RTX

IThe business

RTX makes things that are hard to certify, expensive to build and then need parts and work for thirty years. It is one of the largest aerospace and defense companies in the world, with about 180,000 employees, of whom about 54,000 are engineers, and 2025 sales of US$88.6B.³ It earns that money in three businesses. Collins makes the systems inside an aircraft. Pratt & Whitney makes the engines. Raytheon makes missiles, air defense and sensors. About 48% of sales go to commercial customers, mostly airlines and aircraft makers, and 52% to governments.³

It was not built in one step. In April 2020 United Technologies spun off Otis and Carrier, and on the same day merged with Raytheon Company in an all-stock deal, keeping its own aerospace businesses, Pratt & Whitney and Collins (formerly UTC Aerospace Systems, joined with Rockwell Collins in November 2018).¹¹ The company changed its name from Raytheon Technologies to RTX in July 2023.⁵ Every number on this page before 2020 has to be read with that history in mind, so where I show a ten-year series I add the pieces up as if they had always been one company, and I say how.

All figures are in US dollars. The price is the New York Stock Exchange close of 2 November 2026, the latest completed session when I wrote.¹² The latest results are the third quarter of 2026, reported on 20 October, and the 10-Q filed the same day.⁴⁶

Three ways it earns money

Building the original equipment. New engines, new avionics, new missiles. This is the part that is sold once and often at a thin margin or a loss, because the money is made later. In the third quarter of 2026 Collins commercial original equipment sales rose 23%, while Pratt & Whitney’s fell 5% on the mix of large engines, a reminder that the two businesses do not move together.⁴

Servicing what is already flying. Spare parts, repair, overhaul, upgrades and long-term maintenance contracts. The filings label this “services”, and it was US$24.4B in 2025, 27.6% of sales, and 27.6% again in the first nine months of 2026 C (calculated).³⁶ The company does not publish commercial aftermarket sales in dollars in the 10-K, so I use what is filed and the growth rates it reports. In the third quarter, commercial aftermarket sales rose 21% at Pratt & Whitney, driven by Geared Turbofan engines, and 9% at Collins.⁴ This is the annuity inside the business: an engine is sold once and then needs shop visits for decades, and its owner has few places to go.

Defense. Raytheon’s sales rose 14% in the first nine months of 2026, to US$23.2B, led by the Patriot, Standard Missile and Tomahawk programs.⁶ It also makes Coyote, a low-cost, rail-launched effector against drones that pairs with a radio-frequency sensor and is a key part of the US Army’s LIDS counter-drone system; in the third quarter Raytheon booked US$0.6B of LIDS for international customers.¹²⁶ Collins and Pratt & Whitney also supply the military, most of all the F135 engine, which is the only engine for the F-35.³ The government is therefore both a customer and a rule-maker, and it is the next section’s subject.

Sales in 2025 were 37.6% to the US government, 7.6% to foreign governments through the US government’s foreign military sales program, 6.9% to foreign governments directly, and 48.0% commercial C (calculated).³ Airbus alone was about 14% of RTX’s total sales, before discounts and incentives, and 29% of Pratt & Whitney’s.³ I come back to those two numbers in Risks, because they decide what a bad year looks like.

The order book

Backlog is the value of contracts signed and not yet delivered. It was US$340B at 30 September 2026, up from US$268B at the end of 2025: commercial US$176B and defense US$164B.⁶ Defense bookings alone were US$59B in the third quarter, most of it at Raytheon, whose own backlog went from US$75B to US$130B in nine months.⁶ The big awards were a US$23B Tomahawk contract and a US$14.3B AMRAAM contract, both booked in the quarter.⁶

Not all of it is settled money. The company counts a defense order as booked when it is firm, even if Congress has not yet appropriated the funds.³ The Department of War awarded four multi-year contract actions with ceilings of up to about US$23B for Tomahawk, US$20.7B for AMRAAM, US$24.4B for Standard Missile-6 and US$6.3B for Standard Missile-3 Block IB. Each is undefinitized, meaning the price is not yet fixed, and each is funded year by year by Congress.⁶ The Tomahawk and AMRAAM awards, US$23B and US$14.3B, are in the third-quarter bookings, so about US$37B of the US$340B is undefinitized and not yet appropriated.⁶ The Standard Missile awards, with ceilings of US$24.4B and US$6.3B, are not in the list of third-quarter bookings, and I do not count their ceilings. If the money is not appropriated, RTX says it has contractual recovery remedies, not a guarantee of the revenue.⁶ So the jump in backlog says more about what the government intends than about what RTX will deliver.

