General Dynamics

IThe business

General Dynamics makes four different things and sells most of them to one customer. In 2025 revenue was US$52.55B, up 10%, operating earnings US$5.36B, and diluted earnings per share US$15.45.³ About 117,000 people work there.³ The head office is in Reston, Virginia. 68% of revenue came from the US government, 15% from US commercial customers, 8% from foreign governments and 9% from foreign commercial buyers.³ The US government customer is mostly the Department of War, the name the filings now use for the Department of Defense.³

The mix has moved. Marine Systems is the largest segment at 32% of revenue and holds two thirds of the backlog, and Aerospace has gone from 20% of revenue in 2023 to 25% in 2025.³⁴ I cover them in the order of the questions they raise, not by size.

Marine Systems: submarines, and the limits of the people who build them

Marine Systems had US$16.72B of revenue in 2025 at a 7.0% operating margin, and nuclear-powered submarines were US$12.6B of it, up from US$8.6B in 2023.³ Its three yards are Electric Boat in Connecticut and Rhode Island, Bath Iron Works in Maine and NASSCO in San Diego.³

Columbia and Virginia. Electric Boat is the prime contractor and lead yard on every US Navy nuclear submarine programme. The Columbia class is a programme of 12 ballistic-missile submarines, rated at the highest national priority, with the first boat expected in 2028 and a government programme value above US$125B.³ The Virginia class is bought in blocks. General Dynamics and one other yard are building Blocks IV and V together, and the 10-K lists 14 Virginia-class boats in backlog for delivery through 2034, ten of them with the Virginia Payload Module, an 84-foot hull section that adds four launch tubes.³ The company says it expects the workforce to keep growing so that it can build one Columbia and up to two Virginia boats a year as the supplier base grows to match.³

The third quarter of 2026 was a very large one. Orders in the three defense segments were US$93.8B, 8.5 times their revenue, driven by Virginia-class and Columbia-class awards.⁴ Marine Systems backlog went from US$65.2B at 5 July to US$146.9B at 4 October, and the unfunded part, which means work the Navy expects to pay for but Congress has not yet appropriated, went from US$22.8B to US$99.6B.¹²⁴ The documents I opened give the dollar totals only, so I cannot tell you the number of boats or which block in this award. I looked for them and could not verify them.

Destroyers and the rest. Bath Iron Works is one of two builders of the Arleigh Burke (DDG-51) destroyer. A DDG-148 award in 2025 took its backlog to 11 ships with deliveries through 2032.³ NASSCO builds fleet oilers and an expeditionary sea base; seven T-AO-205 oilers are in backlog into 2030, with the eleventh and twelfth awarded in 2025.³ Those programmes are steady, not exciting.

The supply chain is the constraint. The company says demand for Navy shipbuilding, particularly submarines, has put pressure on a shipbuilding industrial base that was already short of workers after the pandemic, and that it is working with the Navy to stabilise and grow the supply chain.⁵ Marine Systems’ 2024 margin of 6.5% was hurt by supplier quality problems and late deliveries that caused cost growth and delay on the Virginia Block IV contract; the margin recovered to 7.0% in 2025 and 7.5% in the first nine months of 2026, helped in the third quarter by a favourable adjustment on that same programme as it nears completion.³⁵ So the demand is not in doubt. What I cannot judge is how fast the whole chain of forgings, valves, welders and engineers can grow, because that is outside the company’s control, and because it also decides how fast a fixed-price backlog turns into revenue and cash.

Margins are set by the contract. In the defense segments, the company says, new long-term production contracts usually start with lower margins and improve as the learning curve is climbed.⁵ A very large new submarine award therefore starts a long period of low-margin work. Marine Systems earned 8.3% to 9.0% of revenue in 2017 to 2021 and 7.0% in 2025 C (calculated).⁷¹⁵³ I model 7.6% for the base case. I do not model a return to 9%.

Aerospace: Gulfstream and the services that follow it

Aerospace had US$13.11B of revenue in 2025 at a 13.3% margin. Aircraft manufacturing was US$9.41B, up from US$5.71B in 2023, and aircraft services US$3.70B.³ Gulfstream delivered 158 aircraft in 2025, 136 in 2024 and 111 in 2023.³¹⁷ In the first nine months of 2026 it delivered 119, 101 large-cabin and 18 mid-cabin, against 113 a year before.⁴

The new family. The G500 and G600 entered service in 2018 and 2019, the G700 in 2024 and the G800 in 2025 after FAA certification in April 2025.³ The G650, the model the G800 replaces, made its final deliveries in 2025.³ Two more are coming: the large-cabin G400, in development, and the super-midsize G300, announced in the second half of 2025, which replaces the G280.³ The company describes this as completing a nearly two-decade effort to rebuild the whole line.³ The filing also says the mid-cabin aircraft are assembled by a non-US partner, and that the Israel-based supplier of mid-cabin airframes has caused delays because of the conflict in the Middle East.³⁵

Where it is built. Large-cabin aircraft are designed, developed and built in Savannah, Georgia, where Gulfstream has added manufacturing buildings, service capacity and a parts distribution centre.³ The properties list also shows a site in Mexicali, Mexico, under Aerospace and also under Marine Systems.³ The filings give no output figure or ramp rate for Mexicali, so I cannot say how much of the delivery growth comes from it. Treat it as an unknown, not a driver.

