What would you build in space tech today?
In our space economy deck, you will find everything you need to understand the market
SUMMARY
We would build an orbital intelligence and autonomy company for defense and large satellite operators today, with software that detects unusual spacecraft behavior, predicts what happens next and helps operators decide how to respond.
Space startup funding is back, but it is not a broad return of the 2021 space boom. Capital is clustering around defense, satellite infrastructure, in-space mobility and companies that already have contracts, flight heritage or a believable path to production.
The buyer mix matters more than the headline size of the space economy. Upstream businesses still depend heavily on governments, so a startup that cannot sell into defense or civil procurement may discover that its practical market is much smaller than the top-down numbers suggest.
Launch is a poor default entry point because incumbents already have years of operational learning and large amounts of capital behind them. The better startup position is often one layer above launch, where every additional mission creates more demand for communications, navigation, maneuvering, servicing or orbital intelligence.
Satellite manufacturing is going through the same shift. Apex, K2 Space and Muon Space are industrializing production fast enough that another generic bus company needs a genuine discontinuity rather than a slightly better spacecraft.
Defense space stands out because procurement is moving toward proliferated architectures and multiple suppliers. That creates room for startups in surveillance, maneuvering, missile tracking, resilient navigation and other capabilities where the customer actively wants redundancy and new vendors.
Some of the strongest opportunities sit above increasingly commoditized hardware. Basic object tracking, ordinary imagery and standardized satellite buses are getting crowded, while behavior prediction, sensor fusion, interoperable optical links and trusted navigation remain harder.
Satellite servicing has crossed an important line from demonstrations into contracts. The cleaner startup wedges may still be docking hardware, rendezvous software, refueling interfaces and relative-navigation systems rather than financing a complete servicing spacecraft on day one.
The Moon is becoming a real customer market, but NASA is still creating most of that demand. Communications, navigation, power and ruggedized lunar infrastructure look more investable than mining because they solve needs attached to missions that are already being funded.
The weakest ideas are those that need several future assumptions to become true at once. Hyperscale orbital data centers, another commercial space station and large-scale lunar mining may all work eventually, but each asks a new entrant to finance too much before customer demand is proven.
The common pattern is straightforward: the most interesting companies benefit from a busier orbital economy without having to bankroll the entire economy themselves. We would rather own a difficult operational bottleneck than another capital-intensive platform whose success depends on winning a scale race against better-funded incumbents.
This market map, featured in our space economy deck, highlights top companies and startups in the space economy
Why are people building space startups again?
Space startups are attracting serious capital again, but the money is concentrating around defense, satellite infrastructure and businesses that already have a credible customer.
Seraphim Space’s latest quarterly index shows how abrupt the recovery has been. Private space-tech investment reached $23 billion over the latest trailing twelve months, compared with $9.7 billion over the equivalent previous period. The latest quarter alone brought in $7.5 billion across 141 deals. Deal count has not exploded at the same rate, which tells us investors are writing much larger checks rather than suddenly funding every space idea they see.
The companies raising the biggest rounds also show where investors think the real opportunities are. True Anomaly raised $650 million to expand its space-defense spacecraft and software. K2 Space raised $500 million after building a large-satellite business around commercial and national-security demand. Impulse Space raised another $500 million for in-space mobility. Just days ago, Muon Space closed a $250 million Series C after launching seven satellites in the first half of the year and opening a factory designed to produce as many as 500 satellites annually by 2027.
That mix is useful. Investors are paying heavily for companies that can manufacture, maneuver, communicate with, protect or operate satellites. Much less of the current funding boom depends on tourists, asteroid mining or some hypothetical mass consumer economy in orbit.
| Recent company | Recent financing | What the company is scaling |
|---|---|---|
| True Anomaly | $650M | Defense spacecraft and space operations software |
| K2 Space | $500M | Large, high-power satellites |
| Impulse Space | $500M | In-space mobility and orbital transfer |
| Muon Space | $250M | Satellite constellations and manufacturing |
| Apex | $200M+ | Standardized satellite buses |
Who actually pays for space tech today?
