The Rise of Launch Plus: Inside the $67B Space Economy Boom

Zach Martin

Here's a number to hold onto for the next hour. Sixty seven point six billion dollars. That's what Space Capital's Space IQ data says was invested across the space economy in the first half of 2026. Now set it next to the full year of 2025, which came in at fifty five point three billion. So half a year just outran an entire year. I'm Zach Martin, this is SPACE CAPITAL on KMKT IR Hub Radio, and it's autumn 2026. Today I'm not going to replay the SpaceX debut from June. We covered the market reaction to that in an earlier episode. Today we go underneath it, to the plumbing that turns a loud headline into an asset that lasts. Because a big number is a scoreboard, and a scoreboard doesn't tell you how the team actually plays.

Zach Martin

Let's get the scoreboard straight first, and I want to keep the qualifiers attached. Space IQ is Space Capital's quarterly report, manually curated from more than three thousand companies. By its count, more than four hundred eighty eight billion dollars has gone into the space economy since 2009, sorted into three layers: Infrastructure, Distribution, and Applications. Inside this year, the first quarter brought in a record thirty six billion dollars across one hundred forty eight companies. The second quarter brought thirty one point six billion across one hundred twenty nine companies. Add them up and you get the sixty seven point six billion for the half. Notice something, though. The second quarter came in below the first, and across fewer companies. I'd read that as one sign of capital getting choosier, not a stampede. That's my interpretation, not a claim from the data, but it's a useful habit. When a number is a record, ask what happened right after the record.

Zach Martin

Space Capital's founder and CEO, Chad Anderson, has a phrase for where we are. In his January predictions piece, he said the skepticism phase of the orbital economy was effectively over and the industry had entered what he called the industrial phase. He also framed the SpaceX listing as space tech's Netscape moment. Now, keep in mind who's talking. Space Capital is a venture investor in this sector, and it has invested across these layers since 2012. That doesn't make the framing wrong. It means you should treat the framework as a lens from someone with a stake, and then check it against the contracts, the acquisitions, and the hardware. That's what we're going to do. I count four mechanisms doing the real work right now: NASA behaving like a customer instead of a builder, defense buying maneuverability, a sober look at orbital compute, and satellites made the way chips are made. But first, one idea that ties the capital story together.

Zach Martin

The idea is called Launch Plus, and it's a new category in the Q2 Space IQ report, published July 15. Here's the data behind it. Pure launch, companies that only launch rockets, has attracted about eighteen point three billion dollars since 2009. That's roughly four percent of all the capital invested in the space economy. Satellites took eighty one percent. So the money has always known something that the headlines kept forgetting: a rocket is a truck, and nobody builds a business whose only product is the truck. Space Capital has said for years that launch alone doesn't make a great business. What Q2 added is evidence for the corollary. Launch Plus companies led all Infrastructure investment at fourteen point six billion dollars year to date, and they drove seventy one percent of the quarter's top Infrastructure deals, in what Space Capital calls the largest Infrastructure quarter on record.

Zach Martin

What qualifies as Launch Plus? The test is whether a company's business is integrated across launch and at least one other space industry, not merely sitting next to launch. Space Capital describes two forms. The first is integrated operators. SpaceX is the fullest example: launch, satellites, Dragon crew and cargo to the space station, a Starlink based space situational awareness system called Stargaze, a cargo return capsule called Starfall for in space manufacturing, with lunar ahead. Blue Origin combines launch, satellites, and lunar. Rocket Lab combines launch, satellite manufacturing, and satellite communications. Firefly pairs launch with lunar landers and orbital vehicles. Relativity pairs launch with a Mars orbiter program. The second form is digital engineering companies, the ones building the design, simulation, testing, and operations software that rocket and satellite makers stand on. Space Capital names Prometheus, Nominal, Sift, and Revel in that group.

Zach Martin

The edge case is the most interesting one, and Space Capital calls it the Prometheus test. Why would an AI company belong in a launch category? Their answer is that they classify companies by the infrastructure being built and the markets it serves, not by whether the product is hardware or software. Prometheus raised what Space Capital describes as the largest deep tech round ever for physical industrial engineering, to build an AI system that designs, manufactures, and operates physical assets, with aerospace as a stated target. Jeff Bezos has pointed to Blue Origin as the case study customer, and Blue Origin's CEO, David Limp, sits on the Prometheus board. The thesis, as Space Capital tells it, is that the scarce input for physical AI is proprietary industrial data that's very hard to access, and a launch to orbit stack is exactly where that data lives. You can agree or disagree with the taxonomy. It's one firm's classification. But it tells you where an investor with a stake thinks the value accrues.

