BAE Systems

IThe business

BAE Systems designs, builds and supports military equipment and the systems inside it, for governments. In 2025 it had sales of £30.7bn, 8% more than in 2024 on a reported basis and 10% more at constant currency, with underlying earnings before interest and tax of £3,322m, a return on sales of 10.8%, and 111,400 people in more than 40 countries; by 30 June 2026 the workforce was 112,400.⁴⁵ The head office is in London, but the largest customer is the United States: 43% of 2025 sales went to customers there, 27% to the United Kingdom, 12% to Europe outside the UK, 9% to Saudi Arabia and 4% to Australia.⁴ I use the company’s own “sales”, which add the group’s share of joint ventures such as MBDA to the IFRS revenue of £28.3bn.⁴

Air is the biggest sector by sales and by backlog. It makes major airframe units of the Eurofighter Typhoon for its core European customers (13 completed in 2025), has delivered the last two of 24 Typhoons to Qatar, and signed in 2025 to build 20 Typhoons for Türkiye, which the company says is worth £4.6bn to it, with first delivery in 2030.⁴ It also builds the rear fuselage of every F-35 for Lockheed Martin at Samlesbury: 153 in 2025, about 150 a year expected to continue, with the price of Lots 20 to 22, which run to 2030, still under negotiation with Lockheed.⁴⁵ That sentence matters for risk, because the customer is a prime contractor with its own interest. The third strand is GCAP, the Global Combat Air Programme, which the UK calls Tempest. In 2025 BAE launched Edgewing with Leonardo of Italy and JAIEC of Japan as the joint company that designs the new fighter, and in 2026 Edgewing won its first two international contracts, £0.7bn in April and £4.6bn in July.⁴⁵ The company expects the design authority to last beyond 2070.⁴ Air also includes BAE’s stake in MBDA, the European missile maker, which won more than £4bn of orders for BAE’s share in 2025, and a large Saudi business that supports Typhoon and Tornado fleets under the Saudi British Defence Co-operation Programme; that five-year cycle ends in December 2026 and the renewal for 2027 to 2031 is being bid.⁴⁵

Maritime builds the Royal Navy’s nuclear submarines at Barrow-in-Furness. Five of seven Astute boats are delivered; four Dreadnought boats, which will carry the UK nuclear deterrent, are all under construction; and the SSN-AUKUS design work, with mobilisation in Australia, has started.⁵ In July 2026 the company announced a new £5.9bn Dreadnought contract.⁵ It is also building eight Type 26 frigates for the Royal Navy, with five under construction, and the Hunter class for Australia, and in 2025 Norway selected the Type 26 for a government-to-government deal the company describes as worth £10bn.⁴⁵ The sector earned 6.7% on sales in 2025 and 6.1% in the first half of 2026, because, the company says, several first-in-class programmes trade at low margins.⁴⁵ That is the weakest part of the group, and it is a quarter of sales.

Electronic Systems is the profit engine, with a 15.4% return on sales in 2025. It makes electronic warfare equipment for the F-35 (190 Block 4 suites delivered), the APKWS laser-guided rocket (the 100,000th kit has been delivered), and since February 2024 the space business formerly owned by Ball Corporation.⁵ Platforms & Services had a strong run: 17% sales growth in 2025 and return on sales up from 10.2% to 11.4%, on US vehicles (500 AMPV delivered, Bradley, the M109A7 howitzer) and European ones (the CV90 for Czechia, 246 vehicles for $2.2bn, and the Bofors ARCHER and TRIDON air defence systems for Sweden).⁴⁵ Cyber & Intelligence, at £2.4bn and 9.3%, is the smallest and slowest, and from 2027 it will be reported inside a four-sector structure.⁴⁵

Space, and what Ball Aerospace was meant to do

On 17 August 2023 BAE announced it would buy Ball Aerospace from Ball Corporation for about $5.55bn in cash.¹² The deal closed on 16 February 2024 at $5.5bn, about £4.4bn.¹³ The company set its own tests: more than $2bn of annual revenue, revenue growth of about 10% a year for five years, a price of about 13 times 2024 EBITDA after a $750m present-value tax benefit and cost savings, and a return on invested capital above the cost of capital within five years of completion.¹² The part I can check is growth. The business is now called Space & Mission Systems and its sales grew 29% at constant currency in the first half of 2026, with orders of £1.6bn including $373m on the $1.2bn Epoch 2 missile warning satellites for the US Space Force.⁵ Electronic Systems sales rose 35% in 2024 (the year of the purchase) and 8% in 2025, which now includes a full year.⁴ The return test runs to February 2029 and I cannot yet judge it. What I can see is the cost carried in the statements: £414m of amortisation and impairment of acquired and other intangibles in 2025, much of it from this deal, which underlying earnings leave out and basic earnings include.⁴ Space is the part of BAE that is closest to the new US priority, Golden Dome, the missile-defence system, which the company says its technologies line up with.⁵