For an upstream space startup today, governments still matter far more than the size of the overall space economy might suggest.
ESA’s latest Space Economy Report puts the downstream space market at roughly €490 billion. That includes satellite communications, navigation and Earth-observation services embedded across the wider economy, and it is largely commercial.
The market looks completely different when we move upstream to spacecraft manufacturing and launch, which is where many founders imagine building a “space company.” ESA estimates that market at roughly €75 billion and says about 80% of demand is institutional, with defense now taking an increasingly important share.
That should influence the company we start. Building satellite hardware while assuming normal enterprise buyers will quickly replace government procurement is a dangerous plan. A much better setup currently is dual-use: defense or civil agencies can support the expensive first deployments, while commercial customers eventually make the market broader.
This also explains why apparently niche defense programs deserve more attention than enormous top-down estimates of the “space economy.” A startup needs someone who can sign the first $20 million or $100 million contract. In upstream space, that buyer is still very often a government.
As this chart shows, and as featured in our space economy deck, search interest in the space economy has been rising steadily
Would we start another rocket company now?
We would not start a conventional launch company today unless we had a technological breakthrough or a country willing to pay for independent access to space.
SpaceX has pushed the benchmark to an uncomfortable level for everyone else. The company flew 165 Falcon 9 missions in 2025. Its own prospectus reported approximately 620 orbital Falcon 9 launches by the end of the first quarter of 2026, with more than 99% mission success. More recently, Falcon passed 100 launches for the current year while individual boosters are flying thirty or more missions.
A newcomer therefore has to compete against a rocket whose operational learning curve is measured in hundreds of launches. At the same time, Blue Origin, ULA, Rocket Lab, Stoke and other well-funded companies are already chasing the remaining launch demand.
There are exceptions. Europe, India, Japan, South Korea and other countries have strategic reasons to support domestic launch even when the pure economics look worse than buying a Falcon flight. Responsive military launch can also justify different economics, and a genuine leap in reusability could reopen the market.
For a normal venture-backed founder, though, launch gives us the wrong starting position. We would rather build something that benefits every time another rocket flies.
If you want more recent data on this point, please see our latest space economy report.
Is there still room for another satellite manufacturer?
A new satellite manufacturer can still win, but “we build a better satellite bus” has become a much weaker startup pitch lately.
Look at how quickly the leading new manufacturers are scaling. Apex has raised three roughly $200 million financings in about fourteen months, with the latest taking its valuation to $2.3 billion. Its existing factory is designed around high-rate production, and the company is increasingly bringing subsystems in-house.
K2 Space has gone even bigger. Its latest $500 million round valued the company at $6.8 billion after it passed $1 billion in signed commercial and government contracts. K2 says it is preparing to manufacture up to 100 large satellites a year.
Muon Space is approaching the market from another direction. Instead of selling only the spacecraft, it combines mission design, satellites, payloads, software and operations. Its newly announced $250 million round takes total equity funding above $386 million. Muon says it has more than 50 satellites in development and is building toward 500 satellites of annual manufacturing capacity.
Those three companies alone are adding huge amounts of capacity while accumulating flight heritage and customer relationships. Another generic bus company would arrive behind competitors that have already spent hundreds of millions industrializing the product.
We would enter satellite manufacturing only around a clear discontinuity: much more onboard power, unusually fast production, radically cheaper payload integration, extreme maneuverability, a sovereign supply chain, or a form factor that enables missions existing buses cannot handle.
| Company | Recent financing | Manufacturing direction | What makes entry harder |
|---|---|---|---|
| Apex | $200M+ | High-rate standardized buses | Productized spacecraft with flight heritage |
| K2 Space | $500M | Up to 100 large satellites/year | Large, high-power platform already winning major contracts |
| Muon Space | $250M | Up to 500 satellites/year by 2027 | Full mission stack from spacecraft through operations |
This chart, featured in our space economy deck, illustrates yearly venture capital funding for space economy startups
Is defense space the best market for a startup right now?
Yes. Defense space currently has the strongest combination of large budgets, urgent problems and willingness to buy from newer companies.