Orders and backlog. Aerospace orders were US$14.1B in the first nine months of 2026, 1.4 times revenue, against US$10.4B and 1.1 times a year earlier.⁴ Aerospace backlog was US$25.4B at 4 October 2026, up from US$21.8B at the end of 2025.⁴³ 68% of Gulfstream’s 2025 orders and 61% of its backlog were from US customers.³ Services, which are maintenance, parts, and the fixed-base operator (FBO) airport terminals, were US$3.70B in 2025; the company says the 3,000-plus Gulfstreams in service, a 30-facility FBO network and Jet Aviation’s fleet-management arm feed it.³ Two headwinds: Russia sanctions have cut off part of the market, and tariffs reduced Aerospace operating margin by 30 basis points in 2025, although the company says current tariffs are not material to its results, and a February 2026 Supreme Court ruling led to refunds of some tariffs paid.³⁵

Combat Systems: shells, propellant and vehicles

Combat Systems had US$9.25B of revenue in 2025 at a 14.4% margin: military vehicles US$4.97B, weapon systems and munitions US$3.10B, and engineering services US$1.17B.³ Vehicles means the Abrams tank, the Stryker wheeled vehicle and a family of wheeled and tracked vehicles for Europe, Canada and others.³ Munitions means large- and medium-calibre ammunition, Gatling guns, artillery projectiles, propelling charges, and propellant, where the company says it is the world’s largest maker of Ball Powder.³

The mix is changing. Revenue from US military vehicles fell US$203M in 2025 and again by US$203M in the first nine months of 2026, as the Army re-plans its fleet and the M10 Booker programme was cancelled, while international vehicles rose US$194M in 2025 and US$108M in nine months of 2026 on wheeled and tracked programmes in Europe.³⁵ Weapon systems and munitions rose US$343M in nine months on more artillery production.⁵ In 2025 the company won US$9.2B of vehicle awards from foreign customers, and backlog at Combat Systems was US$28.7B in October 2026, against US$17.0B at the end of 2024.³⁴ Management expects munitions demand to stay elevated and US vehicle volumes to be down slightly in the near term.⁵ It is adding capacity in artillery projectile and charge lines, extruded propellant, Ball Powder and rocket motors.³ The filings do not give a shell output rate, and I will not guess one.

Technologies: GDIT and Mission Systems

Technologies had US$13.47B of revenue in 2025 at a 9.5% margin.³ It has two businesses and about 40,000 people. GDIT is a federal IT services firm: cloud, cyber, networks and operations for the military and civilian agencies, with CENTCOM, Medicare’s payment systems and the National Geospatial-Intelligence Agency among its customers.³ Mission Systems makes defense electronics and communications, including NSA-certified encryption, submarine fire-control and combat systems, ground segments for satellites and undersea drones.³ IT services were US$9.06B of the 2025 total.³ This segment is the least special of the four. Its work is mostly priced on cost, and 85% of its 2025 orders came from options and extra work on existing contracts rather than new competitions.³ The federal staff cuts and contract changes in 2025, and the government shutdown at the start of the fiscal year, touched IT services, and the company expects some limited effect to continue.³

One customer

US government revenue was US$35.76B in 2025, 68% of the group.³ Fixed-price contracts were 51% of that, cost-reimbursement contracts 44% and time-and-materials 5%.³ That matters in a particular way. When the contract is fixed price, cost overruns fall on the company, as they did on Virginia Block IV in 2024. When it is cost-plus, there is no overrun, but the margin is capped. Marine Systems sells almost entirely to the Navy, so the group’s largest single exposure is one branch of one government.¹⁵ A continuing resolution, signed on 2 September 2026, funds the government to 11 December; under it, programmes of record get last year’s money and no new starts are allowed. The company says it does not expect this one to have a material effect.⁵

The state of play now

The third quarter of 2026 was a very good one by every number the company publishes. Revenue was US$14.57B, up 12.9%, operating earnings US$1.53B, up 14.7%, margin 10.5%, diluted earnings per share US$4.49, up 15.7%, and cash from operations US$1.71B, 139% of net earnings.⁴ Every segment grew; Marine Systems revenue was up 21% and its margin 7.9%, Aerospace revenue was up 9% at a 14.5% margin, Combat Systems margin slipped to 13.8% and Technologies to 9.4%.⁴ For the year the company now expects revenue of US$56.5B to US$56.8B, an operating margin around 10.5% and diluted earnings per share of US$17.20 to US$17.30.⁵ The range implies fourth-quarter revenue of US$14.4B to US$14.7B, against US$14.38B in the fourth quarter of 2025 and growth of 12.9% in the third; I read that as cautious guidance and do not rely on it C (calculated).⁵³ On 7 October the board named Danny Deep, the president, as chief executive from 1 January 2027, with Phebe Novakovic moving to executive chairman.¹¹ The price on 2 November is $373.46.¹

IIThe moat

The test of a durable business is what it earned when conditions were bad, and what it earns on the money it keeps putting in. General Dynamics has two answers, one for each half of the company.