The state of play now

Third-quarter 2026 sales were US$25.0B, up 11% (12% organically), adjusted earnings a share were US$2.26, up 33%, and free cash flow was US$2.1B.⁴ The company raised its full-year outlook again: adjusted sales of US$97.0B to US$97.5B, adjusted earnings a share of US$7.65 to US$7.75, and free cash flow of US$9.0B to US$9.2B. The first outlook, in January, was US$92B to US$93B of sales, US$6.60 to US$6.80 of earnings and US$8.25B to US$8.75B of free cash flow.⁴¹⁶

Part of that raise is not repeatable. Third-quarter adjusted earnings include US$0.25 a share, and about US$0.4B of segment profit, from refunds of tariffs that the US Supreme Court ruled unauthorized in February 2026.⁴⁶ A second item sits the other way: Pratt & Whitney booked a US$363M charge for a customer bankruptcy, which the company excludes from adjusted profit.⁴ I count the first as a gift from the courts and the second as a cost of doing business with airlines, and I take both out when I look for what the business earns in a normal year.

IIThe moat

The honest test of a durable business is what it did when something went wrong. For RTX there are two answers, and they do not agree.

The first is demand. Between the end of 2020 and the third quarter of 2026 backlog went from US$150B to US$340B, and sales from US$56.6B to a 2026 outlook of about US$97B.¹³⁶⁴ Airlines kept flying the engines through a global grounding and a safety issue, and governments kept paying for the missiles through a change in administration. Customers do not appear to leave.

The second is execution. In 2023 Pratt & Whitney told the world that a rare condition in the powder metal used for some engine parts meant it would have to inspect most of a fleet, and it booked a US$2.9B operating-profit charge and a US$5.4B reduction in sales.⁵³ The shares fell from US$97.01 on 24 July 2023 to US$87.10 on 25 July 2023, when the issue became public.¹ In 2024 it paid US$384M and US$580M to settle two cases with the Justice Department and the SEC, and agreed to a US$200M penalty, half of it suspended, under a State Department agreement.⁶ A business with a strong edge has fewer such days, or has the edge anyway. RTX has had both. The edge held; the record did not.

What does protect it

• Certified, sole-source products. Pratt & Whitney is the only supplier of the F135 engine for the F-35, which flew more than a million engine hours in 2025 and drew a US$2.8B production contract.³ The Geared Turbofan powers more than 2,600 aircraft for over 90 operators.³ A new engine takes years and billions of dollars to certify.
• An installed base that needs servicing. Pratt & Whitney’s services sales were 44% of its sales in 2025, US$14.4B C (calculated).³ The Geared Turbofan aftermarket network grew to 21 facilities and shop-visit output on the PW1100 grew about 26% in 2025.³ The engine problem is, strangely, also a source of aftermarket revenue: more shop visits mean more work, though also more cost.
• Government programs with long lives. Patriot, Standard Missile, AMRAAM and Tomahawk are consumed in use and bought again. The US Congress passed US$25.4B for munitions in 2025 and the 2027 budget request sought a total of US$1.5 trillion for defense.⁶ Missiles that are used up are a better business than ships that last forty years.
• Scale in two cycles. Commercial flying and defense budgets do not usually peak together, though in 2026 they both did. I treat the diversification as real but smaller than it looks, since one customer group, governments, sets both the budget and the safety rules.

Where it is thin

Fixed-price contracts. Raytheon’s sales in 2025 were US$16.6B fixed-price and US$11.2B cost-type, 60% to 40%, and the share of fixed-price work has risen from 50% in 2023.³ When costs rise on a fixed-price contract the company, not the government, pays. In 2023 Raytheon’s segment margin was 9.0% on US$26.4B of sales, against 12.8% in 2021.⁵ The new multi-year munitions contracts are undefinitized, which is a polite way to say that the price is still being negotiated and the company is building before it knows what it will earn.⁶ Raytheon’s margin was 14.1% in the third quarter, with about US$85M of one-time benefits in it, so it is not a thin year now.⁴

Concentration is the other. Boeing and Airbus together are 16% of Collins and Airbus is 29% of Pratt & Whitney, and a single government is 37.6% of the company.³ Being hard to copy protects against competitors. It does not protect against a customer with a single vote.

IIIThe competitive advantage

The question here is what RTX can do that a rival cannot. There are three real competitors, one on each leg. GE Aerospace, with CFM’s LEAP engine, is Pratt & Whitney’s main rival for narrowbody engines. Lockheed Martin is the other great defense prime, and it is also a customer of RTX since the F-35 flies on Pratt & Whitney’s engine.