Who pays: the budgets

The demand behind the backlog is political, so I read the budget documents. At the Hague summit on 25 June 2025 NATO allies committed to 5% of GDP on defence and security related spending by 2035: at least 3.5% on core defence and up to 1.5% on related areas such as infrastructure and the industrial base, with a review in 2029.⁶ NATO’s 2026 edition shows European allies and Canada spent more than US$571bn in 2025 in 2021 prices, up nearly 20% in real terms, or 2.3% of GDP.¹⁴ The UK government’s Strategic Defence Review of June 2025 set 2.5% of GDP by 2027 and 3% in the next Parliament when fiscal and economic conditions allow.¹⁵ In the United States the President’s fiscal 2027 budget proposes US$1.5 trillion for defense, made of US$1.15 trillion of discretionary funding (28% more) and US$350 billion of mandatory funding, with US$65.8bn for shipbuilding, funding for the Golden Dome missile defence, and a priority on twelve critical munitions.¹⁶ BAE says it supplies most of those munition types and has already agreed to quadruple output of the infrared seeker for the THAAD interceptor.⁵ I have not opened a European Union spending document, and the UK Defence Investment Plan was not yet available to me, so the UK and European figures here are the NATO and Strategic Defence Review statements, which are commitments to spend, not appropriations. A promise to spend 3.5% is not a contract. The backlog is.

That backlog is £84.0bn, up from £83.6bn at the end of 2025, and the company counts in it its share of joint ventures.⁵ The IFRS order book, which counts only the group’s own unsatisfied contracts, was £62.7bn at 30 June 2026.⁵ I use £84.0bn for the size of the work in front of the business and £62.7bn when I compare with the accounts. Order intake was £36.8bn in 2025 and £16.4bn in the first half of 2026, a book-to-bill ratio of 1.2 in 2025.⁴⁵ Backlog has almost doubled in ten years, from £42.0bn in 2016, but sales rose 61% over the same years, so the extra work is mostly long programmes (submarines, frigates, the new fighter) that earn over a longer time.¹⁷⁴

The “who pays” question has a sharper answer than the geography suggests. The filing says credit risk on defence sales to government customers is considered extremely low.¹¹ It does not give a percentage of sales to any one government. By location, the United States is 43% of sales and the United Kingdom 27%, so two governments are 70%.⁴ I could not open the IFRS 8 customer note in the annual report because the company’s site blocked automated downloads, so I do not state the share of sales to the US government alone. It will be most of the 43%.

IIThe moat

The test of a durable business is what it did when conditions were hard, and for BAE the hardest recent test was 2016 to 2020, when sales were flat, a pension deficit of £6.1bn needed cash, and the company reorganised. It did not lose money in any year. Underlying profit rose from £1,905m in 2016 to £2,205m in 2021, return on sales stayed between 9.8% and 10.5%, the dividend was raised every year from £0.213 to £0.251, and the Group’s share of the net pension deficit went from £6.1bn at the end of 2016 to £2.1bn at the end of 2021 and to a small surplus today.^171018194 That is a business whose results move slowly. The weak spot was cash, not profit: free cash flow was £615m in 2018 and £367m in 2020, with a £1bn payment into the UK scheme in 2020, and over 2016 to 2020 it averaged £759m a year C (calculated).¹⁰¹⁸¹⁹

What protects the business is a position, not a price. Four things make it hard to dislodge. First, incumbency on programmes that run for decades: the Royal Navy’s submarines are built at Barrow, the Typhoon production line and the F-35 fuselage line exist, and GCAP’s design authority is meant to last beyond 2070.⁴ A government cannot easily move a half-built submarine. Second, security rules: when it bought Ball Aerospace, the company noted that more than 60% of its 5,200 employees held US security clearances and said the purchase gave a substantial increase to its US classified revenues.¹² Third, local presence: governments prefer to buy from a company with plants in their own country, and BAE has them in the US, UK, Sweden, Australia and Saudi Arabia.⁴ Fourth, technical depth in electronic warfare, sensors and munitions, where it reports a 15.4% return in Electronic Systems.⁴

The test I would apply to any moat is pricing power, and this is where I am careful. A business with a real edge can raise prices or keep margins above its peers. BAE’s return on sales has been between 9.8% and 10.8% for ten years. Lockheed Martin earned 10.3% on revenue in 2025, Northrop Grumman 10.8%, General Dynamics 10.2% and RTX 10.5% C (calculated).⁷⁸⁹²⁰ The customer sets the price on most programmes, either by cost-plus contract or by competition between a few suppliers, and caps the profit. So the edge shows in volume and stability, not in margin: visibility of work. The backlog is 2.7 times 2025 sales and the IFRS order book 2.2 times C (calculated).⁵⁴ That is worth something, and I price it, but it is a different thing from earning an unusual return.

