Newsletter · · Ashutosh Agarwal
The Defense Money Has Arrived and Now the Fight Is Over Who Keeps It - Aerospace & Defense Weekly - Week of August 22, 2026
How the aerospace and defense podcasts pivoted from whether the money would arrive to who captures and keeps it, covering the surge past $100 billion a year in defense venture capital, an unexpected aerospace IPO window, robot fighter jets carrying live missiles, and a right-to-repair IP fight that could undo the boom, for the week of August 22, 2026.
Aerospace & Defense Weekly
Week of August 22, 2026: The Defense Money Has Arrived and Now the Fight Is Over Who Keeps It
Week ending Saturday, August 22, 2026
For the last two years the running argument in this sector was whether the money would actually show up, whether all the talk of rearmament, drone wars, and empty missile bins would turn into real budgets and real contracts. That argument is over. The money is here, in almost embarrassing quantities. What the podcasts spent these last two weeks on was the next question, which is harder and more interesting: now that the cash has arrived, who captures it, who keeps it, and can any of it be turned into actual hardware fast enough to matter.
You could hear the shift everywhere. A defense-focused investor laid out, in unusual plain terms, why venture money has gone from a rounding error to over $100 billion a year flowing into defense. An aerospace banker described an IPO window that simply did not exist a few years ago. Air Force operators described loading live missiles onto robot fighter jets they no longer "fly" so much as "direct." A drone-swarm software founder explained why he took his company public before it turned five. And underneath all of it, two policy fights, one over ammunition production, one over who owns a contractor's intellectual property, that will decide whether this boom converts into capability or just into higher stock prices.
One habit worth keeping, because it changes how much weight each claim deserves. We flag whether a speaker is an operator, someone actually building the thing, running the company, or buying it for the government, or a pundit, analyst, or investor commenting from the outside. A founder telling you his revenue is real is different evidence than an analyst telling you a stock looks cheap. This fortnight was unusually rich in operators, which is exactly why it was worth reading closely.
1. The Clearest Map Yet of Why Every Investor Is Suddenly Piling Into Defense, and the Trap Sitting Underneath It
If you want to understand the money flooding this sector, the single most useful conversation of the fortnight was on How I Invest with David Weisburd (August 12), where Jake Chapman of Marque Ventures, a defense-focused venture investor, so a pundit with real skin in the game, walked through the numbers behind the frenzy.
Start with the scale of the shift. "If you go back to 2019, 2020," Chapman said, "there were single-digit billions in venture or private equity, early-stage growth dollars going into defense. And now it's well over $100 billion a year." The market underneath it has swelled too: the Pentagon's budget was "around $800 billion" when he started investing in 2019, "over $900 billion" last year, and "this year, they're talking about a $1.5 trillion budget", which, he noted, would still be only about 3.5–4% of the economy, against 6–7% late in the Cold War. Add allied budgets, intelligence, homeland security, and classified "black budgets," and he pegs the whole defense ecosystem at "probably fairly well over a $2 trillion a year market. Not a lot of markets of that size."
Why now? Three reasons, in his telling. First, the world genuinely looks more dangerous, U.S.–China competition, Ukraine, the Strait of Hormuz fight with Iran. Second, and this is the part investors underrate, three companies broke the taboo: "Palantir, SpaceX and Anduril really proved" defense could be a successful investment sector, and now "there's a mafia developing" of ex-SpaceX, ex-Palantir, and soon ex-Anduril engineers founding the next wave. Third, Washington is deliberately making itself an easier customer, aligning research money with actual purchasing, and, "just yesterday," eliminating a cybersecurity rule called CMMC that "imposed hundreds of thousands of dollars in costs on startups" and "made it almost impossible for young companies to do business with the government."
Chapman also gave the best short explanation anyone has offered for why the private valuations look so high. Big funds have a math problem: "if you are T. Rowe Price... you don't have a lot of options at that scale where you have to write a $100 million or $250 million check." There are more giant dollars chasing defense than there are companies large enough to absorb them, "and so valuations shoot up." That is a demand-for-deals story as much as a quality-of-company story, worth remembering the next time an Anduril markup makes headlines.