GE Aerospace is the comparison that hurts. In the third quarter its adjusted operating margin was 22.5% on US$13.3B of adjusted revenue, its commercial services revenue rose 23% and it says its services backlog is over US$170B.⁸ RTX’s whole company runs at an adjusted segment margin of about 15% in that quarter, and 11.8% across 2025, with a good deal of one-time help in 2026 C (calculated).⁴³ Pratt & Whitney’s adjusted margin was 10.5% in the third quarter and 9.2% for nine months. It is a business that has not earned what its rival’s engine business does, partly because the aftermarket mix has moved toward Geared Turbofan work, which the 10-K lists as a drag on margin, and partly because of the cost of the powder metal problem.⁴³

Lockheed Martin is the other side of the comparison. Its business segment margin was 10.8% in the third quarter, and its backlog is US$292.0B, which includes US$52.5B for PAC-3 MSE interceptors, the interceptor missile used with RTX’s Patriot system.⁹ RTX’s Raytheon segment, at 14.1% with about US$85M of one-time help, earned more than Lockheed’s segments did in the quarter. I read that as a good moment for missile makers, not a lasting gap.

So where is the edge? I can find three places, and none is certain.

Breadth. No competitor has all three of a leading commercial engine, a leading aircraft-systems business and a leading missile business. A diversified company is not worth more than the sum of its parts by itself; RTX has to prove that it is. Corporate cost is small, US$248M in 2025, but the group has also shown little ability to avoid problems in one division spilling into the others.³

The aftermarket on its own engines. More than 2,600 Geared Turbofan engines are in service, and the ones with the powder metal issue need shop visits that Pratt & Whitney and its network of 21 facilities do. If the engine becomes reliable, the services stream could be long. I cannot yet see the margin on that work. Reported profit at Pratt & Whitney was 7.9% in 2025.³

Missile capacity. The contracts signed in 2026 lock in volume for five to seven years at the three missile lines that matter most: Tomahawk, AMRAAM and Standard Missile.⁶ A competitor would need years to match the supplier base. But the customer can and does buy the same missile from more than one supplier on some lines, and the price has not yet been fixed.

Where rivals can catch up

GE Aerospace says its LEAP deliveries grew 26% in the year to date, and its CEO says commercial services will grow low double digits in 2027.⁸ If Pratt & Whitney’s Geared Turbofan fleet continues to lose time on the ground, airlines choose the LEAP for new narrowbody orders. In defense, Lockheed’s PAC-3 award and new entrants in drones and low-cost missiles will take part of what RTX counts on. The company itself lists several in its risk factors.³

IVManagement

Christopher Calio has been chief executive since 2 May 2024 and chairman since 30 April 2025. Before that he was president and chief operating officer from March 2023 and chief operating officer from March 2022.¹⁰ He succeeded Greg Hayes, who led the company through the merger and the engine charge. The chief financial officer is Neil Mitchill Jr. A change is coming at the place where the engine problem sits: on 15 September 2026 RTX said that Shane Eddy will step down as president of Pratt & Whitney on 1 January 2027, to be replaced by Jill Albertelli, who now runs its military engines business.¹⁷ The man who ran the segment through the powder metal years leaves just as the airlines expect the disruption to end, and the successor comes from the defense side. I do not read anything into the timing, but I note it.

How they are paid, and what that pays them to do

Mr. Calio’s total pay for 2025 was US$24.8M, against US$18.0M in 2024 and US$12.7M in 2023. In 2025 it was US$1.5M of salary, US$5.1M of bonus, US$10.9M of stock awards and US$6.8M of stock appreciation rights.¹⁰ About 94% of his pay is at risk, and 96% of votes at the 2025 meeting supported the pay plan.¹⁰ The size does not trouble me for a company this large. The measures do.

The annual bonus is set on two numbers, equally weighted: adjusted net income and free cash flow. The results in 2025 were US$9,115M and US$8,448M, against targets of US$8,340M and US$7,500M, so the bonus paid at 164% of target.¹⁰ But the company’s GAAP net income that year was US$6,732M and its reported free cash flow US$7,940M. The bonus numbers add back tax-law changes, tariffs, restructuring, acquisition accounting and “significant and non-recurring items”.¹⁰ That is 35% more profit than the accounts show, and 6% more cash, which is a difference worth knowing. It may be a fair adjustment. It also means the people who set the adjustments are the people paid on them.

The long-term awards are paid on four things over three years: growth in adjusted earnings a share (35%), return on invested capital (35%), and shareholder return against the S&P 500 (15%) and against defense and aerospace peers (15%).¹⁰ The 2023 to 2025 awards vested at 146%, helped by a shareholder return in the 82.7th percentile of the S&P 500, while return on invested capital was 7.6% against a 7.5% target.¹⁰ A return of 7.6% on capital is not a high bar, and the company earned it only after 2025’s results. Pay rewards earnings and returns; none of it measures the engine fleet’s time on the ground or the number of new compliance findings.