The defense half is protected in the plain sense that the customer has very few places to go. The Navy has its nuclear submarines built by Electric Boat and one industry partner. A new builder would need nuclear-qualified yards, a cleared workforce and a supplier network that takes years to qualify, and the Navy is working with the companies on the supply chain rather than on a new builder.³⁵ Columbia has the highest national priority rating.³ The same contracts give the company visibility. Backlog is now 3.9 times the 2026 revenue outlook C (calculated).⁴⁵ In 2020, a year when the jet market stalled and many industrial companies cut guidance, General Dynamics’ revenue fell 3.6% and operating earnings 9.6%, and free cash flow rose to US$2.9B C (calculated).¹⁵¹⁶

The weakness is the margin. A business protected from competition ought to keep its price. General Dynamics’ operating margin was 12.3% in 2016, 13.7% in 2017, 10.8% in 2021 and 10.2% in 2025 C (calculated).⁷¹⁵³ Part of that is mix: CSRA, bought in 2018, added IT services to a segment that has earned 9.3% to 10.2% since, and Aerospace has been carrying the cost and the lower early margins of an all-new line of aircraft.⁷¹⁵³ Part is the customer, which pays either a fixed price it negotiates or cost plus a fee it sets. I read the record as saying the protection covers the volume and the position, but the owner shares most of the gain with the customer.

• Position that cannot be copied quickly. Electric Boat is the prime contractor on all US nuclear submarine programmes, and Bath Iron Works is one of two DDG-51 builders.³
• Visible work. Backlog of US$220.3B, with US$111.8B funded.⁴ The funded part is 2.0 times the 2026 revenue outlook, about two years of revenue that does not depend on a new appropriation C (calculated).⁴⁵
• A product brand. Gulfstream aircraft have set more than 350 city-pair speed records, and the company supports over 3,000 of them in service, which makes the services line grow with the fleet.³ Aircraft manufacturing revenue per delivery rose from US$51.4M in 2023 to US$59.6M in 2025, with completion work included C (calculated).¹⁷³
• Money in advance. Customer advances and deposits were US$11.08B at 4 October 2026, up from US$9.82B at the end of 2025.⁴ Customers fund part of the inventory. That helps while orders grow and reverses when they do not.

Where it is thin

The first thin place is that profit is the contract, and contracts get re-priced. When Virginia Block IV ran into supplier problems, the company took the cost. About half of its US government revenue is fixed-price.³ The second is that growth in the biggest segment comes from awards that lock in margins of 7% to 8% for a decade. The third is the jet market. Business-jet deliveries are discretionary, and the company noted the Russia sanctions and tariffs as constraints.³ The fourth is people. The company says that demand in Europe and in shipbuilding is putting pressure on the supply chain and on hiring skilled workers.³⁵

Return on invested capital is the cleanest single check. The company calculates it at 12.3% in 2023, 13.2% in 2024 and 14.2% in 2025.³ That is a good, not exceptional, return on a base of US$33B of average invested capital.³ It is rising with the margin on Gulfstream, and it is the number the executives are paid on.¹⁰

IIIThe competitive advantage

The question here is what General Dynamics can do that a rival cannot. The answer is different in each segment, and in two of them it is more about being the only one than about being better.

Submarines: the edge is the position

In Marine Systems the company is the prime contractor on Columbia and Virginia, with one industry partner on the Virginia blocks. That is an advantage in volume and in the customer’s need for it, and the proof is the US$93.8B of defense orders in a single quarter.⁴ It does not show up as an unusually high margin, and I do not expect it to. The customer sets a fee, the company has been working to get the learning curve down, and Marine Systems earned 7.0% to 7.9% in the periods I have reported.³⁴ The honest way to read this advantage is as a very long annuity of low-margin volume, with a ceiling on its upside and a floor under its downside, unless costs run away on a fixed-price contract.

Gulfstream: a real margin gap

Where the edge shows up in the numbers is aircraft. Gulfstream earned a 14.5% operating margin in the third quarter of 2026 and 14.7% over nine months.⁴⁵ Textron Aviation, the nearest US rival in the broader business-jet market, earned a segment margin of 10.7% on revenue of US$1.68B, and Bombardier an adjusted EBIT margin of 10.5% in its second quarter.⁸⁹ The three are not exactly comparable, and Gulfstream’s figure includes Jet Aviation’s services, but the gap is about four points on revenue of US$3.5B a quarter, or around US$0.5B a year of operating earnings C (calculated).⁴⁸⁹

Gulfstream delivered 119 aircraft in nine months, with 101 large-cabin ones, and holds US$25.4B of backlog, about 1.8 years of its revenue C (calculated).⁴ The larger the aircraft, the less direct competition there is. At the top of the market the nearest rivals are fewer than at the middle. I do not have comparable data for Dassault or Embraer, so I leave them out.