Zach Martin

Now watch the plus get bought. In Q2, Rocket Lab announced an eight billion dollar acquisition of Iridium, at fifty four dollars per share in cash and stock, expected to close in the middle of 2027. That's pending, so it hasn't happened yet. Why Iridium? Space Capital's answer is spectrum and services, and I think this is the single most important idea in the whole deal. Spectrum is the scarcest asset in the space economy, because nobody can make more of it. Iridium's L band licenses are globally harmonized, meaning they work everywhere on Earth rather than country by country. Add roughly two and a half million subscribers and decades of government and safety of life contracts, and Rocket Lab gets an installed base in the markets heating up fastest: direct to device, positioning and timing, and aviation. Rocket Lab also closed its acquisition of Mynaric, adding laser optical terminals and its first European entity. And Space Capital notes that Amazon paid eleven point six billion dollars for Globalstar in the same quarter, on the same logic.

Zach Martin

Here's my coaching analogy for that, and it's mine, not Space Capital's. Think of a rocket company as a team that's great at getting to the stadium. Wonderful skill. But the money in a stadium is in what you own once you're inside: the seats, the broadcast rights, the concessions. Reaching orbit is the ticket gate. Spectrum and subscribers are the building. Space Capital puts it in one line: the value of reaching orbit is what you own once you're there. And their forecast follows from it. They say standalone launch will atrophy over time, and they expect more high value mergers, more teaming agreements, and more launch companies expanding into services as they reach orbit. That's a forecast, not a fact, so mark it as one. But the Iridium and Globalstar prices are facts, and they point in the same direction.

Zach Martin

Next, the bridge to public markets, because it explains why all this capital can move the way it does. In May, Space Capital formalized a partnership with Nasdaq. Each quarter the Space IQ briefing is filmed and published from Nasdaq MarketSite. Nasdaq's Jeff Thomas said the exchange has been there to support every major technology revolution and expects to play the same role as the space economy comes to the forefront. Then on June 26, the Space Capital Summit convened at that same MarketSite, two weeks after the SpaceX listing. On exits, Space Capital's own accounting says the SpaceX IPO was the largest liquidity event in venture capital history, surpassing the full year exit value of 2021. It also says HawkEye 360 listed, and that Blue Origin acknowledged for the first time that it will need outside capital to scale, without ruling out going public. And currency now flows both ways. SpaceX used its own record priced stock in a sixty billion dollar all stock deal for Cursor, and Rocket Lab is paying for Iridium partly with its own shares.

Zach Martin

A word of discipline here, because I'd be doing you a disservice if I only gave you the cheerful side. Space Capital's own roundup of coverage includes a MarketWatch headline that asks why the SpaceX stock is now cratering. I don't have that article's details, so I won't characterize it. But a public market gives you a daily price, and a daily price goes both ways. Paper equity swaps look wonderful when the paper is rising. So the real question for any allocator isn't whether the window is open. It's whether the underlying assets, the contracts, the spectrum, the hardware, hold their value when the stock chart gets ugly. That's why the rest of this episode is about mechanisms. The Summit that Space Capital hosted had one line that I think is the thesis for today: every durable buildout starts with a customer changing how it buys. So let's go meet the customers, starting with the biggest one in civil space.

Zach Martin

The Summit's opening conversation was NASA's Deputy Administrator, Matt Anderson, talking with CNBC's Morgan Brennan. And the message was blunt. NASA is no longer primarily an operator of space systems. It is a customer for them. In the words of Space Capital's write up, space becomes infrastructure the government buys, not exploration the government performs. Anderson's phrase for the role is anchor customer, not sole builder. Now what does that mean in practice? As described at the Summit, NASA is using milestone based, firm fixed price contracts. You deliver on schedule and on budget, you earn recurring task orders. You fail, and the work gets reallocated. Let me make that concrete with a kitchen remodel. The old model is hiring a contractor by the hour, with the bill rising as the plans change. A fixed price contract is a bid. The contractor eats the overrun. That one change flips who carries the risk, and that's why it reshapes who can compete.

Zach Martin

The Lunar Outpost founders describe the old approach from the inside, and I find it clarifying. Both Cyrus brothers started their careers at Lockheed Martin. Justin worked on military satellite programs. Julian worked on avionics for the Orion human spaceflight program. And what they concluded, according to Space Capital's profile of the company, is that the legacy model could build extraordinary single mission systems, but cost plus contracting rewarded delay and perfection over learning. It couldn't produce iterative, services oriented infrastructure. That's the company's own origin story as told by an investor, so treat it as a perspective. But it matches what NASA now says it wants to buy: services, delivered on a schedule, with a price that doesn't float.

Zach Martin

Let me give you the schedule that Anderson laid out, because the cadence is part of the message. Artemis II flew its crewed lunar flyby in April 2026. Artemis III, targeted for 2027, is described as a low Earth orbit trial that needs the SLS rocket, Starship, and New Glenn all in orbit at the same time. Anderson's shorthand was three rockets, two dockings, and one splashdown. Then Artemis IV, targeting 2028, is the first crewed landing of the new era, at the lunar South Pole. Behind it sits the Ignition plan that NASA Administrator Jared Isaacman announced in March 2026: a ten year, three phase commitment to a permanent Moon Base. The orbital Gateway has been paused. Space Capital's Nasdaq announcement refers to a twenty billion dollar lunar base commitment. And in low Earth orbit, the International Space Station retires in 2030, so NASA expects private stations, such as Vast's Haven-1, which is slated to launch in 2026, to become the habitats where NASA is one anchor tenant among several customers.