Ownership is firmer. Mr. Calio must own shares worth six times his salary. He reports 421,213 shares beneficially owned, of which 291,304 are the net shares from stock appreciation rights he could exercise within 60 days, worth about US$85M at $202.79, or about 56 times his 2025 salary C (calculated).¹⁰¹ All 19 directors and officers together report 1,406,743 shares, about US$285M.¹⁰ His own money is in, though most of it is options that pay only if the price rises. The shares are held by large index and fund managers, not a controlling owner: Vanguard 9.3%, BlackRock 7.8%, State Street 6.8% and Capital Research Global Investors 5.7%, as of 31 December 2025.¹⁰

What they did with the money

Over a full cycle, the evidence is mixed. From 2020 to 2025 RTX generated US$30.4B of free cash flow, paid US$18.8B in dividends and spent US$18.5B on buybacks, which is US$7.0B more than it earned C (calculated).^1113145153 The gap was filled by debt, which rose from US$31.8B at the end of 2020 to US$43.8B at the end of 2023 and has since fallen to US$37.4B.¹¹⁵⁶

The big decision was the buyback of 2023. In October 2023, three months after the engine issue became public, the board authorized US$11B of buybacks and RTX borrowed US$10B to buy back stock through accelerated repurchases, taking US$12.9B in the year at an average of about US$91 a share.⁵ S&P cut the credit rating from A- to BBB+ in August 2023, and both agencies moved their outlooks to negative in October.⁵ In hindsight the shares were cheap, because they have since more than doubled. But the choice to lever up while the size of the engine problem was still unknown was a decision to rely on the best case. It paid off, and I would not give the same credit to a decision that had gone the other way. Since then, buybacks were US$444M in 2024, US$50M in 2025 and nothing in the first nine months of 2026, and the company has used cash to repay about US$6B of debt.¹⁵³⁶

Then there is the compliance record. In October 2024 Raytheon Company entered two deferred prosecution agreements with the Justice Department: one for payments made through a joint venture, Thales-Raytheon Systems, in the Middle East, and one for defective pricing on contracts from 2011 to 2013 and 2017.⁶ RTX paid US$384M and US$580M, and in August 2024 agreed with the State Department to a US$200M civil penalty, half of it suspended, for export-control violations, many of them found when integrating Rockwell Collins.⁶ An independent monitor was engaged in April 2026, and on 21 September 2026 RTX and Raytheon Company signed an administrative agreement with the Navy that relates to all of these matters.⁶ The company says none will have a material effect.⁶ I take the point that these happened in the past and were settled. I also note that the monitor is still in the building.

VFinancials

Four things in RTX’s statements do not mean what they first seem. The company in 2016 was a different company: United Technologies, with Otis and Carrier, no Raytheon and no Rockwell Collins. 2020 is half pandemic and half merger. Reported earnings carry about US$1.15 a share of acquisition accounting that is not cash. And the free cash flow line is flattered by things the company excludes from the owner’s point of view, which I take up below.

With those cautions, the picture is of a business that grew at 4.7% a year across the decade, from US$58.7B on a combined basis in 2016 to US$88.6B in 2025, and that has been through a large reset in the middle C (calculated).^181131920 The pro forma sales for 2019 were US$74.2B before the pandemic took them to US$64.1B in 2020; 2025 was 19% above the 2019 level.¹¹³

Profit tells a harder story. The combined segment margin was 14.4% to 14.8% before the merger, fell to 9.5% in 2021 and 6.7% in 2023, and was 11.8% in 2025 C (calculated).¹⁸¹¹⁵³ Part of the drop is the measure: the early years use Raytheon Company’s operating income, which includes the pension cost the government reimburses, and the later years are segment profit before that item, so the table overstates the fall by a point or so. The 2023 figure also carries the US$2.9B engine charge, about 4 points of margin on its own. The rest is real: the pandemic, a wave of fixed-price cost overruns at Raytheon, supply-chain inflation and the engine issue. After 2023, the margin has recovered, helped by higher aftermarket volume, and in the third quarter of 2026 the adjusted segment margin was 15.2% with US$0.4B of tariff refunds in it.⁴ It is back near where it began, and the price assumes it goes on rising.

By segment, Collins grew from US$21.2B in 2021 to US$30.2B in 2025, and its margin from 11.3% to 16.3%.⁵³ Pratt & Whitney’s sales rose from US$18.2B to US$32.9B, but its segment profit was a loss of US$1.5B in 2023 and a profit of US$2.6B in 2025, a 7.9% margin.⁵³ Raytheon’s sales rose from US$26.6B to US$28.0B while its margin fell from 12.8% to 9.0% and came back to 11.5%.⁵³

Cash flow tells a clearer story than earnings. Free cash flow rose from US$2.5B in 2020 to US$7.9B in 2025, with a dip to US$4.5B in 2024, which carried about US$1.1B of legal settlements and about US$1.0B of engine compensation, and in 2025 it covered the dividend 2.2 times C (calculated).¹¹¹⁵³

Reported profit against adjusted profit

For 2025, GAAP earnings were US$4.96 a share and adjusted earnings US$6.29. The US$1.33 difference is US$1.15 of acquisition accounting adjustments (mostly the amortization of intangible assets created by the Raytheon and Rockwell Collins deals), US$0.14 of restructuring and US$0.04 of other items.¹⁶ I do not mind the first: it is a non-cash cost of past deals. But it is a real cost of having paid for those deals, and I prefer to look at cash.