One of the clearest recent examples is the Space Development Agency’s award of roughly $1.75 billion for 36 additional missile-warning, missile-tracking and missile-defense satellites. That is one procurement program adding dozens of spacecraft rather than a government commissioning another single exquisite satellite.
Space domain awareness is expanding in the same direction. The Space Force initially selected 14 companies for its Andromeda contracting vehicle, which will procure maneuverable reconnaissance and surveillance spacecraft around geosynchronous orbit. Within weeks, the contract ceiling was increased from $1.8 billion to more than $6.2 billion because the service expects greater demand.
What makes defense especially interesting for a startup is the growing number of places where the military deliberately wants several suppliers. Proliferated constellations, commercial imagery, alternative navigation, orbital maneuvering and space surveillance are all being procured through broader pools instead of being automatically handed to one traditional prime.
The opportunity still comes with long procurement cycles, security requirements and the risk of becoming dependent on one government program. Even so, if we are deciding where an expensive space technology can find its first serious customer these days, national security belongs near the top of the list.
If you want more recent data on this point, please see our latest space economy report.
Is space domain awareness already crowded?
Basic satellite tracking is crowded, but understanding what spacecraft are doing and helping operators react is becoming one of the better opportunities in space tech.
LeoLabs shows that customers will pay meaningful money for commercial orbital intelligence. The company finished 2025 with more than $60 million of contract awards across U.S., allied and commercial customers, while its U.S. government business grew 186% year over year. Its radar network now tracks more than 25,000 objects in low Earth orbit.
Slingshot Aerospace is moving further up the stack. It recently won a $69.2 million Space Force contract, the largest in its history, to provide AI-powered mission rehearsal and operational training. That followed an earlier $27 million award and years of work combining orbital data, simulation and software.
The evolution is fairly clear. Knowing that satellite A is at coordinate B is increasingly a commodity. Military and commercial operators want to know whether a spacecraft is behaving strangely, what it might do next, whether another object is a threat, which sensor should observe it and how their own satellite should respond.
The Space Force recently demonstrated how operational this is becoming during the SALITRE exercise, where a commercial satellite executed a real evasive maneuver in orbit during a simulated adversary scenario watched by multinational operators.
We would therefore look above the catalogue layer: behavior detection, predictive tracking, sensor fusion, automated tasking, maneuver recommendations and autonomous fleet operations. Those problems get harder as orbit gets more crowded. Good.
This chart, featured in our space economy deck, shows why SpaceX is leading in the space economy
Are laser links a real startup market in space?
Yes. Optical communications are unusually attractive right now because governments are spending billions on satellite networks that depend on laser links while interoperability is still not fully solved.
The Space Development Agency’s proliferated architecture needs satellites from different manufacturers to pass data between each other using optical communications terminals. SDA has created its own terminal standard because hundreds of spacecraft from multiple vendors eventually need to behave like one network.
The technology has progressed, but the latest GAO assessment still found a meaningful gap. As of the period reviewed, only two of the four initial satellite contractors had demonstrated a cross-contractor laser link in space, and both happened to use terminals from the same supplier. The next tranche uses several different optical-terminal suppliers whose designs had not yet demonstrated full on-orbit interoperability.
We would look at terminals that can reliably interoperate across spacecraft, pointing and acquisition hardware, optical ground terminals, network routing, automated testing and space-to-air or space-to-maritime links.
The larger architecture is already being purchased. A startup does not have to convince the Pentagon that satellite-to-satellite bandwidth is useful. It has to solve one of the technical problems preventing that network from working reliably.
Would we build a GPS alternative today?
We would build around resilient navigation today, but we would avoid financing another full GPS replacement unless our architecture was clearly better than the systems already moving toward deployment.
Xona has gone much further than a navigation concept. The company raised $170 million this year to scale its Pulsar low-Earth-orbit navigation constellation, opened a satellite manufacturing facility and has already demonstrated navigation signals from orbit. It is also working with companies such as Topcon, Furuno and Trimble around positioning and timing applications.
