'Who gets paid to rearm the West?'

Lockheed Martin just raised its 2026 cash forecast to $8.0bn. Rheinmetall also grew sales by 39% in the first half of 2026 but used $1.81bn of cash doing it. Which companies are best at turning defence orders into cash?

On 22 October 2026, Lockheed Martin reported a record order backlog of $292bn during its Q3 earnings call. Lockheed also raised its full year 2026 free cash flow forecast from $7-7.2bn to nearly $8bn.

Rheinmetall, Germany’s largest defence company, grew its sales by 39% in the first half of 2026, and its operating margin rose from 12.1% to 15.0%. But over the same six months, its operating free cash flow was -$1.81bn.

Rheinmetall says its customers paid their advances later than expected, therefore affecting the cash levels in the company. It had also built up inventory for deliveries due later in the year and kept spending on new plants. And on top of that, customers owed it even more money at the end of June after heavy invoicing late in the quarter.

So Rheinmetall is becoming more profitable but its cashflow statement goes the other way.

Both companies sell to governments that are in the race to rearm against future conflicts. 

Lockheed already produces excess cash today. But Rheinmetall is spending today’s cash on growth it hopes to be paid for later. Two defence companies have two entirely different ways of using their heavy order book to future-proof their businesses. 

In February 2026, in Impactfull Weekly #27, we argued that the second leg of the defence trade would reward suppliers whose production lines were already running. Our reasoning was that the countries behind on defence spending couldn’t build what they needed fast enough at home. But the question now is which of those suppliers are capable of turning them into cash returns.

A bigger defence budget pays for more equipment, higher wages and new factories. Only part of that money turns into profit for a supplier, and given defence is a capex-heavy industry, part of today’s profit then has to pay for tomorrow’s factory.

Our view is that the next stage of the defence trade will reward the companies that turn approved production capacity into profitable deliveries. Approved capacity means production the military customer has already tested and signed off on. 

So over the next 3 to 18 months, we expect that money to go to the suppliers of propulsion and sensors, and to the companies that connect and maintain military equipment. For each of them, the billion-dollar question is whether they can produce more while still earning a good return on the money they invest.

In this Impactfull Weekly, we’ll show you why defence orders are growing faster than deliveries, which parts of the supply chain get paid as armies rebuild, and companies where you can see this in their future earnings.

Part 1: Big order books don’t make returns 

Since the summer, two worries have pulled most defence shares down.

In Europe, growth in total defence spending is expected to slow from 20% in 2025 to 9% in 2026 (after inflation). In Washington, Congress has already trimmed President Trump’s $1.5tn budget request for 2027, and the midterm elections on 3 November could cut it even more. 

(source: European Defence Agency)

From 14 August to 21 October, the iShares US Aerospace & Defense ETF fell 18.7%, and by 21 October Rheinmetall’s shares were down 34.7% for 2026. 

Lockheed’s shares were up 14.0% for 2026 over the same period. Although Rheinmetall mainly sells to EU governments, and Lockheed sells mainly to the US government, the main difference between them is how quickly their orders turn into cash, and that’s what this part is all about.

Equipment budgets aren’t completely frozen

The US Congress still hasn’t passed a defence budget for 2027. 

So since 1 October, the Pentagon has been running on a continuing resolution, a “stopgap law” that keeps the government running at the previous year’s budget level until 11 December 2026. 

In practice, that means the Pentagon can keep paying for the programmes it already funded in 2026. What it can’t do is use that money to start producing anything new, or to raise production rates above 2026 levels. If you supply a programme that was already running, your money keeps coming, but any newly announced programs will be waiting for the U.S. Congress to pass the budget. 

On the other hand, what we're seeing in Europe is the slowdown from 20% to 9% of spending growth. That number covers the whole defence budget, and most of it goes on salaries and operational costs. 

To see where and what equipment makers actually get paid, you need to look at a smaller subsection of it. The European Defence Agency (EDA), the EU body that collects member states’ defence data, calls this part defence investment. It’s the money governments spend on buying equipment (procurement) and on research and development (R&D). 

Defence investment grew 23% in 2025, to $150.1bn (€134bn), and the EDA expects it to grow another 23% in 2026, to $182.6bn (€163bn), after inflation.