Return on capital agrees. Net assets of £11.9bn, net debt of £3.8bn and net lease liabilities of £1.7bn at the end of 2025 total £17.5bn; underlying EBIT of £3,322m taxed at 22% leaves £2.6bn, a return of 14.8% C (calculated).⁴ That is above my 8.5% cost of capital, but £12.7bn of that capital is goodwill, much of it from past purchases, so the figure says more about what was paid for the businesses than about how well they use new money. The company has also returned most of its cash, not reinvested it: net capital spending of £957m in 2025 compares with about £0.76bn of depreciation and amortisation of its own assets, after taking out £414m of amortisation and impairment of acquired and other intangibles from the £1,173m total.⁴

The steadiest evidence of quality is per-share growth. Underlying EPS rose from £0.403 in 2016 to £0.752 in 2025 (a compound 7.2% a year) and the dividend from £0.213 to £0.363 (6.1% a year), while the weighted share count came down from 3.17bn to about 3.0bn.¹⁷²¹ The dividend has never been cut in those ten years. Its policy is cover of about two times underlying earnings; 2025 earnings of £0.752 against £0.363 is 2.07 times C (calculated).⁵⁴ I take that as a floor under the payout and a sign that management plans to grow it with earnings.

What would break it? A moat in defence is political more than commercial. If a government decides to buy from someone else, or to buy less, or to build at home, volume goes. The F-35 fuselage depends on a Lockheed Martin contract, Typhoon exports depend on government approval, Type 26 for Norway depends on two governments agreeing, and GCAP depends on three. Nothing in the company’s control makes those decisions. I come back to this in Risks.

IIIThe competitive advantage

The question here is what BAE can do that a rival cannot, measured against the nearest companies. The comparison that is easy to make from filings is the profit margin on sales, and BAE is in the middle of it.

The table says the edge is not in the margin: every large defence prime lands between 10% and 11%. It also says BAE is not a lagging business. The differences are in mix. Electronic Systems at 15.4% is as good as it gets in the group, and it is 25% of sales; Maritime at 6.7% pulls the average down, and it is 22% of sales.⁴ If Maritime earned the group average, underlying profit would be about £280m a year higher C (calculated). If Electronic Systems grows faster than the rest, the group margin will drift up without any change in how any single business is run, which is what I assume in my base case and is a bet on mix. In the first half of 2026 the group margin was 10.8%, up from 10.6%, with Electronic Systems at 15.5%, Air 11.9%, Platforms & Services 11.9%, Cyber & Intelligence 9.0% and Maritime at 6.1%, down from 6.8%.⁵

What BAE has that its US peers do not is range of customers. Lockheed Martin, Northrop Grumman and General Dynamics have mostly US government customers. BAE’s sales are 43% US, with the rest across the UK, Europe, Saudi Arabia, Australia and Asia.⁴ In practice that means when European spending rises, as it is doing, it gains from it directly through Typhoon, Hägglunds and Bofors vehicles, Type 26, MBDA and Edgewing, while still having its US business as ballast. Europe excluding the UK grew from £2.8bn in 2024 to £3.6bn in 2025 of sales, up 28%, and in the first half of 2026 to £2.1bn from £1.6bn.⁴⁵ I find this a real advantage for the next five years. It is not an advantage over the US primes in the US, where BAE is a mid-size competitor in a market it does not control.

Where it is weaker. First, it makes parts for others in some of its best-known work: the F-35 rear fuselage is a subcontract to Lockheed Martin. Second, it is a smaller player in the weapons that the US now wants most. BAE says it is a key supplier to most of the 12 critical munition types, often as the maker of a part such as the THAAD seeker, which gives it a share of the volume but not the economics of the company that owns the finished weapon.⁵ Third, the heavy shipbuilding is in the UK and Australia, where it is the main builder and therefore carries the risk of first-of-class overruns. Fourth, the new fighter is shared: GCAP is a joint company with Italy and Japan, so BAE does not control the programme or take its whole profit.⁵

The other point on advantage is capacity. The company is spending about £1.0bn a year on capital expenditure and £407m on its own research and development, and says it will invest more than $300m over five years in Hägglunds and $135m more in Texas and New Hampshire for guided munitions.⁴⁵ Whether that investment earns more than it costs is the open question for cash. The NATO share of GDP can rise a lot and not help an owner if the extra volume needs capacity that the contract does not pay for. The evidence of the first half of 2026 is encouraging: sales up 9% and profit up 11% with a margin that went up.⁵ I treat it as a test I have not yet seen through a full cycle.

