The Internet Economy’s Evolution and What It Reveals About the Space Economy

As an Amazon Associate we earn from qualifying purchases.

Key Takeaways

• Early internet infrastructure took 15-20 years before generating meaningful commercial returns
• Platform models, not hardware, created the internet’s trillion-dollar companies by market cap
• Space economy investment patterns closely mirror the internet’s 1995-2005 buildout phase

The Infrastructure Nobody Noticed

In 1969, four computers exchanged data packets across telephone lines for the first time. The researchers at UCLA, Stanford Research Institute, UC Santa Barbara, and the University of Utah who built ARPANET weren’t building a commercial network. They were solving a specific defense communications problem, and nobody was calling it an economy for good reason: it wasn’t one yet.

That gap between technical capability and commercial reality defined the internet’s first two decades, and then it collapsed with remarkable speed between 1993 and 2000. That same pattern is now visible in the space sector, where technical capability has outpaced commercial infrastructure for sixty years and is only now beginning to close that gap. The analogy isn’t perfect, nothing that spans different technological eras ever is, but it’s close enough to be instructive in ways that other comparisons simply aren’t.

The internet didn’t become an economy by accident. It became one because a sequence of infrastructure investments, regulatory decisions, and entrepreneurial bets created the conditions for commerce. The same sequence is happening in space, just with rockets instead of fiber cables and orbital slots instead of IP addresses.

How the Internet Economy Actually Developed

ARPANET operated as a research and government network through the 1970s and most of the 1980s. Commercial activity was explicitly prohibited. The transition to a commercial network happened in stages, with the National Science Foundation lifting its acceptable use policy in 1991 and the World Wide Web becoming publicly accessible that same year after Tim Berners-Lee published its specifications at CERN.

What followed was not an orderly commercial development. It was a chaotic, overlapping rush of infrastructure investment, speculative bets, and genuine innovation that occurred faster than anyone predicted and collapsed just as quickly. Between 1995 and 2000, the Nasdaq Composite index rose from roughly 1,000 to over 5,000. Between March 2000 and October 2002, it fell back to around 1,100. Billions in market capitalization evaporated. Companies with genuine users but no path to revenue, like Pets.com and Webvan, closed entirely.

And yet the infrastructure that those investments built, the fiber optic cables, the server farms, the browser standards, the payment processing rails, all of it remained. Amazon survived. eBay survived. Google, which incorporated in September 1998, was barely touched by the bust because it hadn’t gone public yet. When the dust cleared, the companies that had built real services on top of the infrastructure were in a stronger competitive position than before, because their failed competitors had pre-funded the pipes.

This is the moment the space economy is approaching, not the crash, but the inflection point where infrastructure investment becomes dense enough that commercial services built on top of it become viable.

The Dot-Com Boom as a Case Study in Infrastructure Speculation

The dot-com boom is often remembered as a story of irrational exuberance and gullible investors. That memory is incomplete. A significant portion of the capital that flowed into internet companies between 1996 and 2000 went into physical infrastructure, specifically undersea fiber optic cables, terrestrial broadband networks, and data centers. Companies like Global Crossing, WorldCom, and Qwest collectively spent hundreds of billions of dollars laying fiber.

Most of those companies went bankrupt. Global Crossing filed for bankruptcy in January 2002 with $12.4 billion in debt. WorldCom’s accounting fraud collapsed the company in July 2002, at the time the largest bankruptcy in American history. The fiber they’d laid didn’t disappear when the companies did. It was acquired in bankruptcy proceedings at cents on the dollar, and those low-cost assets became the backbone of the modern internet. The companies that built the early 2000s internet economy, including Google, Amazon, and eventually Netflix, were renting bandwidth and rack space from assets that had been written down by 90% or more.

There’s a direct parallel in the current space economy. SpaceX has driven launch costs from roughly $54,500 per kilogram to low Earth orbit with the Space Shuttle to under $3,000 per kilogram with Falcon 9, and is targeting costs below $100 per kilogram with Starship. That cost compression is creating conditions where orbital deployment becomes economically viable for services that would have been laughably unprofitable at Shuttle-era prices. It’s the bandwidth price collapse of the late 1990s replaying in a different medium.

Rocket Lab built a small launch vehicle specifically to serve the emerging small satellite market and went public via SPAC in 2021. Planet Labs operates the world’s largest fleet of Earth observation satellites, around 200 Dove satellites, and generates revenue from selling satellite imagery and analytics to governments and enterprises. These companies exist because launch costs fell far enough to make their business models viable, and they’re the equivalent of early 2000s web companies that could only exist because bandwidth prices had dropped 99% from their mid-1990s peaks.

What “Picks and Shovels” Actually Means

During the California Gold Rush of 1849, the people who reliably made money weren’t the miners. They were the merchants who sold mining equipment, food, and supplies. Levi Strauss built a lasting business selling durable work pants. Sam Brannan made a fortune in general merchandising before most miners had staked a claim.