Combat and Technologies: scale, not uniqueness

Combat Systems makes the Abrams, the Stryker and a wide range of wheeled and tracked vehicles, and says it is the world’s largest producer of Ball Powder. It earned a 14.4% margin in 2025. The rival here is the Army’s own spending plans.³ When the Army cancelled the M10 Booker and cut Stryker volumes, US vehicle revenue fell by US$203M twice in a row.³⁵ Technologies is a scale player in federal IT and communications; its margin of 9.5% is the evidence that competition on price is real.³ These are good businesses, and I do not see them as a reason to expect more than the market rate.

Where rivals can catch up

The place a challenger could hurt most is the Gulfstream price. Bombardier’s backlog of US$21.8B is not far behind Gulfstream’s US$25.4B, and both serve the same buyers.⁹⁴ On the defense side, the Army is moving to next-generation combat vehicles that the company has to win, and it says it is investing for that.⁵

IVManagement

Two people matter, and one of them is about to change jobs. Phebe Novakovic has been chairman and chief executive since January 2013. Before that she was president and chief operating officer, and from 2010 to 2012 she ran Marine Systems.⁷¹⁰ On 7 October 2026 the board elected Danny Deep, 56, chief executive from 1 January 2027; she becomes executive chairman. Mr. Deep has been president since December 2025 and joined the board in August 2026. He ran the Combat Systems segment from April 2024 to June 2025, was head of global operations until December 2025 and before that ran General Dynamics Land Systems.¹¹ The chief financial officer is Kimberly Kuryea.¹⁰

Who owns it

There is no controlling holder. The largest holders on the proxy were Longview Asset Management at 10.0% for clients, Vanguard at 8.9% on a September 2025 filing, BlackRock at 5.4% on a 2023 filing and Newport Trust, the independent fiduciary of the company’s own 401(k) stock fund, at 5.0%.¹⁰ Those are holdings, not views. All directors and executive officers together own 1.4% including options they can exercise within 60 days.¹⁰

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

Ms. Novakovic’s total pay for 2025 was US$25.9M, which is 241 times the median employee’s US$107,485.¹⁰ The company says 93% of it is linked to company or share performance. Her base salary is US$1.75M; the annual bonus was US$5.73M, 177% of target.¹⁰ At the 2026 meeting 97.7% of the votes cast on pay were in favour C (calculated).¹⁸

The yearly bonus is based 70% on three financial measures, earnings per share, free cash flow and operating margin, and 30% on strategic and operational goals.¹⁰ That mix rewards the numbers I would also check. It has one weakness. Free cash flow in 2025 was more than US$700M above the target, and the committee’s own explanation is stronger-than-planned award activity, which means customer advances arriving earlier than planned.¹⁰ That is a gain in timing, and it is paid as if it were a gain in earnings.

Half of long-term pay is performance share units that vest after three years on average return on invested capital, with a modifier for share return against peers. The 2023 to 2025 units paid out at 128% of target because return on capital was 13.2% against a target of 12.6% and the share return ranked at the 61st percentile.¹⁰ Another 30% is options and 20% is restricted stock.¹⁰ Return on invested capital rewards management for earning more on the capital it has and for not spending on things that earn less, which fits a shipbuilder that is also being asked to spend billions on yards. It does not reward growth in the dollars earned, and it can be improved by letting the capital base shrink, so I watch the dollars as well.

The ownership rules are firm. The chief executive must own stock worth at least 15 times her salary, and the other named officers 10 times.¹⁰ Ms. Novakovic owned 802,357 shares in March 2026, worth about $300M at $373.46, and could exercise rights over another 573,565. Her own money is well in. Mr. Deep owned 34,551 shares and could exercise rights over 135,205 C (calculated).¹⁰¹ He will be paid as chief executive from next year, and the company has not yet published his package, so I cannot say whether the incentives change.

What they did with the money

The decade shows a pattern. From 2016 to 2025 the company produced US$30.4B of free cash flow, paid US$12.6B in dividends and US$11.8B for its own shares, and spent US$10.1B on acquisitions in 2018, mainly CSRA, which took total debt from US$4.0B to US$12.4B.^137141516173 Since then it has repaid debt; total debt was US$13.0B at the end of 2020 and US$7.0B in October 2026, and net debt is US$2.6B.¹⁵⁴ The share count fell from 310 million diluted to 272 million, 12% in nine years, and the quarterly dividend went from $0.76 in 2016 to $1.59 now, an annual growth of 7.7% over ten raises C (calculated).¹³³⁵

Buybacks are described as being done to cover dilution, and they have been uneven: US$1.8B in 2021 and US$231M in 2019. In the first nine months of 2026 the company repurchased 1.7 million shares for US$604M, and 5.1 million shares remain authorised, 1.9% of the total.⁵ The 2018 acquisition is the decision I find hardest to judge. CSRA added revenue and scale in IT services, and the segment’s margin has been 9.3% to 10.2% since, below the group average, so the return on that US$10B is not obvious. The company has not published a return on it.