IVManagement

The people who matter are Charles Woodburn, Chief Executive since 1 July 2017, when he succeeded Ian King, who had led the company since 2008, and Brad Greve, the Chief Finance Officer; Cressida Hogg is Chair, and Tom Arseneault, who runs the US business, BAE Systems, Inc., sits on the board.¹⁰⁵ Woodburn has put his name to every results announcement I read from 2017 on. The board is adding Michelle Hinchliffe, a former KPMG leader who sits on the boards of BHP and Macquarie, as a non-executive director from 1 September 2026, with a seat on the Audit and Risk Committee, and the Chair said she was also chosen because the Australian business is growing.²²

How they are paid, and what I could not verify

This is the part of the page where I have a gap. How a chief executive is paid is in the remuneration report in the annual report, and the company’s website blocked automated downloads of the 2025 report, so I could not read the bonus measures, the long-term plan targets, the pay totals or the holding requirements.²³ I will not guess them. What I can say from other filings is limited and I give it for what it is worth. The company guides to four measures: sales, underlying EBIT, underlying EPS and free cash flow, and a three-year cumulative free cash flow floor.⁵ If pay follows those, as in most large companies, it rewards underlying EPS, which leaves out £414m a year of amortisation of acquired intangibles, so an acquisition looks better on that measure than on basic EPS. In 2025 underlying EPS grew 12% and basic EPS 6%.⁴ And the cumulative cash floor rewards the timing of customer advances as much as it rewards earnings, a point I come back to in Financials.

On their own money in the shares, the only dealings I found in the four months before this page are monthly purchases in the company’s Share Incentive Plan: the Chief Executive, the Chief Finance Officer and other senior managers each buy six to eight shares a month, roughly £100 to £150, and receive three matching shares.²⁴ That tells me nothing about conviction, either way. The larger holdings they may have through share awards are in the report I could not read. A cost I can see is what the shares cost the company: the equity-settled share schemes charged £174m in 2025, up from £92m in 2021.⁴¹⁹

What they did with the cash

Over ten years BAE generated £14.9bn of free cash flow. It paid 53% of it as dividends, spent 19% on buybacks, and used 45% on acquisitions, and net debt rose by £2.3bn C (calculated).¹⁷⁴ The decision that matters is Ball Aerospace: £4.8bn in 2024 is 71% of everything spent on purchases and 32% of ten years of free cash flow. The company paid about 13 times EBITDA after tax benefits and savings; the group itself trades on about 16 times 2025 underlying EBITDA (£3,322m of EBIT plus about £0.76bn of depreciation and amortisation of own assets, against an enterprise value of about £66bn) C (calculated).¹²⁴ so on price it was a purchase at a lower multiple than its own. The promised 10% annual revenue growth is being beaten so far; the promised return above the cost of capital cannot be judged until 2029.⁵

The buyback is the point I would give most scrutiny. The stated policy is to return extra cash when the balance sheet allows and when the return is above the cost of capital, and £1.5bn programmes have been announced twice, in 2022 and 2023.⁵ But it barely shrinks the share count. In the first half of 2026 the company bought back 13m shares for £251m and issued 20m shares to employees from treasury, so the shares outstanding for earnings per share rose from 2,988m to 2,995m.⁵ Over 2025 buybacks added £0.003 to a £0.752 underlying EPS.⁴ In effect the buyback pays for the employee share schemes. I count the scheme cost (£174m) as a cost of doing business in my valuation and the buyback as a payment to owners only to the extent that it exceeds the shares issued.

The cultural record is steadiness, and the evidence is the guidance record: the company has raised its 2025 and 2026 guidance at the half-year each time and ended 2025 above the upgraded range on profit, EPS and cash.⁴⁵ The caution is the free cash flow guidance for 2025 of more than £1.1bn given in November 2025, against an actual £2,158m two months later: customer advances arrived late in the year.²¹⁴ A floor is not a forecast, but one that was beaten by about £1bn is not a number I can lean on, and it is why I do not take the 2026 target of more than £2.0bn as a base for the future.