TrustPoint is taking a different technical route. The Space Force recently gave the company $4 million for an end-to-end GPS-independent demonstration involving four satellites and four ground stations. TrustPoint already has three demonstration satellites in orbit and is using C-band rather than the L-band frequencies used by traditional GNSS.
Those companies strengthen the case for resilient PNT while making another complete constellation less attractive to us. The more interesting opening may sit at the receiver and assurance layer.
A product could combine GPS, Galileo, Xona, TrustPoint, terrestrial signals and inertial sensors, then continuously tell an aircraft, drone, ship or autonomous machine which position estimate can actually be trusted. Timing infrastructure for telecom networks and data centers is another possible wedge.
Navigation becomes especially valuable when customers stop assuming GPS will always work. That change is happening now, which gives us a real customer problem rather than a speculative one.
If you want more recent data on this point, please see our latest space economy report.
This chart, featured in our space economy deck, illustrates yearly funding for space economy startups
Is Earth observation still worth entering?
Earth observation is still worth entering if we sell a hard-to-get answer. Another generic imaging constellation would be a poor place to start.
Planet demonstrates that satellite intelligence can now become a substantial business. Its latest full-year results showed $308 million of revenue, up 26%, while backlog jumped 79% to more than $900 million. More importantly, the company disclosed that the $64 million increase in annual revenue came primarily from defense and intelligence customers.
BlackSky shows a similar shift toward government intelligence. It generated $107 million of revenue in 2025 and finished the year with $345 million of backlog, up 32%. International customers have been buying both Gen-3 satellite capacity and recurring access to its intelligence platform.
There is still room for new sensors. The National Reconnaissance Office has deliberately widened commercial procurement beyond normal optical imagery into synthetic-aperture radar, hyperspectral, infrared and radio-frequency data. That gives technically differentiated sensing companies a route to a large buyer.
The harder question is whether a startup really needs a new constellation. For wildfire detection, maritime activity, infrastructure monitoring or military change detection, we would first test whether existing satellite feeds plus proprietary software solve the problem. We would put our own hardware in orbit only where the missing sensor or latency creates the advantage.
Owning satellites can create a moat, but it can also turn a good software business into a capital-intensive aerospace company before product-market fit has been proven.
Is satellite servicing finally a real business?
Satellite servicing has finally moved far enough beyond demonstrations that we would consider building in the market today.
Starfish Space is the clearest recent example. The company won a $52.5 million Space Development Agency contract to dispose of satellites at the end of their lives, followed shortly by a $54.5 million Space Force contract for another Otter servicing vehicle. It also has customers including SES and NASA.
Investors responded after the commercial and government work started to accumulate. Starfish raised roughly $110 million this year to execute contracted missions and expand production. The company says it now has five Otter customers after completing several demonstrations in orbit.
For years, satellite servicing had plenty of technical demonstrations and very few customers willing to pay for an operational mission. We can now point to actual purchases covering deorbiting, maneuvering and satellite servicing.
We would still be careful about building a full servicing spacecraft from day one. Docking hardware, relative-navigation sensors, autonomous rendezvous software, refueling interfaces and mission-planning tools may offer cleaner entry points.
But the underlying market has crossed a line. Operators are starting to spend real money on changing what happens to a satellite after launch.
If you want more recent data on this point, please see our latest space economy report.
This chart, featured in our space economy deck, compares the main business model options for Earth observation satellite operators
Can orbital logistics become a big space-tech market?
Orbital logistics has a credible path to becoming large because cheaper launch does not eliminate the need to move a spacecraft after the rocket drops it off.
Impulse Space is currently the strongest proof point. The company has flown its Mira spacecraft three times, says it has hundreds of millions of dollars in customer contracts and recently raised $500 million, taking total capital raised above $1 billion.
Its larger Helios vehicle is designed to move payloads rapidly from low Earth orbit toward higher-energy destinations. That can matter for geostationary deployments, military missions, lunar payloads and any customer that cares about reaching a particular orbit faster than a standard rideshare allows.