(source: European Defence Agency)

That means Europe’s slowdown is coming from the rest of the budget, which pays for personnel and day-to-day operations. According to the EDA’s figures, that part will only grow by 2.5% in 2026, while the equipment budget grows just as fast as it did in 2025.

The money for equipment hasn’t dried up everywhere. It’s still growing quickly in Europe, and in the US it’s held at 2026 levels until Congress votes for a new budget. 

That’s why we don’t think demand is what’s holding defence companies back or is the reason for the recent sell-off. What matters now is how quickly that money is turning into delivered equipment, and then into cash.

But the money is getting stuck somewhere

A procurement budget (the money a government sets aside to buy equipment) has a long way to go before any of it reaches you.

Parliament has to approve it, the defence ministry has to turn it into contracts and pay the companies, and each company then has to decide how much of that cash to pass on to you as a shareholder. Along the way, it tends to get stuck in three places.

• A budget that hasn’t been signed: 

Lockheed is already feeling this. In its Q3 results on 22 October 2026, Lockheed said its costs on early work for the next batch of F-35s (Lot 20) were running ahead of the money the US government has approved so far. 

It also said its results will suffer unless more funding arrives. Germany has a similar date coming up. On 9 December 2026, its parliament debates Arminius, an order for around 1,800 Boxer armoured vehicles worth $28.0bn (€25bn), of which Rheinmetall’s share will be $13.9bn (€12.4bn). Until those votes happen, the work is “stuck”.

• An order that’s been booked but not yet paid for:

This is where Rheinmetall sits today: the cash it used in H1 2026 went into filling the gap of customer advances that arrived late, into inventory and into new plants. Renk, which makes the transmission for the Leopard 2 tank, is in the same position, but on a smaller scale. Its orders grew 29.7% in H1 2026, and its inventory rose by $79.5m (€71m) to $568m (€507m) as it built gearboxes it hasn’t delivered yet. 

Leonardo, meanwhile, needs to bring in $1.81bn (€1.62bn) of free operating cash flow in the H2 2026 to hit its full-year target. In each case, the cash arrives when the equipment is delivered and the customer pays. The next set of results will show whether that’s happening.

• Customers paying in advance can boost cashflow numbers temporarily:

When a government pays in advance, the company’s cash goes up before it has built anything. 

For example, BAE Systems generated $2.39bn (£1.79bn) of free cash flow in the first half of 2026, after negative cash flow a year earlier, most of it coming from customer advances. 

That cash is definitely real, but it comes with an order book of aircraft, ships & submarines that still need to be produced and delivered, so you can’t really treat it as excess cash to return to shareholders.

And even when the cash does arrive, there’s one more step before it reaches you. 

In January 2026, President Trump signed Executive Order 14372. It lets the Pentagon act against defence contractors that deliver late or underinvest in production while paying out buybacks or dividends.

The Senate’s version of the 2027 defence bill goes one step further. Its section 815 would bar any company that signs a Pentagon contract from 15 June 2027 onwards from buying back shares or paying dividends, unless the Pentagon approves the payout as part of a plan to invest in things like new production capacity. Congress will be settling on the final wording of the bill come December.

But for now, companies are all making different choices. Lockheed did zero stock buybacks in the first nine months of 2026, while L3Harris bought back $525m in H1 2026.

So before you buy a defence company for its order book, it’s worth knowing which of these three places its money is stuck in, and when it’s likely to come out, if it comes out.

Part 2: War is moving to the air

Where a company’s money gets stuck depends partly on what it sells. 

An armoured vehicle order like Arminius, once it’s approved, takes years to build, deliver and pay for. Interceptors (the missiles that shoot down incoming missiles and drones) work differently, because armies fire them and then have to buy them again. 

Since the war with Iran began on 28 February 2026, the US has used about 65% of its Patriot interceptors, according to estimates by CSIS, a Washington think tank. Ukraine, meanwhile, plans to build 7 million drones this year, according to its deputy defence minister. 

Both point the same way: more of the fighting is now happening in the air, and the equipment that finds, jams and shoots down what’s flying is the equipment NATO governments will keep reordering.

