Lockheed Martin

Two days ago we published our Boeing report — the definitive case of management destroying a moat. Lockheed Martin is the natural companion, and we expected to write the contrast: the defence business Boeing does badly, done properly. The research did not support that story. What it supports is something more useful and considerably less comfortable.

Lockheed is the largest defence contractor in the world: $75 billion of revenue, roughly 123,000 employees of whom about 72,000 are engineers and scientists, and a backlog that just hit an all-time record of $230 billion. It builds the F-35 — the largest weapons programme in history, over $2 trillion across its life — plus the F-22, the C-130, Sikorsky helicopters, satellites, and the missiles that have become the story: PAC-3 and THAAD interceptors, HIMARS, GMLRS rockets, JASSM and LRASM cruise missiles.

And here is the fact that this entire report exists to explain. On 19 July 2026, the Chief of Staff of the United States Air Force, General Kenneth Wilsbach, said of the F-35 — a quarter of Lockheed's revenue and two-thirds of its Aeronautics division — that "every aircraft delivered over the last two years hasn't been combat capable," and that aircraft were being delivered without radars. In 2024 Lockheed delivered 110 F-35s, every one of them late, at an average of 238 days. Six were handed to the Marine Corps with ballast in the nose where the radar should have been. Meanwhile the stock is up 36% over twelve months, beating the S&P 500 by roughly nineteen points. Reconciling those two facts is the work of this analysis.

"We need more companies that make nozzles… cases. A third solid rocket motor provider… they'll just have to get to the back of the line." — Chris Kubasik, CEO of L3Harris — an incumbent explaining why new entrants cannot relieve the bottleneck

The thesis of this report is that Lockheed and Boeing are suffering from the same disease at different stages. Boeing's defence unit lost roughly $8 billion on the KC-46 tanker alone — a fixed-price development contract its chief executive calls "a bad contract for the last decade." Lockheed, the supposed adult in the room, has taken four reach-forward losses in three years, two of them on the same classified aeronautics programme, and its own annual report warns it "may need to record additional losses in future periods." The disease is fixed-price development at the technological frontier — promising a price for something nobody has built yet. Boeing hit bottom and is climbing out. Lockheed, on the numbers, is still descending. The difference is that Lockheed's descent is being paid for by the greatest munitions boom since the Cold War, and the market is watching the arsenal instead of the aeroplane.

The unusual difficulty here is not complexity — it is opacity. When Lockheed took a $950 million charge in 2025, the chief executive could describe the programme only as "a highly classified program that can only be described as [a] game-changing capability." Asked how many years of fixed-price commitments the charge covered, he said only that "due to the nature of the classification, we can't say how many years that is, but it is not unlimited." An investor is being asked to underwrite a liability whose size, duration and subject are all secret. That is a genuine limit on the circle of competence, and it is why the annual report's warning that further losses may follow deserves more weight than it usually gets.

Read the two backlog numbers together and you have the whole company. Lockheed's total order book is at an all-time record of $230.4 billion — a fact every bullish write-up leads with. But the record is not being driven by the aeroplane everyone associates with Lockheed. Missiles & Fire Control now carries $87.9 billion of backlog against Aeronautics' $54.4 billion — the missile book is 1.6 times the fighter book, from a segment producing a fifth of the revenue. In seven months Lockheed signed a $58.6 billion multi-year for PAC-3 interceptors (29 July 2026) and a $35 billion THAAD contract — roughly $94 billion, nearly three times the entire F-35 Lot 18–20 award.

This is a company being quietly re-founded. For thirty years Lockheed was a fighter company that also made missiles. On the forward book, it is now a missile company that also makes fighters — and the missile business carries a 14.5% margin against Aeronautics' 9.4%. That rotation is genuinely good news, and it is the reason the stock has done well. It is also why the deterioration in the fighter franchise has gone almost entirely unremarked.

This is the part of the analysis nobody writes, so we will be specific and source it to the Government Accountability Office and the Air Force's own chief of staff.

The aircraft are being delivered without the capability they were bought for. The F-35's modernisation programme, Block 4, was meant to deliver 66 new capabilities by 2026 at a cost of $10.6 billion. Costs grew more than 50% to $16.5 billion by 2021 — and the Defense Department has not published an updated estimate since. The programme has been descoped, not merely delayed: capabilities that require more engine power or cooling have been deferred entirely, and what remains is now due 2031 at the earliest. The GAO records the programme office's own admission that the reduced effort "does not meet the intent of the original Block 4 effort." The list of what survived has never been made public.