VFinancials

Five things in BAE’s statements don’t mean what they first seem. Sales are not revenue. The company’s main measure adds its share of joint ventures to IFRS revenue, so 2025 sales were £30.7bn against IFRS revenue of £28.3bn; in Air alone the first half of 2026 shows sales of £4.9bn against revenue of £3.8bn, because MBDA and Eurofighter come in at the company’s share.⁴⁵ Underlying profit is higher than profit. 2025 underlying EBIT was £3,322m and operating profit £2,925m, and the £397m gap is mostly £414m of amortisation of acquired intangibles, in particular from Ball; underlying EPS was £0.752 against basic £0.688.⁴ Free cash flow has customer money in it, which I show below. The pension surplus of £1,075m is an accounting number, net of withholding tax on a refund and not cash an owner can take out.⁵⁴ And the ten-year series has breaks: profit was EBITA until 2019, IFRS 15 and IFRS 16 began in 2018 and 2019, and 2020 cash includes a £1bn payment into the UK pension scheme.¹⁸¹⁹

Read across the ten years, sales rose 61% (5.5% a year) and underlying profit 74%, with all of the pace in the last five years: from 2020 to 2025 sales grew 8.0% a year and underlying profit 10.3% C (calculated).¹⁹⁴ Return on sales did not change much, from 10.0% to 10.8%, so growth came from volume. Cash is the item with a story. In 2016 to 2019 the company paid deficit contributions into the UK and US pension schemes of £271m (2017) and £231m (2019) beyond the service cost, and put in a further £1bn in 2020; since 2022 the Group’s net position has been a surplus and those payments are gone.¹⁰¹⁸¹⁹ That is one reason cash today is higher than it was, and it does not repeat.

Cash: how much is real

Free cash flow of £2,158m in 2025 and £11,070m over 2021 to 2025 is the number the company guides to and the number most reports use. For an owner the right question is what is left after costs that are real but sit outside it, and what part of it was borrowed from customers. The first is small: the cost of share schemes (£621m over five years) and dividends to minority holders in subsidiaries (£631m) R (reported). Owner cash after those was £9,818m, 89% of reported free cash flow C (calculated).¹⁹¹¹⁴ The second is larger. Working capital, which includes customer advances, added £1,897m over the five years, so owner cash before working capital was £7,921m, an average of £1,584m a year, 72% of reported free cash flow.¹⁹¹¹⁴ In 2025 the working-capital lines took cash out (−£345m), so owner cash before working capital that year was £2,243m, the best of the five. I therefore do not argue that cash is overstated every year. It is lumpy, and in 2026 the lump runs in the other direction.

Contract liabilities are £7.9bn at 30 June 2026, 25% of annualised sales, up from £3.4bn at the end of 2021 (16% of sales).⁵¹⁹ Customers pay in advance for submarines, aircraft and munitions, and the company uses that cash to fund its supply chain. It is real money, but it is the customers’ until the work is done. As long as the backlog grows, the balance grows and free cash flow stays above profit. When programmes mature and deliveries catch up, the balance falls, and cash runs below profit. The first half of 2026 shows it clearly: free cash flow was £1,791m against minus £368m a year before, and the company says that includes net customer advances of £1.6bn, also from MBDA.⁵ Cash at 30 June was £4.2bn.⁵ Management’s full-year target is more than £2.0bn, which implies only about £0.2bn or more of free cash flow in the second half, so the advances will mostly flow out to suppliers.⁵ The cumulative floors tell the same story: more than £6.7bn for 2024 to 2026, more than £5.5bn for 2025 to 2027 and more than £6.0bn for 2026 to 2028. If 2026 comes in at £2.0bn, those floors need only about £1.3bn for 2027 and about £4.0bn for 2027 and 2028 together.⁵⁴ I take them as floors and not forecasts, but they say the company does not promise to repeat £2.5bn a year.

The balance sheet is strong. Net debt excluding leases was £3,844m at the end of 2025, 0.9 times underlying EBITDA by the company’s measure, and £3,173m at 30 June 2026, with £7.4bn of loans against £4.2bn of cash.⁵⁴ The dollar bond of $750m at 3.85% was repaid in 2025.⁵ The cost of money is not a worry. Net finance costs are guided at about £340m for 2026, covered 10 times by underlying EBIT C (calculated).⁵ I do not value the UK pension surplus. It is £1,352m after tax on a refund, and a refund is not within the company’s control, because it depends on the trustees and on the government’s rules on surplus payments. I treat the Saudi end-of-service obligation (£206m) and the US pension deficit (£103m) as debt. I have not seen the funding-basis valuation of the UK scheme, which could differ from the accounting one, and I say so rather than guess.

VIValuation

First, how this goes badly wrong

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

A government changes its mind. Two governments are 70% of sales.⁴ The US spending increase is a request, and a budget request is not an appropriation: the fiscal 2027 proposal of US$1.5 trillion needs Congress, and in November 2025 the company itself warned that a US government shutdown could delay contract funding and payments.¹⁶²¹ In the UK the target of 3% of GDP comes with the words “when fiscal and economic conditions allow”.¹⁵ NATO’s own 5% commitment is to be reviewed in 2029.⁶ A budget that grows 3% instead of 8% for a few years leaves the backlog intact and starves the new orders that replace it. In my tests below, three points less growth a year for three years takes £1.47 a share off the base.