The budgets point the same way too.

From the $26.5bn the Pentagon has asked for to buy Patriot and THAAD interceptors in 2027, only $3.4bn sits in the normal budget. The rest depends on a separate bill for which the Pentagon asked $350bn, and which Congress has so far cut to a paltry $60bn. We expect US spending on missiles and munitions to land at $40-50bn in 2027, well below the $70.5bn requested but still nearly double 2026’s $24.4bn. 

In Europe, the EU has already lent $168bn (€150bn) to 19 countries for defence, and among that, a third of what they asked for was for air and missile defence, missiles and ammunition.

The question is how fast the factories can deliver, and how much of it reaches companies you can actually buy.

Why the air

Air defence spending keeps rising because of cost. A Shahed drone, the type Iran and Russia launch in waves, costs about $50,000 to build, according to leaked documents from Russia’s main Shahed factory. 

A Patriot interceptor costs about $4.4m, based on the Pentagon’s 2027 request for 3,203 of them at $14bn. So when a Patriot shoots down a Shahed, the defender spends nearly 90x what the attackers did.

No army can keep that up for long, and the US learned its lesson trying to shoot down a fly with a bazooka.

That fundamentally changes what governments buy. Instead of one expensive missile for every threat, armies are building layers of defence. 

Radars and other sensors spot what’s coming and identify it. Jammers knock drones off course without firing anything. Guns such as Rheinmetall’s Skynex fire bursts of shells, and interceptor drones such as Ukraine’s Sting cost only $2,100 each. For context, Ukraine was churning out more than 10,000 Stings a month by March 2026.

So, if a defence can’t tell the difference between a $50,000 drone from a ballistic missile, it will end up firing its most expensive weapons at anything in sight.

The same drones are changing the calculation on the ground. In February 2025, RUSI, a British defence think tank, found that drones had caused 60-70% of the damage to Russian military equipment in the third year of the war. This is one of the main reasons why land based military equipment like tanks & other armoured vehicles get deprioritised when comparing costs.

Looking ahead, the limit is shifting from budget to parts availability. 

Every Patriot interceptor needs a radar seeker (the sensor in its nose that finds the target), and only Boeing makes them. Boeing built about 650 to 700 in 2025, and on 5 October it signed a $14.7bn, seven-year deal with the Pentagon to triple that. The rocket motors take about 2.5 years to build, so interceptors ordered this year won’t arrive before mid-2028. 

One level below these parts sits China, which refines 91% of the world’s rare earths and makes 94% of its permanent magnets, the parts that steer these interceptors and drones. 

Its pause on tighter export rules for the US ends on 10th and 27th November unless Beijing extends it. That’s why the new budget is moving towards more cost-effective options that need fewer of these parts. 

(source: S&P Global)

Israel took delivery of the first operational laser air defence, Rafael’s Iron Beam, in December 2025, and the UK plans to fit its DragonFire laser (costs about $13 per laser shot) to Royal Navy destroyers by 2027. 

Over the next three years, we expect seekers, rocket motors and the magnets and chips inside them to be the scarcest parts in air defence, and the companies that make them, or have locked in supply, to get paid the most. 

See it

Every layer of air defence starts with a sensor. 

Before a jammer, gun or interceptor can do anything to a Shahed drone, something has to spot the threat and work out what it is. 

In September 2026, the US Army tested this at its Scarlet Dragon exercise, flying swarms of up to 100 drones at its own defences. Its stated conclusion was that no single sensor is capable enough, and they need several sensors linked together to track a drone swarm. That means several kinds of sensors at every site, and more software to join them up

In plain terms, the Army is saying it will need many more sensors, and one big radar per site will give way to clusters of small radars, cameras and listening devices, all tied together by software.

The orders are already coming. In September, the Pentagon’s new counter-drone task force awarded ten contracts for compact radars, cameras and gun systems, worth up to $4.15bn between them, and the Army wants to spend $994m on small counter-drone equipment in 2027, up from $596m in 2026. 

Unlike most interceptor money, all of this sits in the normal budget, so it doesn’t depend on the bill Congress has been cutting. Expect many small, repeat orders rather than a few large ones, going to companies that can build sensors in volume and connect them to what armies already own.