Zach Martin

Notice the pattern. A customer who says what it will buy, in what quantity, by when, and at a fixed price. For a company, that's revenue visibility. For an investor, it turns a science project into something you can underwrite. And the company I want to walk through to show you the whole mechanism is Lunar Outpost, because its story compresses fifty years into a few paragraphs. Start with the last time an American vehicle drove on the Moon. December 14, 1972. Eugene Cernan drives the Lunar Roving Vehicle to a final position in the Taurus Littrow valley, parks it, and walks away. The vehicle was built for a single mission of roughly three days. Its silver zinc batteries were single use. Its thermal systems were never designed for the swings between lunar day and lunar night. Across the three missions that used them, the rovers drove about ten hours and roughly ninety kilometers in total. And then, for fifty three years, no American vehicle touched the lunar surface.

Zach Martin

Why the gap? Space Capital's answer, in its profile of Lunar Outpost, is that the infrastructure problem went unsolved because no commercial framework existed to solve it, and legacy cost structures made bespoke, single mission hardware the only thing anyone knew how to build. I'd add my own reading: a rover you abandon after three days can't have a business model. A rover that runs for a year, across many missions, can. That's the whole bet. Now the company. Lunar Outpost was founded in 2017 by brothers Justin and Julian Cyrus, with Forrest Meyen joining soon after as a third founder. Meyen has heritage on MOXIE, the experiment that produced the first oxygen ever generated on another planet, aboard NASA's Perseverance rover. AJ Gemer became chief technology officer in the company's first year, after a decade at the Laboratory for Atmospheric and Space Physics at the University of Colorado Boulder. Justin's master's degree was in Space Resources at the Colorado School of Mines, the first academic track built around the engineering and economics of using resources beyond Earth.

Zach Martin

I need to give you a disclosure that the source makes plainly. Space Capital backed Lunar Outpost at the Seed stage in 2022. So the account I'm drawing on is an investor describing its own portfolio company, and it's written admiringly. I'm giving you the facts it reports, and when I interpret, I'll say so. One of those facts is a delightful one. In 2021, then NASA Administrator Bill Nelson presented Lunar Outpost with a ten cent check, representing ten percent of the company's one dollar winning bid to collect a sample of lunar regolith. The dollar amount was nothing. What it established was that a commercial company had been paid by the U.S. government for material to be extracted from another world, which Space Capital describes as the first sale of space resources in human history. Discipline starts small. A one dollar bid is a way of saying, we'll prove the transaction before we prove the profit.

Zach Martin

The company's early rover line was called MAPP, the Mobile Autonomous Prospecting Platform, a ten kilogram class rover. On February 26, 2025, Lunar Voyage 1 launched aboard Intuitive Machines' IM-2 mission, carrying MAPP toward the lunar South Pole. On March 6, the Athena lander touched down near Mons Mouton and came to rest at an off nominal attitude, and the mission window closed within twenty four hours, before MAPP could roll onto the surface. So, to be clear, the rover never drove on the Moon. And yet the company's account is that the rover validated a lot in that window. Its navigation computer, its autonomous thermal control, its stereo cameras, and a patent pending deployment system called Talon all worked. It stood up the first commercial 4G LTE network on the lunar surface, in partnership with Nokia. And its Stargate command and control software reached technology readiness level nine, the highest level NASA assigns, meaning it had been proven in the environment it was built for.

Zach Martin

I want you to take the coaching lesson out of that, because it's the same one I give about business growth. Small, honest repetitions compound. A flight that doesn't fully succeed can still be a data set, if you built the next one to use it. Next up, Lunar Voyage 2, slated for the second half of 2026 aboard IM-3, carries MAPP to study the Reiner Gamma magnetic anomaly, and brings the first onboard GPU compute to the Moon. By integrating NVIDIA edge processors, the rover can handle its own LiDAR, map hazards, and plan routes without waiting on a signal from three hundred eighty four thousand kilometers away. Lunar Voyage 3 follows in 2027. An Australian rover, co developed with the Australian Space Agency, is built to turn lunar regolith into usable resources. And Lunar Voyage 5 deploys with Artemis IV in 2028, as the first robotic rover to work alongside astronauts on the surface. That's part of ten contracted missions launching before 2030.

Zach Martin

Then comes the step change. On May 26, 2026, NASA selected Lunar Outpost and Astrolab to deliver the first human rated lunar terrain vehicles to the surface. It narrowed three Phase 1 finalists down to two. And this is the line from the source that matters most: it was the first time NASA had selected commercial companies to provide astronaut surface mobility as a service rather than build it in house. Lunar Outpost's piece came as a two hundred twenty million dollar firm fixed price task order to advance Pegasus, the lighter variant of its vehicle. Pegasus carries two astronauts side by side in a low profile form that echoes the Apollo rover. It drives three ways: crewed, teleoperated from Earth, or fully autonomous. It's described as capable of more than nine miles per hour across a nine hundred kilometer range, working for a full year on the surface, and compatible with both Blue Origin's Blue Moon Mark 1 and SpaceX's Starship landers. The company is charting a path to deliver the flight vehicle to NASA in November 2027.