The hardware that Block 4 depends on ran three years late. TR-3 — the processors and displays without which the new capabilities cannot run — halted deliveries for about a year. Lockheed then delivered 174 aircraft that were not combat capable, and in 2024 shipped 110 aircraft, all late, at an average of 238 days, up from 61 days the year before. Test pilots found, in the GAO's words, that "TR-3 software did not reliably start up."

And then there is the radar. The new AN/APG-85 is late and is not mounting-compatible with the radar it replaces — it needs new software and a bulkhead redesign. So in early 2026, six F-35Bs were delivered to the Marine Corps with ballast in the nose instead of a radar, with Air Force and Navy aircraft to follow. A Marine Corps captain explained that officials made the decision "with full understanding of the risk of having production aircraft ahead of the Block 4 capabilities."

The root cause is eighteen years old. In 2008 Lockheed discovered the aircraft's cooling system needed more air from the engine than designed. It asked for an engine redesign in 2013; the programme refused as too late and too costly. The consequence, per the GAO: the cooling system is "overtasked, requiring the engine to operate beyond its design parameters," and the resulting wear has added $38 billion to the programme's life-cycle cost. The fix — a new engine core — reaches production in 2031, with the power and cooling upgrade in 2033, the same year the capabilities need it. The GAO notes this leaves zero schedule margin.

The fleet is getting harder to fly. Mission capable rates fell from 67% in 2021 to 44% in 2025; fully mission capable from 38% to 25%. And the customer is buying fewer: the enacted 2026 budget funded 47 aircraft, only 24 of them for the Air Force — roughly a 45% cut to that line, the first time in a decade the Air Force received less than half the Pentagon's F-35 purchase.

We set out to write that Lockheed does fixed-price development well. Here is what the filings actually show.

The 2025 quarter in which the largest of these landed cost shareholders 8.5% in a single day. And the annual report's language on the classified aeronautics programme is unresolved, not closed: "we continue to monitor this program, and we may need to record additional losses in future periods if we experience further performance issues, increases in scope, or increases in costs from prior estimates."

This is precisely Boeing's disease. Boeing's defence unit has lost roughly $8 billion on the KC-46 tanker, took another $280 million on Air Force One in the second quarter of 2026, and is renegotiating that contract as we publish. Lockheed has taken $3.1 billion across four programmes in three years. Neither company's problem is a lack of engineering talent — both employ tens of thousands of superb engineers. The problem is committing to a fixed price for something that has never been built, on programmes where the customer's requirements evolve and the contractor absorbs the difference. Lockheed's own filings name the competitive consequence with unusual candour: "Competitors may be willing to accept more risk or lower profitability in competing for contracts than we are."

The difference between the two companies is stage, not diagnosis. Boeing hit the bottom — negative equity, $33.7 billion of cash burned, a criminal charge — and is now visibly climbing out. Lockheed never came close to that, but its trajectory over the same window has been downward: operating margin from 14.3% in 2019 to 10.3% in 2025; earnings per share from a peak of $27.55 in 2023 to $21.49 today, a fall of 22%. Boeing is a wreck being repaired. Lockheed is a very good company that has been quietly getting worse — and only one of those two facts is in the share price.

The fashionable argument is that Anduril, Palantir and cheap drones are dismantling the defence primes, and that Ukraine proved a $400 quadcopter beats an $80 million fighter. The evidence is more interesting than either the boosters or the sceptics allow.

What genuinely protects Lockheed — and it is more physical than commercial:

1 · The solid rocket motor chokepoint. The US industrial base for rocket motors went from six producers in 1995 to two. A single company, AMPAC, is the sole US source of ammonium perchlorate — a literal single point of failure for every missile in the arsenal. Nozzles run 7–10 month lead times. The best evidence comes from a rival incumbent: L3Harris's chief executive on new entrants — "A third solid rocket motor provider… they'll just have to get to the back of the line." Start-ups are winning $11–14 million awards against Lockheed's $58.6 billion PAC-3 contract.

2 · Clearances and the cold-start problem. Security clearance processing runs 156 days for Secret and 227 days for Top Secret, and a new firm faces a genuine catch-22: you need a facility clearance to work on classified contracts, but you cannot get one without being sponsored for classified work. None of the 2025–26 acquisition reforms touch this.