Europe has reached the same conclusion, but it is about a year behind the US on orders. 

The US set up a single buyer for counter-drone equipment, JIATF 401, in August 2025, and by September 2026 it had already awarded ten contracts for small radars, cameras and gun systems. 

An L3Harris VAMPIRE system showing an electro-optical/infrared (EO/IR) sensor ball & laser-guided rockets:

(source: L3Harris)

The European Commission only published its equivalent plan in February 2026, calling for drone defences built from several kinds of sensors joined by AI software, and its EU-wide drone defence network isn’t due to be switched on even partially until the end of 2026.

Europe is now playing catch up in three ways. 

The first is money: EU countries asked for about $6.7bn (€6bn) of their SAFE loans for drones and counter-drone systems, and some of it is already turning into contracts, such as Romania’s Skynex guns from Rheinmetall and its 12 Ground Master radars from Thales. 

The second is factories: Rheinmetall plans to raise output of its Skyranger, a vehicle that carries a radar, cameras and a gun together, from 70 to 100 turrets a year to as many as 400. Germany has ordered 19 so far, and reports put a follow-on order of 500 to 600 at up to $9bn (€8bn), not yet confirmed. 

The third is with law: In early 2026, Germany gave its armed forces the power to shoot down drones over its own territory. That points to where European orders go next. Fewer separate parts than the US buys, and more complete systems from the companies that build them, with the largest contracts still to be signed in 2027.

Jam it & connect it

Jamming has historically been the cheapest way to stop a drone. 

Most drones are flown by radio, and a jammer floods those frequencies with noise until the pilot loses control. Nothing is fired, so each use costs almost nothing. In 2023, RUSI estimated that Ukraine was losing around 10,000 drones a month, mostly to Russian jamming.

But the attackers have adapted, and fast. 

Fibre-optic drones are steered through a thin cable unspooled behind them, so there’s no radio signal to jam. In April 2026, Ukraine’s deputy head of the Presidential Office said they made up 32% of the strike drones Ukraine uses each day. By September, the commander of Ukraine’s National Guard said 70% of his force’s drones were fibre-optic, up from about 20%. Ukraine’s defence ministry received 374,000 of them in 2025 and had already received more than 92% of that number by April 2026, nearly three times the previous year’s pace.

A fibre-optic drone also gives off no radio signal, so the only ways to find it are to see it, on radar or camera, and the only ways to stop it are to hit it, with a gun, an interceptor drone or a laser. 

So the spread of these drones moves the US & EU procurement spending away from jamming and towards the sensors in the previous section. Hensoldt, for example, signed a framework in December 2025 to supply its SPEXER radars for Rheinmetall’s Skyranger, with potential volume in the high hundreds. Jamming won’t disappear, because most drones are still radio-controlled. But it’s becoming one layer among several rather than the one stop answer.

The emergent growth driver now is the need to connect all these layers. 

A radar that spots a drone has to pass its position to the gun or interceptor fast enough for one of them to act, over links the enemy is trying to jam. That’s why the European Commission’s February plan calls for European-made command software. The radios and networks underneath are a quiet, profitable business. 

Picture of Hensoldt’s SPEXER 2000 radar:

(source: Hensoldt)

There’s also almost no pure way to own this layer: electronic warfare and military communications sit inside defence giants like L3Harris, Thales, Hensoldt and Leonardo, so you buy them as part of a larger company.

Stop it

Stopping what’s in the air now splits into two markets. 

One is the expensive interceptor fired at ballistic missiles and jets, where a single missile costs millions. The second is the “cheap stuff” fired at drones: guns, small missiles, interceptor drones and, soon, lasers. The money, the buyers and the timing have little to do with each other.

The interceptor market is where the biggest contracts are. 

In 2026, the Pentagon signed Lockheed Martin to seven-year deals worth up to $58.6bn for Patriot interceptors and up to $35bn for THAAD. Lockheed’s missile division grew sales 24% in Q3 2026, and the company expects about 20% growth for the year. 