Zach Martin

Now the number everyone quotes: the broader Lunar Terrain Vehicle Services program carries a four point six billion dollar ceiling through 2039. Read the qualifier carefully. It's a ceiling, and Space Capital describes it as structured as per mission task orders rather than one fixed contract. A ceiling is the most NASA could spend, not what it will. The firm money today is the two hundred twenty million. The rest depends on what I'd call earning your next order, which is exactly the anchor customer model working as designed. Space Capital's argument is that the company is positioned to capture a substantial share of the demonstration and the recurring service missions that follow. That's a positioning claim, not a guarantee. If you're allocating capital, that distinction between a ceiling and a backlog is the difference between a story and a balance sheet.

Zach Martin

How did a company that built ten kilogram rovers scale to a vehicle for astronauts? By carrying the same loop up the ladder. Within twelve months, per the source, it built seven distinct engineering prototypes of its larger Eagle vehicle, each designed to stress a different subsystem. A consortium called Lunar Dawn formed around it: General Motors for battery and electric drivetrain technology, Goodyear for non pneumatic metal mesh tires designed for regolith, MDA Space for robotic manipulation, and Leidos for human factors and mission logistics. In late 2025, Eagle went on a road tour, covering eight thousand five hundred miles over thirty one days, from the Capitol lawn to the banked oval at Talladega, with more than one hundred twenty five thousand visitors. Then Pegasus was iterated down from Eagle using multiphysics simulation and digital twins, and two full scale prototypes have been through two rounds of suited testing at Johnson Space Center with former astronaut John Grunsfeld. Eagle remains the long duration flagship, with Pegasus reaching the surface first.

Zach Martin

Here's why the capital story isn't just about wheels. Space Capital's argument is that mobility activates every other layer. A power station nobody can drive to is a stranded asset. A communications relay without endpoints is a tower in the dark. The forty kilowatt fission surface power system NASA is targeting for 2030 arrives on a lander but has to be towed across kilometers of terrain to its site. That's the writer's logic, and I find it persuasive as engineering, though it's still an argument rather than a measured market size. And Lunar Outpost isn't selling only vehicles. It has a three part software suite: Stargate for command and control, Starweave for swarm software that lets different robots act as one fleet, and Spark, an AI mission operations assistant. A thirty million dollar Series B is funding those teams. The engine behind Starweave, developed under Air Force Research Laboratory and Space Force SpaceWERX contracts, gets its first orbital test this fall. MARS-1 launches on a Falcon 9 via an Exotrail transfer vehicle, and two small spacecraft will run decentralized swarm operations in low Earth orbit.

Zach Martin

Look at the dual use pattern, because it's the risk management inside the business. The company's first product was Canary, an air quality sensor built for lunar habitat monitoring under Lockheed Martin's NextSTEP program, which then became a real terrestrial business in methane detection and urban air quality. The swarm software was funded by defense contracts and aimed at the Moon. And the thermal work, a European Commission funded in orbit demonstration called TACOS, helped qualify the hardware that earned Lunar Outpost Europe, its Luxembourg entity, a role as thermal and systems engineering lead for ESA's Moonraker lunar polar orbiter, launching around 2030. Same discipline each time: one technology, more than one customer. Space Capital also notes the surface is filling up, with six different programs pursuing lunar surface efforts, led by China and joined by India, Japan, South Korea, and ESA, and China developing a crewed rover for a 2030 landing. An anchor customer matters most when there's a race on.

Zach Martin

One more detail, because it's a nice image of where this industry's head is. The Summit ended with a Lunar Outpost lunar terrain vehicle parked in Times Square. A vehicle built for the Moon, standing outside the building where the market prices the companies that build such things. I don't want to oversell a photo op. But it makes the point about what changed. The government is buying a service, and the vendors are public enough, and funded enough, to park hardware in front of allocators. That's the first mechanism. The second is the highest stakes customer of all, and it's changing how it buys faster than almost anyone expected.

Zach Martin

For sixty years, the United States has defended itself mostly from the ground: radars, sensors, and interceptors rooted to the Earth. That center of gravity is moving to orbit. I'm drawing now on the Summit session called Shield and Spear, where Space Capital's Chad Anderson interviewed Marcia Holmes, Deputy Director of Golden Dome for America, and Eric Romo, President and COO of Impulse Space. Disclosure again: Impulse Space is a Space Capital portfolio company. Anderson opened by saying a lot of how the work is being done can't be discussed in a room like that, so, as he put it, they weren't there for the secrets, they were there for the direction. That's the right way to hear this chapter. We're tracking direction, from public statements, not classified detail.

Zach Martin

Holmes started with the threat, and she didn't soften it. She said the threat is real, complex, and growing. Russia has fielded hypersonic and Avangard systems and used them on the battlefield in Ukraine. China, she said, is projected to field over one thousand intercontinental ballistic missiles, doubling its arsenal. Cruise missiles, hypersonics, and ballistic missiles can now range the United States directly. Her words: our homeland is no longer a sanctuary. And the implication she drew is that ground based systems can't see the newest threats early enough. You need to see them from launch all the way through their flight trajectories, and you need to move interceptors into position before a missile deploys its warheads or decoys. Space based assets, she said, are the only thing that will enable that. Those are her statements as a government official, and I'm reporting them as hers.