VFinancials

Four things in these statements do not mean what they first seem to. The segment structure changed: Technologies was created in 2020 from Information Technology and Mission Systems, so the earlier years are the sum of the two. The 2018 acquisition of CSRA lifts revenue from 2018 on. Revenue recognition rules changed in 2018 and the 2018 10-K restated 2016 and 2017: 2016 revenue was first reported as US$31.4B and is US$30.6B here.¹³⁷ And the fiscal year ends on 31 December but the quarters end on a Sunday near the quarter-end, so the third quarter of 2026 ended on 4 October.⁷⁴

With those cautions the picture is steady growth with a falling margin. Revenue went from US$30.6B in 2016 to US$52.6B in 2025, 6.2% a year, and operating earnings from US$3.7B to US$5.4B, 4.1% a year C (calculated).⁷³ Earnings per share rose faster, from US$8.29 to US$15.45, 7.1% a year, because the share count fell and the tax rate dropped, from about 27% of pre-tax earnings in 2016 to 17.5% in 2025 C (calculated).⁷³ The operating margin was 12.3% in 2016, 13.7% in 2017 and 10.2% in 2025. Nine months into 2026 it is 10.5%.⁷³⁴

Read across the segments. Aerospace revenue grew from US$7.8B in 2016 to US$13.1B in 2025 while its margin fell from 17.8% to 13.3%, because the company put four new aircraft into service between 2018 and 2025, and new models start at lower margins.⁷³ Marine Systems doubled revenue from US$8.1B to US$16.7B and its margin fell by 0.4 points C (calculated).⁷³ Combat Systems grew from US$5.5B to US$9.2B at a steady 14% to 15%. Technologies grew from US$9.1B to US$13.5B at about 9.5%, the lowest-growth part after CSRA. The corporate line, mostly equity pay, was US$175M in 2025 against US$17M in 2016.⁷³

Cash is where the business looks best. Free cash flow was 91% of net earnings over the ten years, US$30.4B against US$33.4B, and 94% in 2025 C (calculated).³ The cash is lumpy by year: US$1.8B in 2016, US$3.4B in 2017, US$2.0B in 2019, and US$4.0B in 2025.

What an owner can really take out

Free cash flow is not the same as what an owner can take. The company defines it as cash from operations less capital spending. Equity-based pay, which the company issues to employees instead of paying cash, is a real cost to owners, but cash from operations adds it back. Owner cash, here, is free cash flow less that pay: US$3.76B in 2025 and US$17.0B over the five years 2021 to 2025, against net earnings of US$18.0B C (calculated).³ So reported earnings are a fair guide to the cash an owner can take; the gap is 6% over five years.

One more item could argue against me. Capital spending was US$1.16B in 2025 and is running at US$918M for nine months of 2026, 66% more than a year before, with the fourth quarter “expected to remain elevated”.³⁵ Depreciation of property, plant and equipment was US$680M in 2025, so spending above depreciation was US$481M. The proxy says the elevated investment is especially in the shipyards.¹⁰ I count capital spending in full and do not add back growth, so my owner-cash numbers are conservative if the new capacity earns a return. The memo lines let you substitute your own view. What I cannot see is how much of the new spending is for the new submarine awards and how much is the cost of just keeping up.

Now the 2026 cash. Nine months of free cash flow was US$4.83B against US$3.01B a year before, and that includes a US$1.23B increase in customer advances and deposits, against US$45M in 2025.⁴ Take that increase away and the figure is US$3.60B, up 21% C (calculated). It also includes a US$500M contribution to the pension plans, which reduces the deficit.⁵ The advances are real money, and they fund the inventory of aircraft and submarine parts. But they unwind as the work is delivered. I use the lower figure when I look at the run-rate and let the model count advances only as a small share of added revenue.

The tax line is steady. The effective tax rate was 17.6% in the first nine months of 2026 and 17.2% a year earlier.⁵ I use 18% in the base case and 20% in the bear.

The balance sheet

Net debt was US$11.5B after CSRA at the end of 2018, US$7.1B at the end of 2024 and US$2.6B at 4 October 2026.⁷³⁴ Total debt was US$7.0B and cash US$4.4B, which is more cash than the company needs, and the company repaid US$500M in June and US$500M in August.⁴⁵ The notes are cheap: a weighted coupon of about 3.5%, with the first big maturity, US$750M, in April 2027 and none larger than US$1.0B in any year.³ There is a US$4B committed bank facility that expires in March 2027, which the company says it may renew or replace.⁵ The pension plans are not fully funded: the funded status of the qualified plans was a deficit of US$763M at the end of 2025, after US$537M of contributions that year.³ I count US$0.3B as a claim, which assumes the US$500M contribution in 2026 closed most of the gap and that returns and discount rates did not move it; I cannot see the current figure.