A first-of-class programme goes wrong. The Maritime sector, about a fifth of sales, earns 6.1% to 6.7% because several first-in-class programmes trade at low margins.⁴⁵ The first Type 26 frigates have not yet gone to sea trials, Dreadnought is a new design with a first boat still being built, and GCAP is in design with two partners.⁴⁵ I do not know which contracts are fixed-price and which are cost-plus; the filings do not say, and I treat that as something I can’t see. A £1.5bn overrun is not out of scale for programmes of this size, and the cash would go before it comes back.

I pay a full price for cash that is partly the customers’. At £20.75 the shares are 27.6 times 2025 underlying earnings, about 20 times 2025 underlying EBIT and 18 times 2026 EBIT if the company reaches the middle of its guidance, and the dividend yields 1.7%.¹⁴ The free cash flow of the last two years has had customer advances in it, and 2026 has had £1.6bn so far.⁵ A defence contractor at a high multiple is exposed to two things at once: a growth rate that disappoints, and a multiple that falls even if growth does not.

What an owner can really take out

In 2025 owner cash was £1,898m, or £0.63 a share, 3.0% of the price C (calculated).⁴ Before working capital it was £2,243m, or 3.6%. Over 2021 to 2025 it averaged £1,584m before working capital, 2.5% of the price C (calculated).¹⁹¹¹⁴ Reported free cash flow for 2025 was £2,158m, and underlying earnings to shareholders £2,253m. In 2024 owner cash before working capital was £1,527m against underlying earnings of £2,065m, 74%; in 2025 it was 100%.⁴ The ratio moves a lot, which is why I do not pick one year. In my base case owner cash is £2,109m in 2027, £3,362m in 2030 and £4,016m in 2035. At this price an owner takes about 3% a year of cash out, against a 10-year UK government bond that pays 5.25% and Bank Rate of 3.75%.²⁵²⁶ The rest of the return has to come from growth.

What I can’t judge

I can’t say how much of each government’s promised increase will become orders, or when. I can’t say which contracts carry fixed prices, what the first Type 26 and Dreadnought boats will cost to finish, or how much of the £7.9bn of customer advances stays as the backlog grows. I can’t say what Lockheed Martin will pay for F-35 Lots 20 to 22, which are still under negotiation, or whether the UK Typhoon support contract that ends in December 2026 and the Saudi support programme that is being bid for 2027 to 2031 will be renewed on today’s terms.⁵ I could not read the funding-basis pension valuation, the pay report or the customer note in the annual report, and the UK Defence Investment Plan and any European Union spending document were not available to me. For that reason I show a ladder on return on sales and a cash-conversion line and not one number.

How the cases are built

I build each case year by year in pounds. Sales grow from £30,662m at a rate I set for each year. Underlying EBIT is sales times a return on sales. Cash before working capital is a share of EBIT: I calibrate it to 2021 to 2025, when cash after tax, capital spending, leases and joint-venture dividends, with net interest (excluding lease interest, which is already a cost of leases) added back and working capital taken out, was 76% of EBIT, and I use 74% in the base case because capital spending of about £1.0bn is above the old run rate.⁴ From that I subtract the share-scheme cost (£174m in 2025, growing with sales) and minority dividends (£86m, growing 3%), and add working capital: an outflow of £350m in 2026, £600m in 2027, £400m in 2028 and £200m in 2029 as the advances unwind, then zero. The 2026 figure counts only the second half, because I take net debt at 30 June: full-year cash on the company’s definition of about £2,043m, in line with its guide of more than £2.0bn, less the £1,791m already in that balance sheet. Claims ahead of the shares are net debt of £3,173m at 30 June, plus the £450m interim dividend declared (the shares went ex that dividend before the price date, so the cash is owed and the price excludes it), plus up to £188m of shares bought since 30 June, plus the Saudi end-of-service obligation of £206m and the US pension deficit of £103m, less £190m of other investments: £3,930m in all.⁵²⁷ I do not deduct lease liabilities, because lease payments are already in cash. I value no UK pension surplus. The share count is the 2,998.99m voting shares at 30 October, less about 11m held in trust, plus 25m for employee share schemes: 3,013m.²⁸⁵

I discount at 8.5%: the UK 10-year government yield of 5.25% plus 3.25 points for budget risk, programme risk and the things I can’t see.²⁵ The premium is a judgement, not a measurement. I stop at 2035 and grow the last year’s cash at 2.0%, 2.5% or 3.0% in the bear, base and bull.

