% at our entry price, and it is covered several times over by free cash flow. This is not a balance-sheet story and there is no solvency question here; it is the opposite of the going-concern names elsewhere in our coverage. The financial risk in RTX is not that it runs out of money. It is that the market has already paid for several years of the growth, which is a valuation question, not a liquidity one.

We are paying up, and we will not pretend otherwise. At $193.92 RTX trades at about 27 times the midpoint of 2026 adjusted EPS guidance, a low-thirties trailing multiple, and roughly 2.7 times sales. That is a premium to the market and a premium to RTX's own history, and at least one widely followed fair-value model puts intrinsic value closer to $155, well below the current price.

The sell side is more constructive but not euphoric, and the spread is instructive. The mean price target is about $228, roughly 17% above our entry. Bernstein sits at $214 and recently trimmed from $223, Barclays started at Overweight at $223, and Citi carries a Buy at $240, with BNP Paribas and RBC positive but in the low-to-mid $200s. So the most bullish target implies about 24% upside and the most cautious about 10%, over the next year. That is a respectable but not extraordinary return for a stock at this multiple, and it tells you the easy re-rating has happened. What is left is the multi-year compounding of a backlog that converts as capacity comes online, plus a 1.5% dividend, which is the return we are actually underwriting.

The honest framing is that at this price, time is working for us only if the capacity ramp delivers and the orders keep converting. Buy it expecting a quick move to $240 and you have mispriced the trade. Buy it expecting earnings to grow into the multiple over three to five years while the resupply cycle runs, and the entry looks reasonable. On this name, patience is the position.

We went long RTX common stock on Friday 16 October 2026. We hold shares, not options, and we hold them for a three-to-five-year horizon. The position disclosure at the top of this note gives the terms, and we will update it as a dated entry if anything changes.

One calendar fact has to be stated the way we state every uncomfortable one. RTX reports third-quarter results on 20 October 2026, four days after we opened and three days after this note publishes. We are not positioned for that print, and holding common stock through an earnings date is a very different thing from the short-dated options calendar that defined our IBM trade note. There is no expiry here and no premium decaying against us. If the quarter disappoints and the stock falls, our thesis is measured in years and a single quarter does not break it; if the quarter is strong, we were long into it by coincidence of timing, not design. We mention the date because a reader deserves to know a catalyst sits immediately ahead of them, not because it is the reason we are in the name.

Sizing follows the stance. Constructive, in our framework, is a real position taken with conviction but with the acknowledgement that the entry price already reflects a good deal of the thesis, so it is sized to be added to on weakness rather than to be maximal on day one.

This is where the note earns its keep. The thesis is not fragile, but it is not free of real risks, and three of them are specific.

The geared turbofan. Pratt & Whitney's powder-metal issue on the GTF engine is the one open wound on the balance sheet that is not about defense at all. It cost roughly $150 million in powder-metal-related compensation in the second quarter alone and has been a multi-billion-dollar overhang across the fleet-management campaign. It is being worked down rather than resolved, and a reader underwriting RTX is underwriting Pratt's execution on that remediation as surely as they are underwriting Raytheon's interceptors. It does not threaten the company, but it caps the free-cash-flow upside in the near term.

Valuation and the multiple. We said it in section 08 and we list it as a risk here because it is the most likely way this position disappoints over one to two years. At 27 times forward earnings the stock has little margin for a growth stumble. A missed capacity milestone, a slip in GTF remediation, or simply a rotation out of defense could compress the multiple faster than earnings grow, and the stock could be lower in a year while the three-to-five-year thesis stays intact. That is the scenario we are most prepared for.

The budget and the political cycle. Defense revenue depends on appropriations, and appropriations depend on a functioning budget process. Continuing resolutions, shutdowns and shifting administration priorities can delay the conversion of even signed contracts, and the framework agreements that are not yet in backlog depend entirely on multi-year funding actually being enacted. A peace settlement in Ukraine, a de-escalation in the Middle East, or a domestic pivot away from procurement would all slow the order pace, even if the installed under-capacity means they would not reverse the replacement cycle already underway.

Below those three sit the ordinary ones: concentration in a handful of large programs, the commercial aerospace cycle turning down and dragging Collins and Pratt, and execution risk on the capacity ramp itself, which is hard, capital-intensive and dependent on a supply chain RTX does not fully own. None of these is a reason to avoid the name. All of them are reasons to size it as a patient position rather than a conviction maximum.

The open questions we consider the next highest-value work on this name, published as openly as the conclusions.

1. 01

   Quantify the capacity ramp. Translate "more than doubled output" into unit rates where disclosed: Patriot, GEM-T, Standard Missile, Coyote. The thesis lives or dies on the slope of this line, and management has so far given multiples, not numbers.

2. 02

   Size the framework agreements. Put a probability and a dollar range on the five agreements not yet in backlog, and the funding path each depends on. This is the upside the $289 billion figure excludes.

3. 03

   Map the interceptor supply chain. Identify the single-source choke points behind GEM-T and Standard Missile, the solid-rocket-motor constraint in particular, and who else supplies it. The bottleneck is only as durable as its narrowest link.

4. 04

   Model the GTF drag precisely. Build the powder-metal remediation into a free-cash-flow bridge so the defense upside is not quietly offset by a commercial liability we have treated qualitatively here.

5. 05

   Watch the Q3 print and the capacity commentary on 20 October. Not for the quarter, but for any hard numbers on production rates and any change in the framework-agreement timeline.

6. 06

   Test the ceasefire scenario. Model what order flow and the multiple do under a Ukraine settlement, to pressure-test the claim that resupply survives peace.

• RTX Q2 2026 results, reported 23 July 2026
• RTX Q2 2026 earnings call transcript
• RTX records $289 billion company backlog
• RTX full-year 2026 guidance raise
• Raytheon $3.7B Patriot GEM-T interceptors for Ukraine, April 2026
• RTX seven-year Tomahawk contract, $22.9B
• SM-6 and SM-3 Block IB awards, October 2026
• Counter-drone LIDS and Coyote awards, US Army, Qatar, Jordan
• First domestic GEM-T production order in over thirty years
• RTX price, market capitalization and dividend data
• RTX short interest and ownership
• Analyst price targets: Bernstein, Barclays, Citi
• Pratt & Whitney GTF powder-metal remediation
• RTX balance sheet and Form 10-Q