What an owner can really take out

Free cash flow was US$7.94B in 2025, the best year in the series, and the 2026 outlook raises it to US$9.0B to US$9.2B.³⁴ That is the number the company talks about. It is not what an owner can take. Three things are inside it that I do not count as free. The first is stock compensation, US$519M, and the 401(k) match, US$573M, which RTX now pays in shares, so that its cash is higher and its share count is also higher.³ The second is “increase in other intangible assets”, US$492M in 2025, which is money paid to airframers for the right to supply new aircraft and to partners on engine programs; the business needs it to stay on its platforms.³ The third is one-off: about US$1.0B of the 2025 free cash flow was paid out or credited to airlines for the engine issue, and that is a cost of 2025 and 2026 rather than a permanent drain, so I leave it in 2025 and watch it fall to about US$0.7B in 2026.³⁶

After those, owner cash after interest was US$6.36B in 2025, which is 94% of net income of US$6.73B, and 70% of the US$9.1B of adjusted net income used for the bonus C (calculated).³¹⁰ Over the five years 2021 to 2025 the total was US$21.8B against free cash flow of US$27.8B and net income of US$23.8B C (calculated). Reported earnings are close to the cash over the cycle. Adjusted earnings are not. In 2026 the guided free cash flow of about US$9.1B becomes US$7.3B of owner cash once I remove the tariff refunds, the stock and 401(k) shares and the intangible payments, and that is 2.6% of the market value. The gap between those two numbers, about US$1.8B, is the point of this section.

Capital spending is not the problem: US$2.6B in 2025 against depreciation and amortization of US$4.4B, of which about US$2.0B is acquisition accounting that is not a cash need.³ Working capital is. In the first nine months of 2026 contract assets rose US$2.2B and inventory US$1.7B, because sales are rising 11%, while contract liabilities, the advances customers pay, fell US$0.7B. That is what growth looks like and it is a use of cash, not a one-off.⁶ Company-wide, factoring of receivables added about US$0.3B to nine-month operating cash flow compared with a year earlier.⁶

The balance sheet

Total debt was US$31.8B at the end of 2020, after the merger and the Otis and Carrier separation, and was almost unchanged through 2022.¹¹¹⁴ It jumped to US$43.8B in 2023 on the US$10B accelerated buyback and has fallen since, to US$37.4B at 30 September 2026, with cash of US$9.2B and net debt of US$28.2B.⁵⁶ On my estimate of 2026 EBITDA of about US$16B, that is about 1.8 times E (estimate). Debt is mostly fixed-rate with an average maturity of about 12 years, and the revolving credit line of US$5.0B, renewed in August 2026 to 2031, is undrawn.³⁶ S&P rates it BBB+ and Moody’s Baa1, both with a positive outlook.⁶

Three items sit outside that debt. Minority interests of US$1.7B and the net pension and retirement liability of US$1.9B I add to the claims ahead of the shares.⁶ The third is a commitment for commercial aircraft financing and other contracts of about US$12B, which is not a liability today but is a promise to customers that I cannot value.⁶

Dividends and buybacks

The dividend has grown every year since 2021, from US$2.005 a share to US$2.67 in 2025 and about US$2.87 in 2026, by my arithmetic from the declared amounts, with US$0.73 a quarter now C (calculated).³²¹ In 2025 it cost US$3.57B, 45% of free cash flow and 56% of owner cash C (calculated).³ At the price it yields 1.44%. It is well covered. Buybacks, by contrast, are an option the company can use or leave. There is US$0.6B left in the 2023 program and the company has not bought in 2026.⁶

The company says that an executive order issued after the new defense acquisition rules may limit dividends, buybacks and executive pay incentives at defense contractors that perform badly, and that such a limit could hurt the share price.⁶ This is the first time I have seen the largest customer write into the rules a link between its supplier’s delivery speed and its owners’ cash. It matters for how I think about the dividend, since 37.6% of sales and a much larger share of the profit growth now depend on one buyer that can set conditions.

The price in numbers

On those numbers the shares trade at 26.3 times 2026 adjusted earnings, 19 times my estimate of 2026 EBITDA, and the free cash flow yield is 3.3% of market value, or 2.6% on owner cash C (calculated). The shares have risen from US$49.93 on the day of the merger, during the pandemic sell-off, to $202.79, which is 24% a year over six and a half years before dividends, and from US$84.14 at the end of 2023 to US$183.40 at the end of 2025.¹ This year the shares are up 11%, while the midpoint of the earnings outlook has risen from US$6.70 in January to US$7.70, 15%, so this year’s gain is earnings and not a higher multiple C (calculated).¹⁴¹⁶ The last close is the base from which I judge everything below.