Three comparisons help. My base 2026 underlying EBIT is £3,693m, 11% above 2025 and inside the guidance of 10% to 12% growth; the company beat its 2025 guidance on profit, EPS and cash, and raised 2026 at the half-year.⁵⁴ My base then grows EBIT at 5.5% a year to £6.0bn in 2035, against 10.3% a year from 2020 to 2025. Second, my bull reaches £5.9bn in 2030, a 12.0% a year pace for five years, which would be faster than the last five and needs a return on sales of 12.6%, higher than any year in the last ten.⁴ Third, my bear is not a collapse: EBIT stays between £3.4bn and £3.8bn for ten years, the return on sales falls to 9.5%, and cash conversion drops to 70%.

Put in plain words, the price needs about 2.8 points a year more sales growth than my base for nine years, or a return on sales of about 14.1% by 2030, which no large defence prime in my table earns, or cash conversion of 88%, or all of these in part. It also corresponds to a discount rate of 7.4% on my base cash flows, which is only 2.2 points above the gilt yield. None of these is silly alone. Together they say the market is paying for a company that grows at the rate of the last five years for another ten, holds a higher margin and has little risk of the budget turning. That is possible, and it is the reason my bull case sits above the price.

The margin of safety, and the price against the range

My margin-of-safety price is £11.68, US$15.67. It is 70% of my base value. At that price an owner would have a 30% cushion against being wrong about the budget, the programmes and the advances, and the price would be 40% above my bear. It would be about 16 times 2025 underlying earnings and a 5.4% owner cash yield on the 2025 figure. At £20.75 there is no margin of safety. The price is 24% above my base, 13% below my bull, and 2.5 times my bear. My weighted value, with 25% on the bear, 50% on the base and 25% on the bull, is £16.43.

The price reflects a real debate, and the case for the market being right needs naming. It says that NATO allies have just agreed to spend at least 3.5% of GDP on core defence by 2035, that the US request raises discretionary defense funding by 28% in a year, that BAE sits in every important domain (submarines, fighters, vehicles, electronic warfare, space, munitions), that its backlog is 2.7 years of sales and keeps growing, that book-to-bill was 1.2 in 2025, and that a business with that visibility deserves a low discount rate. If I give that its full weight, my bull case is the answer. At 7.5% and 2.5% growth my base value rises to £20.36, close to the price. The case against says that return on sales has not moved in ten years, that every large prime earns the same, that the cash of 2024 to 2026 has been supported by customer money, that the growth must come on top of a base that is already 61% larger than in 2016, and that the price is already £4.06 above my base. 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 pound

The nearest comparisons I have worked through on this site are cash-generative companies with long contracts. Enbridge has a base return to October 2036 of 9.9% a year in my work and Berkshire Hathaway 6.6%; there is no other defence prime on the site, and Palantir is a software company with a price far above its value. BAE’s base return is 5.6% a year E (estimate). The UK 10-year government bond pays 5.25%, and I ask 8.5% from a risk like this one.²⁵ BAE’s base return beats the bond by 0.4 points and falls short of what I ask by 2.9. It differs from the 7.4% 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. BAE is a good business with a long order book. The price is a separate question. A good business at a price that needs more than my base is a poor purchase for the same reason a poor business at a low price can be a good one. I give neither a grade.

The bull case needs many things at once, so I multiply them out. Sales growth near 10% for two years (about 30% likely, by my guess), a return on sales of 12.6% (25%), no budget disappointment in ten years (50%), cash conversion of 80% (40%) and perpetual growth of 3% (50%). Together that is under 1% 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 have to be near £45bn and underlying EBIT about £5.2bn in 2031, and the value is then £19.86 a share, about the same as the price today, because five years of cash have gone out and the rest has grown. At ten years the base reaches £24.09. In the bear it is £11.78; in the bull it is £35.92. The four inputs that drive it are sales growth, return on sales, cash conversion and the 8.5% rate. These are estimates of value, not forecasts of the share price. The market price can sit above or below value for years; FT data show the shares up 11.35% over the past year.²

VIIThe case for the price

What most people miss

Three things cut in BAE’s favour and get little attention. First, Europe is already in the numbers: sales to customers in Europe outside the UK rose 28% in 2025, to £3.6bn, and 35% in the first half of 2026.⁴⁵ Second, the pension swung from a £6.1bn deficit at the end of 2016, with £271m a year of extra payments in 2017, to a £0.8bn accounting surplus, and I found no deficit payments in the 2025 or 2026 statements.¹⁷¹⁰⁴ Third, the company signed a seven-year framework with the US Department of War to quadruple output of the THAAD interceptor’s infrared seeker, and the fiscal 2027 request makes twelve critical munitions one of its highest priorities.⁵¹⁶