And he told the origin story of the whole prime-contractor structure, which is really the story of why the incumbents look the way they do. After the Cold War, with budgets falling, the Pentagon summoned defense CEOs to a dinner now known as "the Last Supper" and told them, in Chapman's paraphrase, "look to your left, look to your right. One of you is going to be out of business." The result was the consolidation into "a half dozen major defense primes", Lockheed, Raytheon, Northrop Grumman, General Dynamics. His sharp point: the bureaucracy created the oligopoly, not the other way around. Surviving multi-year budget cycles and mountains of red tape forces you to become a company with "a small subset of engineers... but giant teams of lobbyists and lawyers." As he put it, the primes "become financial firms, basically."
Why it matters: this is the bull case stated with unusual honesty, and the caveat is baked right into it. The demand is real, the budget trajectory is up (Chapman argues even a smaller budget just shifts money toward the new, asymmetric stuff, away from aircraft carriers and toward things like Anduril's Barracuda cruise missile), and the flywheel of talent and capital is spinning. But the sky-high private valuations are partly a symptom of too much money chasing too few scaled companies, a classic setup for disappointment if the government's actual purchasing does not catch up to the private investment. Chapman said it himself: "remains to be seen if that investment is met with actual acquisitions from the government." That sentence is the entire risk, in eleven words.
2. The IPO Window Nobody in Aerospace Expected, and Why "Boring" Is Suddenly the Hottest Word in the Sector
For decades, aerospace was an M&A-and-debt business, not an IPO business. That has flipped, and the reversal is a genuine capital-markets event. On The Aerospace Executive Podcast (August 13), Jefferies banker Nick Fazioli, an operator on the deal side, explained what changed.
First, the universe of public aerospace companies had quietly shriveled: "over the last 15, 20 years, there were 40% fewer public aerospace related public companies than prior because of mergers, take privates, LBOs." Then two things happened at once. Companies got big enough to go public on their own, and, in a twist, investors soured on software. "Now you got this doubling of all of a sudden software and tech and things that are AI risks from an existential-risk standpoint," Fazioli said. "And now what used to be boring industrial manufacturing production companies" look "very AI proof. And in some ways AI positive." The industrial IPO backlog, he said, "is huge... at times that's like half our business now," in a business where "if you were an aerospace banker, IPO was not a product we were looking at."
The poster child is VSE Corporation. When CEO John Cuomo took over, "the market cap was something to the effect of like 400 million, 450 million... involved in five unrelated businesses", today it is a "$6.3 billion" company. Cuomo shed everything that did not fit and doubled down on aerospace aftermarket, and the access to capital that followed is the tell: "for that PAG deal, we raised a billion dollars of equity for him in a day. When he took over, his whole market cap was less than half of that." (That PAG business, Fazioli noted almost in passing, was itself a "$150 million EBITDA" operation by the time it changed hands.)
The most sought-after corner of all is component MRO, maintenance, repair, and overhaul of individual parts, because it looks like the holy grail: "a little bit higher margin, lower CapEx... less of a heavy workforce." He is fielding constant calls: "there's more money out there chasing component MRO than there are deals." Standard Aero (taken public by Carlisle) actually ran a full sale process, "looked at the IPO valuations and just said, this is... the better move." Arxis "went public yesterday." Fazioli's summary of the exit market: "It's full steam ahead."
He was candid about the vertigo, too. This has been "basically a 16-year bull run," with roughly "a two to two and a half turn multiple expansion on most deals" since before the pandemic. What could break it? "The obvious thing... is if things really deteriorate", the Iran conflict feeding high fuel prices, squeezed consumers, sticky interest rates. "Everything kind of flows down from the airlines," he warned, "ultimately to the supply and MRO base."
The contrast to all this growth-stock enthusiasm came on Investing Experts (August 17), where analyst Steven Cress (a pundit) made the case for the old guard as the defensive play. In a jittery, "risk-on, risk-off" market, he likes a "barbell", pairing growth names with income names, and put Lockheed Martin on the income side: a "2.3%" forward dividend yield, "significantly higher than the S&P 500" and above even the Vanguard High Yield ETF at "about 2%," with "a lot of analysts revising their numbers up" on the back of low global munitions stockpiles and multiple wars. RTX, he noted, was up "27%" in his income-oriented portfolio.