The interesting part is how this market interacts with cheaper launch. Very large rockets can make bulk transport to orbit cheaper while creating more payloads that still need to reach different destinations. In that world, orbital transfer looks closer to a last-mile problem.
We still would not fund a generic “FedEx in space” pitch based only on future launch volume. We would want a mission where the customer already puts a price on faster deployment, higher orbit, repositioning or maneuverability.
Defense provides several of those missions today. Commercial constellations may add many more if launch volume keeps rising.
Is the Moon finally a real market?
The Moon has a real market now, but NASA is still creating most of it.
NASA recently ordered four additional commercial lunar deliveries from Astrobotic, Firefly Aerospace and Intuitive Machines for nearly $600 million in total. Astrobotic received roughly $298 million for two missions, Firefly $144 million and Intuitive Machines $148 million. NASA says the awards bring its planned lunar surface deliveries across providers to 17.
That is more convincing than a one-off moonshot because the agency is buying repeated missions from several suppliers and asking them to reuse and improve lander designs.
Communications and navigation could become especially interesting. NASA’s Near Space Network procurement for geostationary-to-cislunar relay services carries a maximum potential value of $4.82 billion over its full term, with Intuitive Machines selected for the first award. Future landers, rovers and crews will need connectivity even when Earth is not directly visible.
We would therefore look closely at lunar communications, navigation, power, autonomous operations and ruggedized equipment. Each solves something required before a larger lunar economy can function.
Mining is much harder to justify today. A useful mineral deposit does not create a business by itself. Someone still needs to mine it, process it, transport it and pay enough for the material to cover the extraordinary infrastructure cost. We would rather sell picks and shovels to the government-funded lunar buildout than bet the company on lunar commodities.
This chart, featured in our space economy deck, shows revenue breakdown by customer segment in the space economy
Which space-tech ideas look too early right now?
Orbital data centers, another commercial space station and large-scale lunar mining are the three areas where we would currently resist the hype.
Orbital computing has become more credible at small scale. Processing satellite data before transmitting it to Earth can save bandwidth, and Planet is working with Google on an R&D project around space-based data centers. That is worth watching.
The hyperscale version still asks us to assume too many things at once: extremely cheap launch, huge power systems, efficient thermal management in vacuum, radiation-tolerant computing and economical hardware replacement. A terrestrial data center can swap a failed GPU without launching a rocket. That difference remains enormous.
Commercial stations have a different problem. NASA clearly wants them. It recently returned to industry for input on the next phase of its commercial-station strategy, while Axiom, Starlab, Blue Origin’s Orbital Reef and Vast are already years into development. Vast has even secured a NASA private astronaut mission to the ISS as it builds toward its own station.
A new station company entering today would have to catch those teams, develop human-rated hardware and spend heavily before knowing how much non-NASA demand really exists. Selling equipment or services to the station companies looks much more attractive.
Lunar mining sits even further out. NASA is now spending serious money on lunar logistics, but we still cannot point to a recurring industrial buyer for tonnes of lunar resources. For now, infrastructure around lunar activity has a much firmer business case than extracting commodities from the Moon.
So what would we actually build in space tech today?
We would build an orbital intelligence and autonomy company for defense and large satellite operators, focused on detecting unusual spacecraft behavior and helping operators decide what to do next.
The product would sit between the sensors watching orbit and the satellites being operated. We would ingest radar, optical, RF, telemetry and public orbital data, combine those sources into a continuously updated picture of nearby spacecraft, detect maneuvers or behavior that looks unusual, predict what those spacecraft could do next and recommend or automate a response.
A first customer could use the system to protect high-value satellites in geosynchronous orbit. Another could manage a proliferated low-Earth-orbit constellation. A third could rehearse how its fleet should respond to a hostile spacecraft approaching one of its assets.
The reason we choose this over another satellite bus or rocket is simple enough. More satellites create more objects to track, more maneuvers to interpret, more potential threats, more links to manage and more operational decisions. The software problem gets larger every time the hardware market succeeds.