The dividend is the part of the record I find most straightforward. It has been raised every year for 29 years, the quarterly rate is $1.59, and the yearly cost, about US$1.7B, is 37% of the 2026 earnings outlook and about 29% of last-12-month free cash flow C (calculated).⁵ At $373.46 the yield is 1.7%. It is a secure payment, not a large one.

The price in numbers

On those numbers the shares trade at 17.7 times 2026 operating earnings and 21.6 times the earnings outlook, and the last-12-month free cash flow yield is 5.7%, but that last figure has the US$1.2B of customer advances in it. Ten years ago, on 2 November 2016, the shares closed at $149.67; they are 2.5 times that, plus dividends C (calculated).¹ They are 5.7% below the August high.¹

VIValuation

First, how this goes badly wrong

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

The submarines cost more than the contract pays, or arrive more slowly than the backlog implies. Marine Systems is a third of revenue and 67% of the backlog, and it earns about 7% on fixed-price and cost-plus work whose price was set before the supplier chain could be tested.³⁴ In 2024 supplier quality problems and late deliveries took Virginia Block IV costs up and the segment margin down to 6.5% from 8.1% to 8.6% in earlier years.³ If that repeated across a backlog now more than twice the size, a point of margin on US$19B of 2026 revenue is about US$190M of operating earnings, and two points is 6% of the company’s total C (calculated).⁴ A second part of the same risk is that US$99.6B of the Marine backlog is unfunded and depends on appropriations, with the government on a continuing resolution to 11 December.⁴⁵

The one customer changes course. 68% of revenue is the US government.³ The Army is re-planning its fleet and cancelled the M10 Booker; federal staff cuts and contract changes hit IT services in 2025.³⁵ If the next budget moves money away from ships, vehicles or IT, or ties it to terms the company cannot meet, there is no other buyer for a submarine.

I pay a full price for a good result. The shares are 11% higher than at the end of 2025, when they closed at $336.66, and trade at 17.7 times 2026 operating earnings and 21.6 times earnings.¹ The jet business has had its best deliveries in the ten-year record, 158 in 2025, and the third quarter ended with the largest backlog in the ten-year record.³⁴ A high price for the best results leaves less room if either the multiple or the earnings moves the wrong way.

What an owner can really take out

In 2025 owner cash was US$3.76B, or $13.74 a share, 3.7% of the price C (calculated).³ That is 89% of the 2025 net earnings of US$4.21B. For the nine months of 2026 the figure without the increase in customer advances is US$3.43B, against US$2.82B a year before. In my base case cash before interest and dividends is US$4,444M in 2027, US$5,122M in 2029 and US$5,474M in 2030 E (estimate). That cash is after equity pay, after capital spending of 2.2% to 2.5% of revenue, and after tax at 18%, and before interest, which is small at US$215M.⁵ Reported earnings per share of US$17.25 and my cash per share are close, because depreciation and capital spending are about equal; if the new capacity needs to be replaced early, my cash falls.

What I can’t judge

I can’t say how fast the submarine supply chain can grow, and neither can the company. I can’t say what margin the new Virginia and Columbia awards carry; the company says new production contracts begin lower and improve, and the Q3 documents give no figure.⁵ I can’t say what Congress will do after 11 December. I can’t say whether the 2026 rise in customer advances reverses; it is US$1.2B in nine months. I can’t say what Mr. Deep will do differently. And I can’t say where the business-jet cycle is: deliveries are at their highest in ten years, and the order book is longer, but demand for a discretionary purchase can turn in a year. I show ranges where I can’t be exact.

How the cases are built

I build each case year by year in US dollars, from 2026 to 2036, then a value after that. For Aerospace I multiply aircraft delivered by revenue per aircraft, starting at US$60.7M, add services that start at US$4.15B, and apply a margin. For the other three segments I grow revenue and apply a margin. Corporate costs are US$195M in 2026 and grow 3% a year. Fourth-quarter 2026 is my own estimate, set so that full-year revenue is at the top of the company’s outlook, and only that quarter’s cash counts, because the net debt I deduct is the balance at 4 October. Cash is operating earnings after tax (18% in the base case), plus depreciation and amortization of about US$930M a year, less capital spending of 2.2% to 2.5% of revenue, less 6% of any added revenue for working capital. Equity pay is left in as a cost. I discount at 8.5%: the 10-year Treasury yield of 4.91% on 2 November 2026 plus 3.6 points.⁶ It is more than a bondholder would ask and a little less than I would ask of a business without a government customer. After 2036 the last year’s cash grows at 2.0%, 2.5% or 3.0% in the bear, base and bull.