But look at how those contracts are written. On its 22 October results call, Lockheed said the profit terms still aren’t agreed, and that the Pentagon is keeping the deals cash-neutral while Lockheed builds new factories. In other words, Lockheed won’t make much cash from them until deliveries ramp up, and most of the 2027 money depends on the bill Congress has cut. 

The constraint now lies in the parts, and the companies that make them are getting contracts of their own. To feed the Patriot ramp up, Boeing has a $14.7bn deal to triple its seeker output, L3Harris a $4.7bn deal for the rocket motors, and Northrop Grumman $1.6bn for propulsion and other components. From these contracts, L3Harris is the one where rocket motors form a meaningful share of the revenue, and its new plants in Arkansas are due to open in 2027.

The “cheap stuff” market moves faster, and Europe is buying into it now. 

In January 2026, Kongsberg won a contract worth about $1.68bn (NOK 16bn) from Poland for 18 counter-drone batteries combining guns, guided missiles and interceptor drones. 

Germany is doing the same with Rheinmetall’s Skyranger. These systems are cheaper, quicker to build and less tied to US budget politics. The catch is that the market already knows. Kongsberg’s shares trade at nearly 39x operating profit, and much of the rest of this market belongs to companies you can’t buy, such as MBDA and Diehl because they are not listed equities.

Feed it

Every layer above needs ammunition. 

Guns fire shells, interceptors carry explosive warheads, and rocket motors burn solid propellant. All of it depends on a handful of chemicals, and this is where the shortage is most physical. 

The key one is nitrocellulose, the base of the propellant that fires artillery shells. In 2025, the European Policy Centre (EPC) put Europe’s entire output at 4,500 to 10,000 tonnes/year, against a need of about 20,000 tonnes including Ukraine.

The shortage shows up in the shell count. The US Army aimed to make 100,000 artillery shells a month by October 2025. By March 2026, it was only making 36,000, according to the Pentagon’s Inspector General, which blamed one contractor’s failure to make shell casings to specification. 

The US Army now hopes to reach 140,000 a month by December 2027. Europe has pledged more than a million shells a year by 2027, and the EPC has warned it can’t get there without more nitrocellulose.

This is the part of defence where demand is most certain, and also the hardest to buy. 

Most of the companies making explosives and propellant are state-owned, like Poland’s PGZ and France’s Eurenco, or part of much bigger groups. The one clean listed exception we’ve found is Colt CZ, the Czech arms maker. 

In 2025, it bought control of Synthesia, a Czech nitrocellulose producer that can make about 6,000 tonnes a year. In H1 2026, that business earned an operating margin of 54.5% before depreciation. Colt CZ raised its 2026 forecast in September and is spending about $45m (€40m) to lift Synthesia’s capacity to 7,000 tonnes.

Elsewhere, the shortage is mostly turning into costs before it turns into profit.

Britain’s Chemring is building new explosives plants in Norway and Scotland, spending about $147m (£110m) a year, more than the cash its business generates. Earlier in the essay we also spoke about Rheinmetall and their cashflow hitting a snag in H1 2026 due to a double whammy of late advance payments and investing into capacity expansion.

Feeding the war is a real business, but for most companies the cash comes very late.

Casualties of the air warfare

If more of the fighting happens in the air, less of the money goes to what sits on the ground. 

The US Army already acted on that. 

In May 2025, it stopped buying Humvees, its JLTV light trucks and the M10 Booker light tank, and began retiring older Apache attack helicopters, moving the money into drones and counter-drone equipment for its frontline units. 

Its 2027 request goes further: Apache purchases fall from $362m to $1.5m, and Black Hawk helicopters from $913m to $39m. The one large new armoured vehicle, the XM30, gets $547m for 19 vehicles.

Europe hasn’t cut land spending completely, but is choosing to delay it. 

Germany still plans to buy large numbers of Boxer armoured vehicles and Leopard tanks, but in September Bloomberg reported that it was approving small batches while the large orders waited. 

The biggest, the Arminius Boxer order from earlier in the essay, isn’t due before parliament until 9 December. South Korea, which sold more tanks to Europe than anyone in 2022-23, is suffering from the same rotation. 