VIValuation

First, how this goes badly wrong

Before any numbers, I ask what would hurt an owner most. I count three ways.

The engine problem was not the last one. The powder metal issue cost US$2.9B of profit and US$5.4B of sales when it was booked in 2023, and the accrual for customer compensation is down to US$0.2B.⁵⁶ That does not mean the cost is over. The company expects aircraft on the ground to stay high through 2026, and expects about US$0.7B of cash cost this year after about US$1.0B in each of 2024 and 2025.³⁶ It also says other engine models contain parts made with the affected powder, with no significant impact expected at this time, and that the impact rests on assumptions about shop visits, parts and negotiations with customers.⁶ The securities class action was dismissed in 2025 and is on appeal, the shareholder derivative suits are open, and the SEC is investigating the 2023 disclosures.⁶ An unknown amount sits in the maintenance contracts. I cannot size it, so a scenario below takes US$4.5B of extra cash out of my base over 2027 to 2029.

The cycle and the budget turn together. In 2020 combined sales fell 14%, from US$74.2B to US$64.1B, and free cash flow fell to US$2.5B.¹¹ Defense helped then, since governments kept paying. Now 52% of sales are governments, and the same buyer sets the rules on distributions and on fixed-price terms. A downturn in air travel at the same time as a budget fight is not likely, but it is the case that hurts.

I pay a full price for a good year. The shares trade at 26.3 times 2026 adjusted earnings, and 2026 earnings carry a tariff refund that will not come back.⁴ A business at a high multiple is exposed to two things at once: earnings falling and the multiple falling. That is not about the business. It is about what an owner pays.

What an owner can really take out

In 2025 owner cash after interest was US$6.36B, or US$4.66 a share on 1,365M shares, 2.3% of the price C (calculated).³ In my base case, owner cash before interest is US$8.7B in 2026, US$11.8B in 2028 and US$14.3B in 2030. Take off about US$1.4B a year of interest after tax, and an owner has about US$12.9B in 2030, or US$9.4 a share E (estimate). Reported free cash flow for 2026 will be higher, about US$9.1B, and adjusted net income higher still. The distance between the numbers the company talks about and the one I use is the point of this section.

What I can’t judge

I can’t say what the engine problem will cost. The filing gives me a cash figure for this year and an accrual, not a limit. I can’t say what the new munitions contracts will earn: they are undefinitized, ceilings are not orders, and Congress funds them a year at a time.⁶ I can’t say what Pratt & Whitney will earn on the aftermarket, because the company does not break out the profit on Geared Turbofan work. I can’t say how long the 2026 level of government attention to munitions lasts: the company’s own filing says the future depends on awards, timelines and budget decisions.⁶ I can’t say what the Court of International Trade or the new tariffs will do to the US$0.5B of tariffs RTX has paid, or how much of the refund is still to come.⁶ For those reasons I show a range, and I give the odds in the last section as guesses.

How the cases are built

I build each case year by year in US dollars. Sales for 2026 start at US$97.25B, the midpoint of the outlook, and grow by case. Adjusted operating profit is a margin on sales; the margin includes the FAS/CAS pension adjustment, which was US$0.75B in 2025 and has been shrinking from US$1.13B in 2023, so a 16.0% margin in my base is a little lower once that fades, and needs the segments to earn 15.5% or so.³⁵ Tax is 19%, close to the 19.1% of 2025. Owner cash is after-tax operating profit times a conversion: 78.6% in 2026, calibrated to the bridge above, which carries about US$0.7B of engine cash and a working-capital build, rising to 90% in the base once the engine cash stops. It is a simplification, and it is the one I can defend with filed numbers. I discount at 9.0%. That is the ten-year Treasury yield of 4.91% on 2 November plus 4.1 points for an equity risk I cannot hedge: the engine fleet, a single dominant customer in the government, and a share of profit set by fixed-price terms.⁷ The credit is BBB+ with a positive outlook, so I could argue for less, and the table below shows eight and a half percent.⁶ Cash for the last three months of 2026 counts, 30% of the year’s; net claims are as of 30 September 2026. After 2036 the last year’s cash grows at 2.0%, 3.0% or 3.5% in the bear, base and bull. The odds in this section are my guesses; the weighted value of $141 uses 25%, 50% and 25%.

Three comparisons help. My base owner cash is US$8.7B in 2026 and grows 7.8% a year to US$18.4B in 2036, faster than sales because the margin rises two points, from about 14% to 16%. That is a lot to ask, and it is below what the third quarter looked like. My bull assumes adjusted operating margins of 17.5%, near where GE Aerospace is now on a much richer business, and sales growth of 8% a year to 2030.⁸ And my bear takes sales down 3.5% in 2028 and margins to 11.5%, a repeat of the shape of 2020 and 2023, not worse.