Why it matters: the capital story has two distinct doors, and money is pouring through both. The venture door (Section 1) is buying growth and disruption at nosebleed prices. The public-markets door is doing something subtler, re-rating unglamorous industrial and aftermarket businesses precisely because they are seen as safe from the AI wrecking ball, and treating a prime like Lockheed as a bond substitute with a defense tailwind. Both doors depend on the same thing continuing: that the demand narrative holds and the 16-year bull run does not meet its "black swan." Fazioli named the swan directly, a consumer-and-fuel shock flowing up from the airlines. That is the number to watch even for investors who only care about defense hardware.
3. The Robot Wingmen Start Carrying Live Missiles
The single most concrete sign that autonomous air combat is leaving the slide deck came on The Aerospace Advantage (August 15), where Air Force officers from the Experimental Operations Unit, the operators actually standing up the program, described flying the YFQ-44 Collaborative Combat Aircraft, the drone "wingmen" meant to fly alongside crewed fighters.
The details are the story. Just this week at Creech Air Force Base, one of the officers said, he got "to load, help the maintenance and weapons teams load AMRAMs on the aircraft", that is the AIM-120 air-to-air missile, and then, on one flight, "was the lead controller for one of the aircraft on its way out to the airspace." Note the word: controller, not pilot. And that is the crux. "It is growing in the amount of autonomy that the aircraft does on its own," he said, and everyone will "need to get comfortable with the fact that it is semi-autonomous... the humans are not the pilots. We are not the ones flying the aircraft. It's the algorithms that humans have been a part of designing and developing... the human piece is just controlling." The near-term goal is one person "tasking and controlling more than one aircraft."
The training math is what makes this an industrial story, not just a technology one. Teaching a human pilot is painfully serial, the officer, a former pilot-training commander, described "two to three years" to take someone from a T-6 trainer to combat readiness, one airman at a time. Software does not work that way: "we are teaching 1,000 CCA how to do that right now simultaneously," refining behavior across millions of synthetic repetitions the way a game-playing AI trains itself. The Air Force has also, tellingly, dropped the term "loyal wingman", a deliberate signal that it wants to invent new tactics rather than treat these as robotic tag-alongs.
And the officers connected it straight back to the ammunition question that has dominated this sector all summer: with these aircraft, "if we have the munition stocks and we build up these munition stocks, we're going to have the ability to flood the battle space with weapons. So we can really get the mass that we need."
Why it matters: this is the "cheap mass" thesis moving from argument to live-fire reality, real missiles, real aircraft, real operators in the loop. For the companies building CCAs (General Atomics and Anduril won the first round) and the software agents that fly them, this is validation. But listen to the operators' own dependency: the whole concept only delivers "mass" if the munition stocks exist. Autonomous jets that can flood the sky are worthless without missiles to put on them, which is exactly why the production fight in Section 5 is not a separate story. It is the same story.
4. The Land Grab That Actually Matters Is in Software: The Swarm Operating System
While the hardware gets the headlines, the operators building autonomy argued this fortnight that the durable value is in the software layer coordinating thousands of cheap machines. The sharpest version came on The Drone Ultimatum (August 21), with Alex Fink, founder and CEO of the drone-swarm software company Swarmer (an operator).
His framing is a clean analogy to computing history. "I liken this to the 1981 PC market," he said. "You go from five big manufacturers, vertically integrated, everybody makes their own processor... to 200 different companies make roughly identical PCs and you need one operating system." The drone industry, he argues, is fragmenting the same way: in Ukraine today there are "more than 500 different drone vendors," "more than 250 different UGV" (unmanned ground vehicle) vendors, and "about 2,000 hardware manufacturers." When hardware commoditizes, "margins get compressed. The software is where it's at."
The volume numbers are staggering and explain why humans-in-the-cockpit does not scale. "The industry used to make something like 30,000 drones a year for the entire industry worldwide. And this year, Ukraine by itself is making 7 million drones." At that scale, he said, "you don't have enough pilots. You're not going to have enough pilots. So the whole point is, this has to work without pilots." Swarmer's software keeps drones coordinating even when jammed, "point-to-point multicast" between machines that "works even if the packet loss is about 90%."