Current procurement makes that thesis much stronger. Space Force spending is moving toward persistent surveillance and maneuverable spacecraft. Commercial orbital-intelligence companies are landing much larger contracts. Military exercises now include real evasive maneuvers by commercial satellites. At the same time, the government is deliberately buying from several suppliers rather than assuming one prime contractor will own the full stack.
We would make the first version as hardware-light as possible. Existing commercial sensors can provide much of the initial data. Proprietary sensors or spacecraft could come later if customers reveal a gap that cannot be solved any other way.
If our founding team were unusually strong in photonics, optical communications would be our second choice. A robotics and guidance team could make satellite servicing our strongest option instead. Resilient PNT and specialized Earth intelligence also remain attractive, particularly when the company can begin with receivers, software or data rather than financing an entire constellation.
The common thread is that we would build around the problems created by a much busier orbital economy. Launch capacity and standardized spacecraft are becoming more available. Knowing what is happening, moving data reliably, navigating under interference, maneuvering spacecraft and servicing them remain harder.
That is where we see the best space-tech startup territory today.
| Rank | What we would build today | Why we like it | Main risk |
|---|---|---|---|
| 1 | Orbital intelligence and autonomy | More satellites and more military activity make operations harder every year | Defense sales and strong emerging competitors |
| 2 | Optical communications infrastructure | Large funded constellations need interoperable high-bandwidth links | Difficult photonics engineering |
| 3 | Satellite servicing and autonomous rendezvous | Customers are finally paying for operational missions | Mission failure can be very expensive |
| 4 | Resilient PNT receivers and assurance | GPS interference creates an immediate customer problem | Competing standards and architectures |
| 5 | Specialized Earth intelligence | Government demand for actionable intelligence is growing quickly | Generic imagery is already crowded |
| 6 | Orbital mobility | Launch growth creates more need for post-launch maneuvering | Capital intensity and uncertain commercial volume |
| 7 | Lunar communications and navigation | Repeated NASA missions are creating early infrastructure demand | Commercial demand still depends heavily on government |
If you want more recent data on this point, please see our latest space economy report.
This chart, featured in our space economy deck, shows how satellite internet platform technology has evolved over time
OUR METHODOLOGY
This analysis asks where a new space-tech startup still has an attractive entry point today. We compare capital flows, customer demand, procurement activity, competitive intensity, technical bottlenecks, market maturity and the amount of infrastructure a new company would need to finance before reaching a paying customer.
We prioritized recent evidence because the structure of the space market is moving quickly. Funding rounds show where private investors are concentrating capital; contracts and procurement programs show where customers are committing real budgets; revenue, backlog, manufacturing expansion and operational deployments help separate businesses gaining traction from categories that are still mostly prospective.
No single funding round or government award determines a conclusion by itself. We gave more weight when several forms of evidence pointed in the same direction: for example, a large financing paired with signed contracts and production expansion, or a technical bottleneck paired with an active procurement program.
We also judge opportunities at the specific layer where a startup would compete. A crowded satellite-imaging market can still contain an opening in differentiated sensing or intelligence software, while an increasingly competitive spacecraft-manufacturing market can still create room for components, subsystems or capabilities that solve a new bottleneck.
Categories received less weight when the business case still depends on several future assumptions becoming true at the same time. That is why orbital data centers, new commercial space stations and large-scale lunar mining rank below markets where customers are already paying for surveillance, resilient navigation, in-space mobility, servicing or lunar infrastructure.
Key sources used for this analysis include Seraphim Space’s investment index, ESA’s 2026 Space Economy Report, the Space Development Agency’s Tranche 3 tracking-layer award, Space Systems Command’s Andromeda procurement, LeoLabs’ 2025 contract update, GAO’s assessment of SDA optical communications, Xona’s Series C and Pulsar expansion, Planet Labs’ financial results, BlackSky’s FY2025 results, Starfish Space’s servicing contracts and funding updates, and NASA’s latest commercial lunar-delivery awards.
In our space economy deck, we identify pain points entrepreneurs should prioritize
Related blog posts
- Who are the top investors in the space economy?
- Is the Space Economy growing now?