Put in plain words, to be worth $202.79 the business needs to earn an adjusted operating margin of about 21% by 2030, from about 14% this year and 12.8% last year; or to grow sales 3.6 points a year faster than my base for ten years, to US$220B in 2036, 2.5 times 2025; or to be a business that deserves a discount rate of 7.6%, which is close to what an investment-grade, low-cycle company earns. None of these is silly in isolation. Together they say the market pays for a company that is better than GE Aerospace in the commercial engines and better than Lockheed in missiles, and has no more engine surprises E (estimate).

The margin of safety, and the price against the range

My margin-of-safety price is $104. It is 70% of my base value. At that price an owner would have a 30% cushion against being wrong about the engines, the margin and the budget, and the price would be 58% above my bear. It would be about 14 times 2026 adjusted earnings. The shares last closed below US$105 on 22 July 2024, at US$103.77.¹ At $202.79 there is no margin of safety. The price is 36% above my base, 1% above my bull, and 3.1 times my bear. My weighted value, with 25% on the bear, 50% on the base and 25% on the bull, is $141.

The price reflects a real debate, and the case for the market being right needs naming. It says that RTX is the one large company with a leading position in commercial engines, aircraft systems and missiles, that backlog is US$340B and growing, that the Geared Turbofan will become a long annuity once the fleet is repaired, that governments in the US and Europe will spend more on munitions for a decade, that free cash flow is rising toward US$9B and more, and that all of this deserves 26 times earnings. If I give that all of its weight, my bull case is where the price lands. If the credit improves enough for me to use 8.0%, my base value rises to $184, which is still below the price. The case against says that margins are at a peak with a tariff refund in them, that fixed-price missile contracts are being priced as they are built, that one customer is writing conditions into the rules, and that the engine fleet is not finished. I give the second side more of the weight, and I would not be surprised if the first side were right for a while.

The next best use of the same dollar

Price only means something next to what else the money could do. The ten-year Treasury yields 4.91% with no risk of default.⁷ The closest comparison I have worked through on this site is GE Vernova, another large US industrial that builds heavy machines and then services them for decades, whose base return to December 2036 is 5.8% a year in my work. The ASML page, a sold-out machine maker with a large service fleet, gives 5.2% to September 2036. RTX’s base return is 5.4% a year to October 2036 E (estimate). That beats the bond by about 0.5 points, sits between those two pages and falls short of the 9.0% I ask by 3.6. It differs from the 7.6% above because that is the rate at which my base cash flows equal the price, while this return assumes the value I put on the business in October 2036. RTX is a very good business. The price is a separate question. A good business at a price that needs more than my bull is a poor purchase for the same reason a poor business at a low price can be a good one: what matters is what you pay for what you get, and this looks like a good business at a full price.

The bull case needs many things at once, so I multiply them out. Sales growth of 8% a year to 2030 (about 30% likely, by my guess), an adjusted margin of 17.5% (30%), conversion of 92% of after-tax operating profit to owner cash (50%), no recession or budget fight in ten years (30%), and no second engine problem (60%). Together that is about 0.8% if they are independent. They are not, since a strong market helps four of the five, so I call it 4% to 6%, not 25% E (estimate). My 25% weight on the bull is generous on that arithmetic.

Where this could be in 5 and 10 years

For the base to be right at five years, sales need to be about US$130B in 2031, the adjusted margin about 16%, and owner cash about US$15.2B, and the value is then $181 a share, which is below today’s price because five years of cash have gone out and the rest has grown. At ten years the base reaches $215. In the bear, owner cash stays near US$7B to US$10B and the value is $87; in the bull it is US$23B in 2036 and the value $306. The four inputs that drive it are sales growth, margin, conversion and the 9.0% rate. These are estimates of value, not forecasts of the share price. The market price can sit above or below value for years, and since the merger the shares ranged from US$49.93 to US$225.49.¹

VIIThe case for the price

What most people miss

Three things cut in RTX’s favour and get little attention. First, the engine accrual is almost gone. The company booked US$2.8B of accrued liabilities for customer compensation in 2023, and at 30 September 2026 US$0.2B remained, with cash paid or credited of about US$1.0B in each of 2024 and 2025 and about US$0.7B expected in 2026.⁵³⁶ Second, customers fund a large part of the business. Contract liabilities, mostly advances and progress payments, were US$21.6B, about 8% of the market value.⁶ Third, the two halves of the company do not peak together, and both credit agencies moved their outlooks to positive in 2026, Moody’s in February and S&P in May.⁶

The bull case

In plain words: RTX holds leading positions in engines, aircraft systems and missiles at a time when, the argument runs, each market is short of supply. Airlines that cannot get new aircraft keep old ones flying and pay for parts and repair, and governments are asked to replenish missile stocks; the President’s June request included about US$21B for that.⁶ The company says it received more than US$81B of new awards in the third quarter, raised its 2026 outlook again and ended the quarter with a record backlog; chief executive Chris Calio called it a “record backlog of $340 billion.”⁴ The 2025 tax law added US$156.2B to the defense budget through 2029, including US$24.4B for the Golden Dome missile defense project, and both rating agencies now have positive outlooks.⁶