The business model is the genuinely novel part, and it is worth understanding because it is a bet on how this whole market gets structured. Swarmer does not sell drones; it licenses its software to the drone makers, "a per unit license fee for the drone maker when they make their own. It's almost like Dell has to pay Microsoft for a copy of Windows," plus "every year... a maintenance fee." Fink was refreshingly honest that this is a wager: "we're making a bet on a certain restructuring of the industry... Maybe the market will prove us wrong." He has also done something rare, taken the company public young. He recalled telling a startup accelerator, "we already filed the S1, the IPO's next month." Being public, he said, "allows us to move faster... we can raise capital faster." The structure is unusual too: "about two thirds of the team is in Ukraine," Poland second, Austin the headquarters and third, and because the IP sits in Europe, "we're exporting out of Europe into the U.S.," sidestepping the U.S. arms-export rules that hobble American-based rivals.
Fink also name-checked where the real, working competition is: on interceptors, he pointed to Merops, an Eric Schmidt–backed company, as "the only one that is working at decent volumes and hitting shaheds autonomously right now", everything else "is either piloted or single-unit autonomy." His logic on defeating swarms is grimly simple: "the only way to take a swarm is with a swarm of interceptors. There's nothing else you can do."
The same "coordination is the product" theme ran through Cogs of War (August 19), where three operator-veterans tackled the least glamorous but most decisive problem, logistics in a war zone the enemy can see. Paul Wynn, CEO of Havoc AI, described building "collaborative autonomy" so "thousands of unmanned systems... work together": in two and a half years his company has "built 100 autonomous vessels" and acquired both a UAV and a ground-vehicle company. David Tuttle of Ruin Technologies builds the software that predicts what forward units need before they run out. And Dr. Tom Garvey of CACI played the systems integrator, the company that stitches everyone's kit together. The stakes, Wynn said, are basic and brutal: dispersed troops across "100 islands" in the Pacific "still have to eat" and still need "food, bullets, blood," and helicopters and crewed ships are "the first thing they're going to shoot out." The answer is truly autonomous resupply that does not need to phone home to a command ship that can be jammed.
Why it matters: if the hardware really is commoditizing, and 7 million drones a year from one country at war suggests it is, then the profit pool migrates to whoever owns the coordinating software, the way it migrated to Microsoft and away from PC assemblers. That is the bet behind Swarmer's licensing model and Havoc's collaborative autonomy. For investors, it reframes the neo-prime thesis: the winners may not be the ones with the biggest factories but the ones whose software everyone else's hardware has to run. The catch is that this is still a bet on industry structure, Fink said so plainly, and the incumbents (and Anduril's Lattice) want to own that same layer. Whoever wins it wins the most durable position in the sector.
5. The Ammunition Arithmetic, and an Operator Inside Government Saying We Are Already Losing the Drone Race
Everything above assumes the West can actually build the weapons. Two conversations this fortnight put hard numbers on how far it has to go.
On STRAT (August 7), retired Marine intelligence officer Hal Kempfer (an analyst with an operator's background) laid out the Patriot-missile math, and it is sobering. "The U.S. only produces about 600 of these missiles a year," he said, against "thousands and thousands of missiles that these things are supposed to intercept." That rate "is only sustainable during peacetime." In the Gulf fight with Iran, he noted, allies have been firing "three, four, I've heard up to five Patriot interceptors for one ballistic missile," and the West has even been using expensive Patriots to knock down cheap drones, an exchange that "will literally wear us out economically." There is a funded plan "to increase the production rate to about 2,000 missiles a year," but ramps take years. He also flagged the awkward Ukraine subplot: a tentative offer to license a cheaper Patriot for Ukrainian production, floated in a Zelensky–Trump meeting, that Trump then walked back, plausibly, Kempfer speculated, over fear that "crown jewel secrets" in the missile could be "stolen... by an adversary intelligence service." Europe's own interceptors, Germany's IRIS-T, France's SAMP/T, are "very capable" but built in limited numbers.