3 · The budgeting machinery has not changed. The commission that recommended overhauling Pentagon budgeting reported in March 2024; the budget-structure reform targets the 2028 cycle. The tell: the Comptroller's own reform page still lists its implementation timeline as "coming soon."

4 · The state is capitalising incumbents, not replacing them. The Defense Department put $1 billion of its own money into L3Harris's rocket-motor spin-off. That is the opposite of disintermediation.

Where it is genuinely eroding, and we should be honest: Boeing beat Lockheed to the next-generation fighter in March 2025 — the loss of a generation-defining programme. Contracting through flexible "other transaction" authorities grew from $1.8 billion to over $18 billion in eight years, and the 2026 defence act now forces 14-day decisions on alternate suppliers. Seven of the twelve firms picked for Golden Dome interceptor prototypes are non-traditional. And most tellingly, the Pentagon's 2027 budget puts 49% of munitions funding into weapons costing under $600,000, rising to 70% by 2031.

The synthesis, and it is the sentence to remember: the disruption is happening at the effector-price tier, not the prime tier. Lockheed's moat is being repriced, not breached — and Lockheed is doing much of the repricing itself. In July 2026 it unveiled PAC-3 ACE, an interceptor at roughly half the price of its own MSE, and it built a $150,000 cruise missile — a tenth the cost of a JASSM — flying both variants within ten months. A company deliberately halving the price of its own best product is not a company in denial. I score the moat a 7: physically deep and genuinely hard to enter, marked down for the NGAD loss and for a demand curve bending toward cheaper effectors.

The sharpest single finding in this research is a story about a missile that worked and a contract Lockheed did not win.

In March 2025 Lockheed unveiled the CMMT — pronounced "comet" — a cruise missile designed to cost about $150,000 against the $1.5 million of a JASSM. A tenfold reduction. It flew both an unpowered and a powered variant within ten months of concept. The production philosophy was radical for a prime: "Our concept is to have a factory that fits inside of a room." Twenty-five of them fit on a pallet where nine JASSMs fit.

Then the Air Force created the programme these were built for — the Family of Affordable Mass Munitions, $12.6 billion for roughly 28,000 weapons at a target unit cost near $100,000. The primary contractors are Anduril, CoAspire and Zone 5. Lockheed's CMMT, per trade reporting, "remains separate from the formal FAAM program."

Lockheed fixed the cost engineering and did not fix the channel. The prime can now build cheap — it demonstrably can, in ten months. What it still cannot reliably do is win through the fast-acquisition pathways that the new entrants were designed around. And Lockheed's own annual report names the mechanism with striking candour: such an award "may be subject, in certain cases, to the condition that a significant portion of the work under the OTA is performed by a non-traditional defense contractor." The company is describing, in a regulatory filing, the rule by which it is being designed out of certain competitions.

Our read: the market is right that Lockheed is not being disrupted, and wrong about why. The 2026 Iran war settled the empirical question — when a real high-intensity conflict arrived, the United States burned through more than a thousand Tomahawks and eleven hundred cruise missiles in thirty-nine days, expended over half its pre-war inventory of four out of seven critical munitions, and responded by ordering Lockheed to triple and quadruple production. That is not an industry being disintermediated. But the same episode revealed the vulnerability: at a PAC-3 consumption rate of 225 a day against production of 1.7, the arithmetic of exquisite interceptors against cheap threats does not work, and the budget is already migrating to cheaper effectors. Lockheed's revenue will grow; its average selling price will fall. That is a repricing, and Lockheed is leading it — which is the correct response, and a margin headwind its own chief financial officer has flagged.

The standard bull case on Lockheed is that the F-35 is a decades-long, high-margin aftermarket annuity. The duration is real. The quality of it is not what investors assume, and this is the least-understood part of the company.

Three things spoil the annuity story. First, in November 2023 the Defense Department decided against a fixed-price performance-based logistics contract and stayed on annual cost-plus arrangements, citing data-quality problems. Cost-plus with a capped fee is the opposite of the pricing power the annuity thesis assumes — and note the structural irony: production is fixed-price incentive, where Lockheed keeps any underrun, while sustainment is cost-plus, where it does not. The common claim that sustainment carries higher margins than production appears to be exactly backwards, and we could find no company statement supporting it.

Second, the GAO found that the programme office and Lockheed "reconciled" performance metrics in 19 of 39 measurement periods — roughly half — to clear higher fee thresholds, while readiness rates "generally stagnated or worsened." The auditors' June 2026 recommendation is to restructure the incentives including penalties or elimination.