Zach Martin

Why now? Romo's answer began where every space conversation begins: launch costs and access to space are dramatically different than they've ever been. The last time the country seriously discussed intercepting missiles from space was the 1980s, when reaching orbit cost a fortune. A constellation of a thousand spacecraft in low Earth orbit was a pipe dream until Starlink and others made it routine. And that maturity trickles down. Romo said of the manufacturing base that he's not doing this for the first time, he's doing it for the second or third time and scaling up. Think of it like a restaurant opening its fifth location. The first one is a miracle. The fifth is a process. That analogy's mine, but the point is his: repetition is what makes a defense architecture buildable on a budget.

Zach Martin

Now the doctrinal shift, which is the center of this chapter and the reason it's called the Age of Maneuver. The Pentagon's strategy for low Earth orbit has been proliferation: put up so many satellites that an adversary can't take them all out. That works in low orbit. But it has limits. Romo made a physical point that I think is the cleanest in the entire discussion: medium Earth orbit and geostationary orbit are, in his words, literally a million times bigger than low Earth orbit. Blanketing volumes that vast with enough satellites to cover every point at every moment is, in his word, completely untenable. And yet those high orbits are where much of the most critical infrastructure lives: GPS in medium Earth orbit, missile warning and secure communications in geostationary orbit. Those assets, he said, are under pressure for the first time.

Zach Martin

So if you can't defend by numbers, what's left? Assets that can move. Romo pointed to recent reports of three Russian craft maneuvering around a U.S. imaging satellite over Ukraine as a sign that everything up there now has to be able to move. Back in January, Chad Anderson's prediction piece put it as a market thesis: in 2026, maneuverability becomes the most valuable currency in orbit, and he told readers to watch for mergers and contract awards focused on guidance, navigation, and control, and on non Earth imaging. He also wrote that defense is no longer interested in sitting duck spacecraft. That's a prediction, and Space Capital, as an investor, benefits if it's right. But now we can hold it up against what Holmes and Romo said in June, and the two line up.

Zach Martin

Impulse Space is already flying parts of the capability. Its Victus Surgo and Victus Salo vehicles, both for Space Domain Awareness missions, were in the clean room at the time of the conversation. They're built for fast, high energy repositioning, one bound for a geostationary transfer orbit and one for low Earth orbit. And Impulse, with Starfish Space, completed a fully autonomous rendezvous in orbit, in December 2025: one vehicle approaching another using a single camera, no joystick. Romo drew a distinction I keep thinking about. The ability to reach out and touch an asset when it wants to be touched is one thing. When it does not want to be touched is another. And he was candid that none of it is easy. Closing velocities run to thousands of meters per second. Hitting a target is not so much a bullet hitting a bullet, he said, as maybe a hundred times harder than that.

Zach Martin

Here's where the capital story comes in, and it's about arithmetic. A recent Congressional Budget Office estimate pegged space based interceptors at roughly twenty two million dollars a shot. In historical context, Romo noted, that isn't outrageous, because ground based interceptors such as the SM 3 and the GBI run from thirty million to nearly one hundred million dollars. But multiply twenty two million by the thousands of interceptors a real architecture would need, and the number gets very large, very fast. And Holmes was direct about the constraint. General Guetlein, who leads the program, has said publicly that if space based interceptors cannot be bought affordably and at scale, they will not be part of Golden Dome. Let that sink in. The program's own leader is saying the technology can be proven and still be cut, if the unit price doesn't work. That's the same logic as the fixed price lunar contracts, applied to national defense.

Zach Martin

Her proposed fix is procurement itself. Because Golden Dome is net new, it can start from a clean sheet rather than fifty years of cost plus habit. In her words, we're giving you prize based competitions, rewarding demonstrated performance instead of paperwork and dealing directly with builders to compress timelines. Romo was a fan: you set an ambitious goal, and if you get it, you get paid a bunch. If you don't, sorry. I like that as a discipline test. It means the company has to be able to carry the risk until the result shows up. That favors well capitalized, vertically competent teams, and it explains why the venture and public capital arriving in this sector matters to defense, not only to investors. Balance sheets are now part of national security.

Zach Martin

There's a second half of the shift that's easy to miss. Part of it is away from purely kinetic defense. Holmes said the goal increasingly is to buy decision makers time and hand them better options, with AI surfacing recommendations faster. Strategic assets in space actually buy back time, she said. Romo made the stakes concrete. Intercepting a missile in its boost phase means detecting a launch, relaying it, getting a human decision, and reaching the target, all inside a window measured in seconds, against something that may be accelerating through the atmosphere behind a wall of plasma. Persistent overhead coverage is the only way to compress that chain, and that's the argument for a large, maneuverable fleet. Hypersonics are harder still, he added, because they stay in the atmosphere and can bank and dive away from an interceptor. As Holmes put it, these are very hard problems.