The blunter, higher-authority version came from an elected operator. On the Hudson Institute Events Podcast (August 18), Senator Rick Scott delivered a stark comparison: Ukraine "will build over 5 million drones this year" while the U.S. builds "less than 100,000." His warning on the Navy was unusually direct, "there is no question that our Navy must move decisively to integrate unmanned systems into the fight. The time is now and the Navy is too late to the game already." He argued China "has put unblue [untold] resources into space and we haven't," that Beijing is "way ahead of us as far as the amount of drones," and that the answer has to be mass and swarms: "we need to be able to swarm them... and the only way you can do that is not with people. You've got to do that with unmanned systems." His fix for the sclerotic buying process was to "make decisions faster... not in five years," with a defense-funding reconciliation package he expected "hopefully next week."
Why it matters: these two conversations are the demand-and-urgency floor under the entire investment thesis. A 600-per-year Patriot line against thousands of incoming threats, and a 100,000-versus-5-million drone gap with a wartime adversary's supplier, is precisely why capital is flooding in, and precisely why the production-capacity names, the interceptor makers, and the domestic component base have a multi-year tailwind that does not depend on any single conflict. But it is also the warning: the gap is this wide after years of alarm. Money is necessary but not sufficient. As Scott's own "not in five years" plea makes clear, the binding constraint is the speed at which dollars become hardware, the same execution risk that shadows every bullish section of this issue.
6. Boeing Quietly Exits Autonomy, and the Flying-Taxi Crowd Runs to the Pentagon
A revealing corporate shuffle this fortnight showed how the commercial and defense worlds are bleeding into each other. On The Road to Autonomy (August 14), the hosts (analysts) unpacked Boeing's decision to hand its autonomous-aviation assets to Archer Aviation. In an all-stock deal signed August 10, Archer acquires Boeing's WISK, Arrow, In-Situ, and SkyGrid; Boeing takes "a Class A equity stake of up to 19.75%," commits "to invest up to $55 million" in Archer's next raise, and gets a cross-licensing agreement. The prize inside the deal was In-Situ, "a profitable defense drone business... that generates over $200 million in annual revenue across 35 countries." As one host put it, the deal was "not just about the potential in WISK and SkyGrid, but also the need for Archer to gain some immediate revenue." Tellingly, Boeing "sold them an autonomous eVTOL design, but kept the autonomous software for themselves."
The same migration, money-losing air-taxi hopefuls buying their way into steady defense cash, played out on Motley Fool Hidden Gems Investing (August 12), where the analysts dissected Joby Aviation's purchase of a defense business (Blade) for "$500 million" with "$450 million cash." The market was unimpressed, Joby's stock fell "about 3%," while Archer's similar move was "loved." The debate was whether this is prudent hedging or a distraction. One host's read: it "appears to be a deliberate move to diversify," an implicit admission that commercial eVTOL "is much harder than initially expected and will likely take a lot longer." Another was sharper: "they're trying to be both commercial companies and defense contractors. And it's really hard to do both when you don't have either completed at this point." The bull framing was that defense is "steady... cash generative" and can "give yourself a lifeline" without repeatedly tapping the equity markets. The line that stuck: the thing that "scared me the most is when [they said] a year or three" late on the core air-taxi timeline.
For a contrast, Valley of Depth (August 13) featured Isaac Roberts, chief strategy officer of Skyways (an operator), whose company represents the road not taken by the eVTOL crowd, defense-first from day one. Skyways has "about $24 million" in funding and roughly the same in defense revenue, building autonomous cargo drones for one narrow, deadly problem: "one third of casualties in the Iraq and Afghanistan wars were in logistics." Its V3 aircraft targets "100 pounds" of payload over "1,000 nautical miles" using a heavy-fuel engine, and it has climbed from a two-unit Marine Corps buy to a Navy contract under the Blue UAS program to an Air Force deal, beating Boeing along the way "on performance." Roberts' explanation of why a focused startup beats a prime is a small masterclass: a prime has "a million and one things going on... resource fighting," and "it's hard to fight for a project that hasn't yet been birthed." And his cultural jab at the flying-car set doubled as strategy: the defense "customer doesn't care about ESG. The customer's like, you need to be able to go a thousand miles with 80 pounds. How do you do this?"