I found no named analyst or fund manager whose published view on RTX I could verify in the time I had, so the bull case here is management’s. The holders I can document are the big index and fund managers in the 2026 proxy: Vanguard 9.3%, BlackRock 7.8%, State Street 6.8% and Capital Research Global Investors 5.7% at the end of 2025. Those are holdings, not statements of view.¹⁰

Against reality

If RTX hits its outlook, converts the new missile contracts at today’s margins and the engine cost stops, the model gives a value near my base, $149 a share. The price of $202.79 is 36% above that. It is possible, and it needs more than the plan. To be worth $202.79, the business would have to earn an adjusted operating margin of about 21% by 2030 on my sales path, against about 14% this year and 12.8% last year, or grow sales about 3.6 points a year faster than my base for ten years. Even my bull, at a 17.5% margin and sales of US$133B in 2030, comes to $201.

The distance is about the multiple. Today’s earnings are paid for at 26.3 times. At about 19 times, which is what my base value implies on 2026 earnings, the same business is worth $149. I multiplied out the odds for my bull in Valuation: about 1% if the five conditions are independent, perhaps 4% to 6% with the overlap E (estimate). The price asks for more than the bull.

What a 10x would take. Ten times $202.79 is US$2,028, a market value of about US$2.8 trillion. At 25 times earnings that needs about US$111B of net income, 16 times the US$6.7B of 2025 C (calculated). At a net margin of 12%, well above the 7.6% of 2025, that is sales of about US$923B, 10 times today’s; the whole base defense budget the President proposed for 2027 is US$1.15 trillion.⁶ In ten years it would take 26% a year. The shares rose about 24% a year from US$49.93 on the day of the merger, but that began in the pandemic sell-off and rode a re-rating. I know of very few companies that have grown tenfold from a value like this in a decade. For RTX I think it is close to impossible, and I do not model it.

Where I come down

What moved me was the order book, the fall in engine cash and free cash flow rising from US$7.9B to about US$9.1B. What did not move me was the price: I cannot build a case where owner cash justifies $202.79 without a margin no company in this group has earned, and 2026 has a tariff refund in it. What would raise my values is a Pratt & Whitney margin above 12% with engine cash at zero, Raytheon earning 14% on the new fixed-price work, and a clear answer on the distribution rules. Each extra point of steady margin from 2031 is worth about $9 a share.

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest, and a test of each.

• Tariffs. RTX is the importer of record for some products and says it has paid about US$0.5B of tariffs under the emergency-powers law since they began, and has recognized about US$0.4B of approved refunds, of which about US$0.3B has been received.⁶ The government has appealed the refund order and has put new tariffs in place under other laws. The company says it does not expect a material effect, and I take that with the knowledge that it benefited from the refund in the third quarter. My test removes 0.5 point of margin for good.
• Compliance and the monitor. The settlements below total about US$1.2B including the suspended penalty, and the deferred prosecution agreements can be revived if RTX fails to comply for three years from the monitor’s engagement.⁶ A further finding would cost cash and, more, would put at risk the right to bid for US government work, which the 21 September administrative agreement exists to protect.⁶ I cannot estimate the odds.
• Customer concentration. One government is 37.6% of sales, 52% with foreign governments, and Airbus is 29% of Pratt & Whitney.³ A budget fight, a rule on distributions or a change in the Geared Turbofan relationship with Airbus are not losses of customers, but each of them is a loss of leverage.
• Supply chain. The 10-Q lists constraints in raw materials, rare earth elements and microelectronics, and in 2023 China announced sanctions against what was then Raytheon Missiles & Defense.⁶⁵ Higher volumes in 2026 depend on suppliers that RTX is trying to bind with long-term agreements.⁶
• Fixed-price work. Sixty percent of Raytheon’s sales are fixed-price, and the new munitions awards are priced after work has started.³⁶ The company’s own history, 9.0% margin in 2023, shows what happens when costs run ahead.⁵
• Credit. BBB+ with a positive outlook, US$37.4B of debt, US$2.9B of it due in 2027 and US$3.5B in 2028, against cash of US$9.2B.⁶³ Not a worry at today’s cash flows. It is one if the cycle turns while a settlement or a charge falls due.

What a bad year looks like in the model

The surprise is how little the engine problem costs in cash. Even US$4.5B of extra cash over three years takes only $3 off the base, because the value sits in the long run, and US$4.5B is 1.6% of the market value. What matters more is anything that lasts: lost share at Pratt & Whitney takes $16 off, and a 2020-style downturn $40. I read this as saying the one-off liability is the smaller risk and the franchise is the larger. The risk is not a bad year. It is a bad decade.