Zach Martin

Where's the opportunity for the people in that room? Anderson asked exactly that. The answer was a layered defense running from space to land to sea, with a long list of pieces still to be built: low cost kinetic and non kinetic interceptors, AI and autonomy for automated decision making, track fusion, and low latency communications to stitch it together. Holmes described Golden Dome's ecosystem hub as meant to be a single front door for companies to find where they fit, with primes subcontracting to nontraditional companies inside its command and control consortium rather than the other way around. And the Summit's summary of the session put the through line this way: deterrence in orbit is now a commercial supply chain question. The capabilities Golden Dome requires exist in companies, not programs of record, and the procurement system is being rebuilt to reach them.

Zach Martin

Let me add some context on the money, from Space Capital's Nasdaq announcement. The fiscal 2026 U.S. defense budget is roughly one trillion dollars, with a record one point five trillion proposed for fiscal 2027, and space, cyber, and AI explicitly prioritized. Proposed is the operative word there. Space Capital also describes national security spending as providing a floor for the space economy. That's their framing, but it fits what Holmes said: capital markets, policy, and acquisition reform aligned at once. She called it a perfect storm, almost. Golden Dome is meant to field real capability by 2028.

Zach Martin

And now the caution, because Romo gave one, and it's the best advice in the whole session. The early demonstrations are not what pays off, he said. Long term production is, and confidence in it will ebb and flow with news cycles and OMB reports. The work is to stay the course. He added that the country is swinging back pretty hard on manufacturing and building things again in America. So how do you track it? Both speakers said the signal to watch is not press releases but contract awards, test campaigns, and the steady expansion of the space budgets that quietly underwrite it. That's the investor's checklist, and it's the same one I'd give about any business claim. Ignore the announcement. Follow the awards, the tests, and the budget lines. That's what compounds.

Zach Martin

Let's talk about the idea that dominates the conference circuit: data centers in orbit. And I want to do it the way Space Capital's own piece did, which is titled Orbital Data Centers: The Honest Case. The voice is Jonny Dyer, founder and CEO of Muon Space, a company that builds and operates satellites. His session at the Summit was titled When Kepler Met Turing, and his argument was that the boundary between space infrastructure and terrestrial infrastructure is dissolving. Orbital data centers, on board processing, and AI systems trained on real time views of the physical world all treat orbit as an extension of the compute stack rather than a separate domain. He also placed it in a geopolitical frame: the U.S. and China are racing to industrialize the same layer, and the outcome shapes who owns the infrastructure the next generation of AI runs on.

Zach Martin

Start with why anyone would try this. Terrestrial AI data centers are running into bottlenecks: electrical grid saturation, land permitting, and local fights over water for cooling. Space Capital's January predictions put the whole thing in one word, energy. The pitch is that orbit offers continuous solar power, so you avoid the battery storage that ground based solar needs, plus a very cold radiative heat sink and no municipal zoning board. That's the seductive version. The thesis, as Anderson wrote in January, is infinite solar power and the ultimate heat sink. And look at who's pulling at this thread: he named Elon Musk, Jeff Bezos, Sam Altman, and Eric Schmidt as heavyweights involved, and Space Capital's Nasdaq announcement refers to an orbital compute race between SpaceX, NVIDIA, and Google.

Zach Martin

Now the honest part, and it's the most useful number in this episode. According to Dyer, a large orbital data center today runs roughly eight to ten times more expensive than its terrestrial equivalent. Eight to ten times. Not twenty percent. And his conditions for closing that gap are specific. Launch costs need to fall by about five times, and launch cadence needs to rise about ten times, from the current rate of roughly one hundred to two hundred launches per year to something like five hundred to one thousand. Put differently, the physics may be fine, but the shipping math isn't there yet. If you've ever tried to start a business that depends on a supplier who hasn't scaled yet, you know the feeling. The idea is not wrong. The idea is early, and early costs money.

Zach Martin

What about the objections people raise first? Radiation, for one. Cosmic rays flip bits in chips. Dyer's response, as summarized, is that terrestrial hyperscale chips are already engineered against cosmic ray bit flips, so that's less of a barrier than critics assume. Bandwidth is another. Data centers on the ground connect chips with enormous fabrics measured in terabits. Starlink style optical inter satellite laser links run around one hundred gigabits per second, and Dyer's point is that they're steadily narrowing that gap. I'd flag that one hundred gigabits versus a terabit is still a real difference, and the source describes narrowing, not closing. And then there's heat. Anderson's January piece was explicit that the thermodynamics of heat rejection in a vacuum remain a complex engineering hurdle that only certain companies will overcome. A vacuum is a great insulator, which is the problem. You don't blow air over a chip. You have to radiate the heat away, with large surfaces. So 'ultimate heat sink' is the pitch, and heat rejection is the homework.