Why it matters: the direction of travel is unmistakable, commercial autonomy is discovering that defense is where the near-term revenue lives, and defense is discovering it needs the commercial world's speed and software. Boeing offloading a cash-burning autonomy portfolio while keeping the software is a shrewd balance-sheet move that preserves optionality; the eVTOL names buying defense businesses is a hedge against their own slipping timelines. But the market's split verdict, cheering Archer, shrugging at Joby, is a reminder that investors will reward focus and punish what looks like a company trying to be two things at once. Skyways is the cautionary counterexample: the pure-play that beat a prime by refusing to be distracted.
7. The Landmine Under the Whole Boom: Who Owns a Contractor's Crown Jewels
Every bullish thread in this issue runs into one wall, and it is not the budget. On Double Take by BNY Investments Newton (August 6), the hosts opened with the puzzle investors have been chewing on: after defense stocks "went on a tear" when the administration pushed to raise spending "by more than 50%," the rally stalled. Part of the reason was the realization that a 50%-plus increase "would not happen to the level he'd hoped." But the more insidious reason, they argued, is a piece of legislation.
The bill is marketed as a "right-to-repair" law, the Pentagon's answer to a chronic "readiness problem," its struggle to maintain weapons over decades of service. Margaret Boatner of the Aerospace Industries Association, a former Army acquisition executive (industry operator), explained what it would actually do: make "government purpose rights" the default for every contractor "simply as a condition of doing business with the Pentagon." And "don't let the name fool you," she warned, those rights let the government "use or change or share that proprietary information," including "with other contractors, including competitors." Her analogy: imagine telling "Apple or Google or some cutting-edge tech startup that in order to sell a product to the government, they have to be prepared to provide their proprietary information... who can then later share it with potential competitors."
Moshe Schwartz, an acquisition-policy specialist (a pundit), argued the whole premise is a red herring: "roughly 90% of the readiness issues that the department has is not related to intellectual property." The real problems, he said, are physical and bureaucratic, decrepit maintenance depots (the Army's own 15-year, $18 billion upgrade would reach "21st century technology 40 years after that century has begun"), computer systems that "didn't talk to each other," missing parts, too few mechanics. His warning on consequences is the part investors should tattoo somewhere: the defense industrial base has already "shrunk by 40%" since 2010, and many suppliers do "90 to 95% of their work" in commercial markets. Faced with risking that IP for a sliver of defense revenue, companies will do one of three things, "leave the defense market altogether," decline to bid, or offer "not their best technology, but their second best or their third best." Boatner added the financial kicker: because contractors earn much of their return on the long "sustainment tail," weaker IP protection means "you would expect to see additional costs kind of baked in upfront."
Why it matters: this is the direct contradiction at the heart of the sector. Sections 1 through 6 are all about Washington trying to pull more commercial companies and more private capital into defense. This bill pushes them the other way. For the primes, weaker IP protection strikes at the sustainment profits that underwrite the whole business model, and would likely raise upfront prices. For the startups and the venture money behind them (Section 1), the IP is the asset; threaten it and the calculus Chapman described, write a $250 million check into a defense company, gets a lot shakier. It is not law yet, and it helps explain why the defense trade cooled even as every fundamental driver pointed up. It is the single policy item most worth tracking into the fall, because it can quietly undo the capital flood this entire issue has been describing.
The Bottom Line
The through-line of the last two weeks is that this sector has crossed from whether to who. Whether the money would come is settled, over $100 billion a year of venture capital, an IPO window flung open for "boring" industrials, and budget talk reaching for $1.5 trillion. The live questions now are distribution and durability: whether the value accrues to the incumbents or the insurgents, to the hardware or the software, to the primes' sustainment margins or a startup's licensing model, and whether any of it converts into missiles, drones, and interceptors fast enough to close the gaps that operators keep measuring in embarrassing multiples (600 Patriots a year, 100,000 American drones against 5 million Ukrainian ones).
The optimism is earned. Robot fighters are carrying live missiles; a swarm-software founder is running a public company before it turns five; a focused logistics startup beat Boeing on performance. But the two constraints named this fortnight are the ones that decide whether the boom compounds or fizzles. The first is execution, the slow, unglamorous business of turning appropriations into hardware, which Rick Scott begged the system to do "not in five years." The second is a self-inflicted one: a rulebook that, in trying to fix a readiness problem, could seize the very intellectual property that makes companies want to show up. The demand is not the debate anymore. The machinery behind the money is.