The bull case

In plain words: governments in all of BAE’s home markets have decided to spend more at the same time; BAE is already inside their biggest programmes; the backlog is £84.0bn; and the cash is strong. Chief Executive Charles Woodburn said at the half-year that governments are responding to a volatile threat picture with sustained increases in budgets, and that execution, geographic spread, investment and the backlog position the company for long-term growth.⁵ In May 2026 the company listed where it sees opportunities: space, missile and air defence, drones and counter-drone, electronic warfare, combat aircraft and vehicles, frigates and submarines.²⁹ I found no named analyst or fund manager view I could verify, so this case is management’s. One holder I can document is Capital Group, with 10.97% of the voting rights on 20 October 2026; that is a holding in a regulatory notice, not a statement of view.³⁰

Against reality

If BAE hits its guidance and grows as the NATO plans imply, the model gives a value near my base, £16.69 a share, and the price of £20.75 is 24% above that. It is possible, and it needs more than the plan. The price needs about £5.6bn of underlying EBIT in 2030, against £3.3bn in 2025: 10.8% a year for five years C (calculated), compared with 10.3% a year from 2020 to 2025 and my base of 8.5%. That is 11% above my base for 2030 and 5% below my bull, so the price sits between the two. Today’s EBIT is paid for at 20 times; my base value implies about 16 times. The odds for my bull, multiplied out in Valuation, are under 1% if the five conditions are independent and perhaps 4% to 6% with the overlap E (estimate).

What a 10x would take. Ten times £20.75 is £207.50, a market value of about £625bn, or US$839bn. At 20 times earnings that needs about £31bn of net income, 14 times the £2.25bn of underlying earnings in 2025 C (calculated).⁴ At the 2025 net margin of 7.3% that is revenue of about £425bn, 14 times 2025 sales and roughly all that European allies and Canada together spent on defence in 2025, US$571bn in 2021 prices (a rough comparison, because the price bases differ).¹⁴ In ten years it would take 26% a year, against underlying EPS growth of 7.2% a year over the last nine.¹⁷⁴ Companies of this size almost never compound at 26% for a decade, and a defence contractor’s sales are capped by what governments spend. Tenfold from here is close to impossible, and I would not build anything on it.

Where I come down

What moved me was the order book, a return on sales that has stayed between 9.8% and 10.8% for ten years, and a first half in which profit grew faster than sales. What did not move me was the price: I cannot justify £20.75 without a margin above anything the sector has earned, and the cash of the last two years has customer money in it. What would raise my values is a return on sales above 12% with Maritime back above 7%, and the three open contracts (Typhoon support, Saudi renewal, F-35 Lots 20 to 22) agreed on terms like today’s. Each point of return on sales, held from 2028, is worth about £1.72 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. The company’s own list of principal risks has ten headings: government customers, defence spending and terms of trade; contract risk, execution and supply chain; cyber and other security risks; international markets; people; safety; acquisitions including joint ventures; business interruption; legal risk; and climate transition.⁵

• Government concentration. This is the risk behind every other. The United States is 43% of sales and the United Kingdom 27%, and Saudi Arabia, a single government, is 9%.⁴ One of the first two cutting, delaying or re-sourcing a programme moves the group. The Saudi support programme cycle ends in December 2026 and the renewal is in bid; the UK Typhoon support contract also ends in December 2026 and a successor is under negotiation.⁵ The company expects its Saudi in-Kingdom business to remain stable, but I cannot judge it from outside.
• A customer that is also a prime. The F-35 rear fuselage is built for Lockheed Martin, and the value of Lots 20 to 22 is still being negotiated.⁵
• Joint ventures. MBDA, Eurofighter and Edgewing are not controlled; the Group takes its share of profit and receives dividends (£299m in 2025, £202m in the first half of 2026).⁴⁵ GCAP is run jointly with Italy and Japan.
• Currency. The company guides at $1.32 per £1 and says a five-cent move changes sales by about £500m, underlying EBIT by about £70m and underlying EPS by about £0.014.⁵ That is about 2% of EBIT for a 3.8% move; a weaker dollar lowers the sterling profit of the US business.
• Reporting changes. From 1 January 2027 the company reports four sectors, not five, and IFRS 18 will redefine operating profit and move the share of joint-venture results out of it, so the 2027 comparatives will need care.⁵
• Pension. I could not read the funding-basis valuation. The accounting surplus moves with discount rates and asset returns: the Group’s share was £0.8bn at the end of 2025, and the UK surplus alone rose from £1,125m to £1,352m in the first half of 2026.⁴⁵

What a bad year looks like in the model

The surprise is how little a one-off overrun costs, £0.45 on £1.5bn, because the cash goes once and the value sits in the long run. What matters more is anything permanent. Margins staying at today’s 10.8% instead of rising cost £1.49; slower growth for three years costs £1.47; a lower cash conversion costs £1.14. The risk is not a bad year. It is a decade in which the budget promises do not become orders and the margin does not move.