Three comparisons help. My base 2026 operating earnings are US$5,941M, in line with the company’s outlook of about 10.5% on US$56.5B to US$56.8B. My base then grows 4.5% a year to US$9,213M in 2036, compared with 4.1% a year from 2016 to 2025 C (calculated). That is faster than the past decade, and I give it because Marine Systems revenue is already 8 times covered by its backlog and Gulfstream is delivering the new family. My base has the company margin at 10.7% in 2030, half a point above 2025, not back to the 12% to 14% of 2016 to 2018. And the bear takes the margin to 8.9% in 2030 with the submarine problem repeated and a jet downturn together.⁴³

In plain words, the price needs a margin of 12.2% by 2030, about the 12.3% the company earned in 2016 and 2018, or revenue growth of 7 points a year faster than I model for four years, which would be US$87B of revenue in 2030 against US$70B in my base. The first is possible but I can find no sign of it yet: the company margin has stayed between 10.0% and 10.8% since 2021, though it has not fallen.³⁴ The second needs more than the backlog can turn into revenue, given the supplier constraint. The price is therefore a bet on the margin recovering, and it is not a crazy bet. It is one the filings do not yet show.

The margin of safety, and the price against the range

My margin-of-safety price is $241. It is 75% of my base value. I take a smaller cushion than the 30% I would ask of a cyclical or highly indebted company, because the cash flows here are more visible, the balance sheet has net debt of 0.4 times operating earnings, and the work is under contract. The 25% still matters. It covers a one-point error in the discount rate, which moves my base from $321 to $274, and about 1.5 points of margin in every segment. At $241 the price would be 14.0 times the earnings outlook.

The price reflects a real debate, and the case for the market being right needs naming. It says that Navy submarine spending is the most protected line in the US budget, that backlog is now 3.9 times a year of revenue, that Gulfstream has a new generation of aircraft and a margin gap to its rivals, that the company is investing in capacity to match, that the dividend is secure and the buyback steady, and that a leader well known inside the company is replacing the one who built the current record. If those hold, the margin recovers, and the price is fair. My base value of $321 is 16% below the price, so I am asking for either a lower discount rate than 8.5% or a margin above the 2025 level to pay $373.46. I cannot say that is wrong. I can say it is not yet in the numbers.

The next best use of the same dollar

The best comparables I have worked through on this site, from the price against my base, are Berkshire Hathaway at 16% above my base and GE Vernova at 34% above. General Dynamics, at 16% above, sits next to Berkshire, which my page describes as more than half cash and stocks. The US Treasury ten-year pays 4.91%, and I ask 8.5% from a business with this kind of risk.⁶ My base return to October 2036 is +6.5% a year, which includes the dividend and the cash that buys back shares, so the base falls short of 8.5% by about two points E (estimate). The bull gives +9.7% and the bear +0.7%. The dividend itself yields 1.7%.

The bull case needs many things at once, so I multiply them out. Marine Systems growth near 10% a year for four years with an 8.5% margin (about 25% likely, by my guess), 200-plus Gulfstream deliveries with a 17% Aerospace margin (30%), a 14.5% Combat margin with 4% to 5% growth (40%), Technologies growing 4% to 5% at a 10% margin (35%), and no budget stall or shipyard cost surprise in ten years (50%). Together that is about 0.5%, if they are independent. They are not, because the yards, Congress and the Army depend on the same budget, so perhaps 3% to 5% is a fairer chance E (estimate). The base needs fewer things; the price needs most of the bull.

Where this could be in 5 and 10 years

For the base to be right at five years, Gulfstream has to deliver near 180 aircraft, Marine Systems revenue has to reach about US$28B and the company operating earnings US$7.8B in 2031, and the value is then $370 a share, about 1% below today’s price, because five years of cash have gone out in dividends and buybacks and the rest has grown. At ten years the base is $422, and an owner has also received about $206 a share of cash. The bear gives $252 and the bull $564. These values are mine and are not forecasts.

VIIThe case for the price

What most people miss

Three things cut in General Dynamics’ favour and get little attention. First, the debt is cheap and long. The notes carry a weighted coupon of about 3.5%, none is larger than US$1.0B in any one year, and net interest expense is guided at US$215M, 3.6% of 2026 operating earnings C (calculated).³⁵ Second, customers fund the work. Customer advances and deposits are US$11.1B, 11% of the market value, and rose US$1.3B in nine months.⁴ That is money from the Navy and from aircraft buyers sitting in the company’s account until the work is done. Third, the Aerospace order book is mostly firm. Of US$25.4B of Aerospace backlog, US$24.5B is funded, 96%, while only 32% of the Marine backlog is C (calculated).⁴ The jets are the better-funded part of the story.