Zach Martin

Here's how I'd frame the near term, and it's grounded in what Dyer actually described. The first useful orbital compute isn't running a consumer chatbot in space. It's specialized, high value edge computing, where processing the data where it's collected beats sending everything down. His example is Muon Space's FireSat constellation, developed with Cal Fire, Google, and the Environmental Defense Fund. It runs edge AI directly on orbital sensors to broadcast wildfire vectors hourly. Think about the logic. A wildfire alert that arrives after you've downlinked, processed, and analyzed a mountain of imagery is a history lesson. One that's computed up there, and sent as a small message, is a tool. So the first compute in orbit earns its place by being close to the data. That's a business model that doesn't require beating a terrestrial data center on price. It requires beating the delay.

Zach Martin

Now bring in the giant. In the Q2 letter, Space Capital reports that in one quarter SpaceX sold AI compute to Anthropic and Google for roughly twenty six billion dollars annualized, more than doubling the revenue it took twenty three years to build. It will make AI chips with a venture called Terafab to power its orbital data centers. And it launched a twenty five billion dollar notes offering, following big tech into the debt markets to fund its AI buildout. Here's the careful reading. That's real revenue and real borrowing. It tells you the demand for compute is real. It does not tell you that orbital data centers are already economic. Space Capital's own framing is that space and terrestrial infrastructure are converging, pulled together by the two bottlenecks that govern AI: energy and compute. Both can be true: compute demand is enormous, and the specific orbital version still has a cost chasm to cross.

Zach Martin

That's why Anderson's January prediction reads so sensibly in hindsight, and I'll say plainly it's a prediction from an interested party. He wrote that 2026 would be a year of massive hype and modest experimentation, with true industrial scale deployment pushed to 2027. He told investors to temper expectations: high profile partnerships and pilot announcements, yes, but the hard thermodynamics would sort the field. Then look at what happened. The money flowed in anyway, and the second half of the chain, the physical cost curve, is exactly the part Dyer is spelling out. If you want a coaching rule: a narrative can get funded before a cost curve gets solved, but only the cost curve pays you back. So when someone pitches orbital compute, ask three questions. What's the launch cost assumption? What's the launch cadence assumption? And how do you reject the heat?

Zach Martin

Tie this back to capital flows, because that's the lens of the show. The orbital compute story explains part of why launch capital got repriced. If you believe that cost per kilogram to orbit and flights per year are the two dials, then every company that moves those dials gains an option on a very large market. That's the Launch Plus logic from earlier. It's also why the heaviest rockets matter. Space Capital's January piece argued that with mass constraints removed, heavy industrial depots and large scale data centers become economically viable, and it expected a buyer's market in launch as new vehicles like New Glenn and Neutron arrived. Those are forecasts. But notice the cleanest takeaway: the orbital compute thesis is, at its core, a bet on launch economics. If Dyer's five times and ten times conditions don't arrive, the data centers stay niche. If they do, a lot of the rest of the stack gets rerated.

Zach Martin

There's also an important counter intuition in all this. When the honest case says eight to ten times more expensive, it's tempting to hear that as a rejection. I hear it as a spec sheet. Engineers love a spec sheet. It says: here's the gap, here are the two levers, here's the rate you'd need. That's what separates a theory from a program. And it gives you a way to monitor the sector without getting swept up. Watch the launch cadence. Watch the price per launch. Watch for flight demonstrations of heat rejection at meaningful power levels. Watch for revenue from edge workloads like FireSat. And treat every announcement that skips those details as a mood, not a measurement. That's the discipline I'd want you to carry out of this chapter.

Zach Martin

The last mechanism is the one the Summit saved for the end, and Space Capital called it the mechanism underneath everything that came before. Think about what we've covered. NASA wants to buy services on a schedule. Defense wants capability at an affordable unit price. Compute wants launch cheap and frequent. All three share a quiet requirement: a lot of spacecraft, built repeatably, at a cost that doesn't depend on heroics. And Space Capital's write up says none of it scales on bespoke engineering. So how do you build satellites like products? The answer on the final stage was a word borrowed from chips: fabless.

Zach Martin

Here's the analogy, and it comes from the Summit, not from me. Semiconductors scaled when design decoupled from manufacturing. Apple designs its chips. TSMC fabricates them. Apple doesn't carry the burden of owning a fabrication plant, and TSMC amortizes its plant across many customers. The session argued the same separation is arriving in spacecraft. The company at the center is Eclipse Space, founded by Derek Huerta, a former Starlink engineering manager, who emerged from stealth the morning of the Summit. Space Capital is a seed investor in Eclipse, so again, you're hearing an investor describe a company it backs. On stage with him was Tom Whayne. And their pitch, in the write up's words, is that the fabless model turns satellites from bespoke projects into products with real unit economics.

Zach Martin

What's the argument against the obvious path? The obvious path is to copy SpaceX. SpaceX is vertically integrated, and it's the most successful space company in history, so why not do what it does? Huerta's answer, as summarized, is that owning a mega factory creates an anchor. A factory has high fixed overhead, and a factory that's built has to be kept busy. You need demand to amortize it, and if the demand isn't there organically, you end up manufacturing it, which is a polite way of saying you chase volume whether or not the market needs it. The line from the source that I like best is the compact version of the model: own the design and the supply chain, let someone else own the plant. That sentence is both a manufacturing strategy and a capital strategy. It says you can keep the intellectual property and the customer relationship without tying up billions in buildings.