Third, the deadline. Section 142 of the 2022 defence authorisation act requires the transfer of all F-35 sustainment management to the services by 1 October 2027 — inside any reasonable forecast window. Lockheed still leads seven sustainment activities, and the transfer has slipped repeatedly because the government has never obtained the technical data rights it would need, an issue the GAO has flagged since 2014 across 43 recommendations. In practice Lockheed's position is sticky. But this is a live policy risk with a statutory date attached, and it is not in most models.

This table is why the two companies deserve different scores despite sharing a disease. Lockheed has retired 23% of its shares over the decade — and unlike Boeing, it did so from genuine surplus cash rather than borrowed money, never had to reverse it, and never stopped paying a dividend that has now risen for 23 consecutive years. Most tellingly, free cash flow per share ($37.92) comfortably exceeds earnings per share ($27.31) — the mark of a business converting profit into cash, and the exact inverse of Boeing, Eli Lilly and the AI hyperscalers we have covered.

Two honest marks against. The buyback has hollowed the balance sheet in accounting terms: equity of just $6.72 billion against $21.7 billion of debt, which is why return on equity reads a flattering 86% and price-to-book an alarming 14.7×. Neither number means much; the honest one is return on invested capital of 20.1%, which is genuinely good. And the repurchase pace has slowed sharply — from $7.9 billion in 2022 to $3.0 billion in 2025 — as cash went to the munitions build-out instead. That is the right priority, but it removes a support the shares have enjoyed for a decade. I score management a 7: excellent capital allocation and a serious strategic response, marked down because four reach-forward losses in three years is, in the end, an execution failure that no amount of buyback discipline offsets.

The numbers describe a high-quality business in a mild but genuine decline. On the quality side: free cash flow of $37.92 a share against earnings of $21.49 is superb conversion, return on invested capital of 20.1% is well above the cost of capital, the Piotroski score of 8 out of 9 is among the highest on our board, and 23 consecutive years of dividend increases is a record only a handful of American companies can match.

On the deterioration: operating margin has fallen from 14.3% to 10.3% since 2019, and earnings per share peaked in 2023 at $27.55 and are now $21.49 — down 22%. Revenue grew all the way through; profitability did not. Some of that is the charges, and stripping them out gives roughly 10% segment margins rather than 9% — better, but still well below where this company operated for a decade. The rest is mix: the munitions that are rescuing the growth carry different economics from the fighters they are replacing, and the chief financial officer has said the ramp dilutes margins near-term. The single number that captures the whole picture is that Lockheed's revenue is up 59% over the decade and its earnings per share is up 23% — real growth, diluted by margin erosion and only partly rescued by the buyback.

On the numbers alone, Lockheed is reasonably priced — arguably the most sensibly valued industrial we have looked at. Eighteen and a half times next year's earnings, under fifteen times free cash flow, a 2.44% dividend raised for 23 straight years, a discounted-cash-flow model at $831 (+49%), and a record backlog. Compare that with Walmart at 40× for 4% margins, or Costco at 47×. For a business earning 20% on invested capital with an eight-year order book, this is not an expensive stock.

But two things temper it. First, the move already happened. The shares are up 36% over twelve months against the S&P's 17% — and that run is largely a charge-cleanup trade (the stock bottomed near $412 after the 2025 losses) plus the munitions supercycle, rather than any improvement in the underlying fighter franchise. Wall Street sees only 10.8% of further upside, and the split of 21 buys against 15 holds is lukewarm for a company with a record backlog. Second, the estimates assume the deterioration stops. Consensus has earnings recovering from $21.49 to $30.36 next year — a 41% jump — which requires no further reach-forward losses on a classified programme the company says may produce more, and an F-35 delivery rate that recovers from the 51 aircraft booked in the first half of 2026 toward the ~156 a year management guides to. That gap is the single most load-bearing unresolved number in this analysis.

So the verdict is "Own the Arsenal — Watch the Aeroplane." I score valuation a 6: genuinely fair value for genuine quality, with no meaningful discount for a fighter franchise that is shrinking, a statutory sustainment deadline in October 2027, and a company that has warned it may take more charges. This is a fine long-term holding for an income-and-quality investor at today's price, and the dividend does real work while you wait. But the fat pitch was at $412 last year, when the charges had just landed and everyone was selling. Our accumulation zone is $490 — roughly 16× next year's earnings, a level that would put the yield near 2.8% and that this stock has traded below within the past twelve months.