The bull case

In plain words: General Dynamics is the prime contractor on the two submarine programmes the Navy has said it will build for the next two decades, and it has just won the contracts. It has US$220.3B of backlog.⁴ Gulfstream has refreshed its whole line, and the margin is climbing, 14.7% for nine months against 13.6% a year earlier.⁵ Munitions demand is elevated, and Europe is buying vehicles.⁵ The company has said it is poised for continued performance after a decade of investment, in the words of its chief executive in the third-quarter release.⁴ It has put numbers on 2026: revenue of US$56.5B to US$56.8B, a margin around 10.5% and earnings per share of US$17.20 to US$17.30.⁵ It has also said its Electric Boat workforce will grow to support one Columbia and up to two Virginia boats a year.³ If the submarine supply chain can support that rate, and the new awards earn a better margin as the programmes mature, the revenue is there for a decade.

I found no named analyst note or fund letter on this company that I could verify, so the bull case here is management’s. The holders I can document are in the March 2026 proxy: Longview Asset Management at 10.0%, Vanguard at 8.9% and others.¹⁰ Those are holdings, not statements of view.

Against reality

If General Dynamics hits its 2026 outlook and then grows Marine Systems revenue 8% to 9% a year and delivers 180 Gulfstreams, the model gives my base, $321 a share. The price of $373.46 is 16% above that. It is possible, and it needs more than the plan. To be worth $373.46, the company would have to earn an operating margin of about 12.2% in 2030, against 10.5% now, or grow revenue about 7 points a year faster than I model, to US$87B in 2030. The first is the 2016 to 2018 level of the margin, before CSRA and the new aircraft family; the second is what the backlog cannot supply while the suppliers are the constraint.

The distance is about the margin more than the volume. Today’s earnings are paid for at 21.6 times the outlook. At 19 times, which is what my base value implies on the same earnings, the same company is worth $321. I multiplied out the odds for my bull in Valuation: about 0.5% if the five conditions are independent, perhaps 3% to 5% with the overlap E (estimate). The price asks for most of the bull.

What a 10x would take. Ten times $373.46 is $3,735, a market value of about US$1.0 trillion. At 20 times earnings that needs about US$51B of net earnings, 12 times the US$4.21B of 2025 C (calculated). At a net margin of 8.0%, the 2025 level, that is revenue of about US$639B, 12 times today’s. In ten years that is a revenue growth of 28% a year, against 6.2% a year from 2016 to 2025. Even at 25 times earnings and a 10% net margin it needs about US$409B of revenue, 7.8 times today’s. The whole US defense industrial base does not offer that much new work for one contractor, and the company’s own backlog of US$220B is about four years of today’s revenue. This is not a plausible outcome. I know of no large defense contractor that has made its shareholders ten times their money in ten years from this size, and I have not checked every list. The shares are 2.5 times their November 2016 price, with dividends on top.¹

Where I come down

What moved me was the cash and the backlog: free cash flow of US$3.96B in 2025, 94% of net earnings, net debt cut to US$2.6B, a Gulfstream margin of 14.5%, and US$93.8B of defense orders in a quarter.³⁴ What did not move me was the price, because I cannot build a case in which the margin I can see supports $373.46, and the cash flow of the last 12 months has US$1.2B of advances in it. What would raise my values is a Marine Systems margin above 8% on the new awards, a company margin of 11.5% or better for two years with free cash flow still near 90% of earnings without advances, and appropriations that fund the unfunded backlog on time. What would lower them is a Virginia or Columbia cost problem on the scale of 2024, or a budget that moves money away from ships.

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.

• Fixed-price contracts. About half of US government revenue is fixed price.³ The company says the 2024 results of Marine Systems were hit by supplier quality issues and late deliveries on Virginia Block IV.³ I model the effect in the first scenario below.
• Appropriations. The government is on a continuing resolution to 11 December 2026; the fiscal 2027 defense bills have not passed.⁵ US$99.6B of Marine backlog, and US$108.5B in total, is unfunded.⁴ The company says the present resolution is not expected to matter.
• Tariffs and trade. Tariffs reduced Aerospace margin by 30 basis points in 2025. A February 2026 Supreme Court ruling led to partial refunds, and the company says the rest are uncertain in timing and not material.³⁵
• Supplier and partner delays. The Israel-based supplier of mid-cabin airframes has had delays tied to the conflict in the Middle East, and shipbuilding suppliers face a labour shortage.³⁵
• International contracts. A large international tracked-vehicle contract carries US$1.3B of unbilled receivables, with deliveries to early 2028, and variable consideration that the 10-K says could be lower than estimated.³
• Succession. A new chief executive takes over on 1 January 2027, with the previous one as executive chairman.¹¹ The pay design for the new role is not yet published.
• Pension. The qualified plans were US$763M short at the end of 2025 before the US$500M paid in 2026.³⁵

What a bad year looks like in the model

Each of these takes between $6 and $35 off the $321 base, and the worst is a lower margin that stays. What matters is how long it lasts. A two-year stall or a jet downturn that ends costs a few dollars, because most of the value sits in the cash after 2030 and the deposits come back with the recovery. A margin one point lower for good costs $35, and a slower ramp on the biggest segment $12. The risk is not a bad year. It is a lower margin that stays.