Zach Martin

Let me make that concrete with a home analogy, labeled as mine. Imagine two people who both want to start a bakery chain. One builds a giant central bakery before they have customers. The other designs the recipes, the branding, and the supplier contracts, and uses existing ovens. The first one has more control. The second one has more flexibility, and a much lower break even point. If demand surprises on the upside, the first runs out of capacity and the second hires more oven time. If demand disappoints, the first has an idle plant, and the second just cuts orders. In an industry where demand has often been uncertain and politically driven, that flexibility is worth a lot. That's the logic, and it's why the Summit described the model as opening the market to any nation or company that wants orbital capability without building a factory first.

Zach Martin

What does Eclipse actually design? According to the Summit summary, the system architecture, the radio frequency payload, the direct to device communications, and the software, while outsourcing standardized manufacturing. The target customers are regional governments and commercial telecom enterprises that want to own and operate sovereign megaconstellations, and the pitch is that they can do so free of the export control restrictions known as ITAR, without sinking billions into domestic production infrastructure. I'd flag the qualifier here. These are positioning claims from a company that just emerged from stealth. We don't yet have a delivery record to measure, and the source doesn't give one. What we have is a model and a thesis. That's an early stage thing. So treat it as a hypothesis to watch, not a result.

Zach Martin

Still, the logic connects to everything else we've covered. Remember the spectrum point from the Iridium deal: spectrum is the scarcest asset, because nobody can make more of it. A nation that wants its own constellation has to secure spectrum, find customers, and build satellites. Fabless takes one of those hurdles, the factory, off the table, while leaving the ones that matter strategically in the owner's hands. And remember what Romo said about manufacturing: he's doing it for the second or third time and scaling up. Decoupling design from manufacturing is a way to make that repetition a service rather than an in house burden. If it works, it also changes where capital goes. Less into buildings, more into design talent, supply chain relationships, and software. That's an interpretation, and it depends on the model proving out, but it's the direction the Summit pointed.

Zach Martin

It also sits inside a wider consolidation story. Space Capital's Summit write up says consolidation across the three layers is accelerating, and it calls Rocket Lab's eight billion dollar Iridium deal, announced within two weeks of the Summit, the clearest example yet of Infrastructure buying Distribution to own the full stack. Fabless sits on the opposite side of the same coin. One strategy says, own everything from silicon to orbit, as SpaceX is doing. The other says, own the design and the customer and rent the factory. Both can win, in different markets. A vertically integrated giant optimizes for its own constellation. A fabless architect serves many customers who'll never build a factory. I'd expect the industry to run both models at once for quite a while.

Zach Martin

So let me pull the four mechanisms together, and then give you the practical homework. Mechanism one: NASA behaves as an anchor customer, buying services with firm fixed price milestones, and Lunar Outpost's two hundred twenty million dollar task order inside a four point six billion dollar ceiling is the working example. Mechanism two: defense buys maneuverability under prize based, affordability constrained procurement, where General Guetlein's twenty two million dollar per interceptor problem is the arithmetic that disciplines everyone. Mechanism three: orbital compute is real demand meeting an unsolved cost curve, with Dyer's eight to ten times gap, five times cheaper launch, and ten times more launches as the conditions. Mechanism four: fabless manufacturing turns bespoke satellites into products. Four ways of saying the same thing. The customers are changing how they buy, and the suppliers are changing how they build.

Zach Martin

Now the homework, and none of this is investment advice. It's a way to think. First, separate ceilings from backlogs. A four point six billion dollar ceiling is a ceiling. Watch for the task orders. Second, watch contract awards and test campaigns rather than press releases, which is the advice Holmes and Romo both gave. Third, ask the unit economics question of every company: what does one more unit cost, and who owns the fixed costs? Fourth, remember that the pace of capital already flattened a bit from the first quarter to the second, so selectivity is already in the data. And fifth, hold the public market side loosely. A stock can crater on a Tuesday and the contract is still the contract. The reverse is true as well. The assets that last are the ones with customers, spectrum, flight heritage, and a price that works.

Zach Martin

One last thing, and it's about the long game. In its Nasdaq announcement, Space Capital notes that GPS, built from twenty four satellites, generated trillions of dollars of economic activity, and that GPS took forty years to compound into the invisible backbone of the global economy. It says today there are about eighteen thousand satellites in orbit, and projects more than two million by 2040, driven by AI, geopolitics, and the orbital compute race. That's a projection, and a big one, and I'd hold it with some humility. But the pattern of compounding is the part I trust. Small, repeated, honest steps. A rover that teaches the next rover. A rendezvous that teaches the next rendezvous. A task order that earns the next task order. That's how a frontier becomes an economy. I'm Zach Martin. Thanks for spending the hour with me on SPACE CAPITAL, here on KMKT IR Hub Radio. Real assets, real ideas, and a brighter tomorrow. See you next time.