Romania’s order for 216 K2 tanks, worth about $7bn, is still unsigned. The share prices have moved first: by October, Rheinmetall was down 35% this year, the German tank-gearbox maker Renk was down 60% from its high, and Hyundai Rotem, which builds the K2, was down 63%.

The sea tells a more mixed story. 

Surface ships turn out to be vulnerable to cheap drones too: Ukraine’s defence ministry says about 30% of Russia’s Black Sea Fleet has been sunk or seriously damaged, much of it by unmanned boats. 

But naval money isn’t shrinking, at least not in the US. The Pentagon’s 2027 request more than doubles shipbuilding, from $27.2bn to $65.8bn, and almost all of it sits in the normal budget. What’s changing is where the money goes: under water, into submarines, and into air defence for the ships that remain, such as Patriot interceptors on US destroyers and the DragonFire laser on British ones.

So land is the clearest casualty, and it’s a casualty of timing more than of demand. The orders are slower, and the share prices have fallen further than the order books. 

That gap is worth watching. Renk, for example, grew its order intake 29.7% in H1 2026 as its shares fell. So if the large European land orders are signed in 2027, companies like it are bound to recover from this pessimism.

Companies to watch

Link to artifact

Colt CZ (PSE: CZG)

Colt CZ is the cleanest listed way to own Europe’s propellant shortage. 

Its Synthesia unit makes about 6,000 tonnes of nitrocellulose a year, the base of the propellant in every artillery shell, and earned a 54.5% margin before depreciation in H1 2026. 

Europe needs about 20,000 tonnes, including Ukraine’s needs, and makes at most 10,000. Colt CZ raised its 2026 forecast in September and is adding 1,000 tonnes. The risk is net debt, which more than doubled in Q1.

L3Harris (NYSE: LHX)

L3Harris gives you exposure to the slowest part of the Patriot build: the rocket motor.

In September, Lockheed gave it a $4.7bn, seven-year contract for Patriot’s propulsion, and its new Arkansas plants open in 2027. Its radio and jamming division earned a 26.9% margin in Q2.

It trades at 12.8x EV/EBITA, the cheapest US prime, after its shares fell 36% since March on the delayed listing of its missile business and the Senate’s proposed buyback ban. But it carries the highest debt of the five primes, and if December’s final defence bill keeps the ban, its $500m-a-year buyback is at risk.

Thales (EPA: HO)

Thales is the sensors pick. 

Its Ground Master radars sit at the base of many European air defence networks, and Romania has just ordered 12 with SAFE money. Defence orders rose 14% in the first nine months of 2026, and defence sales grew 13% organically. 

Thales also converts more of its profit into cash than any European peer, 1.43x its EBITDA, and holds net cash. At about 15x EV/EBITA, the price still reflects the slower group, not the defence business. The key metric to watch is if its defence book-to-bill falls below 1.

Our take

We think the money from rearming the West reaches shareholders in three stages, with each stage reaching a different point in time.

The first leg of the trade got paid in 2025, when European defence shares re-rated on the promise of bigger budgets. Thales rose 65.7% that year, and the market has already priced the obvious names: Kongsberg trades at nearly 39x operating profit and Hensoldt at about 60x its earnings.

The companies that own the scarce parts, and sell them for money that’s already approved, are the ones being paid now. 

Colt CZ’s nitrocellulose business earned a 54.5% margin in H1 2026, Thales turns 1.43x its EBITDA into cash, and L3Harris has a seven-year contract for the rocket motors every Patriot needs. New capacity only arrives from 2027, when L3Harris’s Arkansas plants open and Chemring’s new explosives plants start producing. Until then, we think these companies will keep their pricing power.

The large US interceptor contracts and Europe’s big land orders will be the last ones to provide returns. Lockheed’s Patriot and THAAD deals stay cash-neutral until deliveries ramp, and interceptors ordered this year only arrive from mid-2028. Most of the 2027 money waits on a reconciliation bill Congress has cut from $350bn to $60bn, and Arminius only goes before Germany’s parliament on 9 December.

So for this rearmament of the West, the cash today sits with whoever owns the scarce parts, and it moves to the primes and the land names from 2027.

https://www.marketscreener.com/news/who-gets-paid-to-rearm-the-west-ce785cdfde8cf026