The internet economy followed a similar pattern. The companies that created the most durable value in the 1990s weren’t the ones chasing the gold; they were the ones selling the shovels. Cisco Systems built the routing equipment that directed traffic across the internet. Sun Microsystems sold the servers that ran web services. Akamai Technologies solved the content delivery problem, figuring out how to get web pages and media files to users faster by caching them closer to where people actually were.

Cisco’s revenue grew from $1.2 billion in 1994 to $18.9 billion in 2000. That’s not speculative. It’s picks-and-shovels economics at scale, and it happened because Cisco didn’t need to predict which websites would win. It needed to predict that more traffic would flow, which was easy.

The space economy is producing its own picks-and-shovels companies. Maxar Technologies manufactures satellite buses and geospatial intelligence products. Rocket Lab is expanding beyond launch into satellite manufacturing and space systems. Iridium Communications completed its second-generation constellation in 2019, a $3 billion infrastructure project, and now sells satellite phone and IoT connectivity services globally. These aren’t speculative bets on a distant future. They’re infrastructure businesses with recurring revenue, and they deserve more analytical attention than the flashier exploration-oriented companies that capture most of the media coverage.

Government as the First Customer

One aspect of the internet economy’s development that tends to get underappreciated in retrospect is how important government spending was in the early commercial years. It wasn’t just ARPANET and the NSF. The federal government was the internet’s first paying customer for commercial services, and that spending provided the revenue that kept early companies alive long enough to reach civilian markets.

Amazon Web Services signed a $600 million cloud computing contract with the CIA in 2013, a deal that represented roughly half of AWS’s annual revenue at the time. That contract, and the subsequent JEDI cloud contract won by Microsoft in 2019, validated cloud computing as a reliable enterprise-grade service in ways that private sector contracts alone couldn’t. Government procurement, in other words, de-risked the technology for the private sector.

The space economy is running this same playbook, explicitly and at scale. NASA‘s Commercial Crew Program paid SpaceXIntuitive Machines and Astrobotic Technology, is doing the same for lunar surface access. Intuitive Machines’ IM-1 mission landed on the Moon in February 2024, the first American lunar landing since Apollo 17 in 1972, and it was built and operated by a private company under a NASA commercial contract.

This pattern, government as anchor customer rather than sole operator, is the most direct policy parallel between the internet economy’s development and the space economy’s current moment.

Network Effects and the Space Economy’s Missing Flywheel

Network effects are what made the internet economy’s leading companies so durable. Facebook became more valuable to each user as more users joined. The eBay marketplace became more liquid as both buyers and sellers increased. Google‘s search quality improved as more people searched, because each search produced data that made the ranking algorithm better.

These are demand-side economies of scale, and they’re fundamentally different from the supply-side economies of scale that traditional industries produce. A steel mill becomes more cost-efficient as it scales. A social network becomes more valuable as it scales. The distinction matters because demand-side economies of scale create winner-take-most dynamics, where the leading platform tends to capture an overwhelming share of the market, leaving little room for viable competitors.

The space economy hasn’t yet produced a business with strong network effects. Starlink has some: a larger constellation can offer lower latency and more consistent service, which attracts more customers, which funds more satellites. But the core value proposition, providing internet connectivity, doesn’t improve as more people use it. It’s a capacity business, not a network effects business. The space economy’s first true network-effects business probably hasn’t been built yet, and when it arrives, it’ll likely be in data: a platform that aggregates satellite imagery and signals intelligence from multiple providers, layers analytical tools on top, and becomes more useful to each customer as more customers contribute insights back to the platform.

Platform Economics and the Space Economy’s Missing Layer

The internet economy’s second act, the one that produced Apple‘s App Store, Google‘s search advertising business, and the social media ecosystem, didn’t come from building better infrastructure. It came from building platforms on top of infrastructure that somebody else had already built and paid for.

Marc Andreessen and Jim Clark launched Netscape in 1994 with the browser as a platform layer between users and the raw internet. The browser didn’t own the cables or the servers. It owned the interface, and interface ownership turned out to be more valuable than infrastructure ownership. Amazon Web Services, which launched publicly in 2006, took a similar approach: rather than building applications for itself, it rented its excess capacity to developers and businesses at a price far below what those developers would have spent building their own infrastructure. By 2023, AWS generated $90.8 billion in revenue, more than the entire rest of Amazon’s business by operating profit.

The space economy’s platform layer barely exists yet. In 1994, you could use the internet to send email and access primitive websites. What you couldn’t do was buy something, watch a movie, hail a ride, or store your files. Those capabilities required a platform layer that took another decade to build, and the companies that built it became the most valuable on earth. Starlink, SpaceX’s satellite internet constellation, is potentially the first space economy platform at meaningful scale, serving over 4 million customers in more than 100 countries as of early 2025. But Starlink is selling a connectivity service, not a platform that other companies build on top of. The equivalent of AWS for the space economy, a service that turns orbital assets into developer-accessible infrastructure, hasn’t been fully built yet, and that gap represents the sector’s largest near-term commercial opportunity.

The Role of Regulation in Shaping Both Economies

The internet’s commercial expansion in the 1990s happened in a regulatory environment that was, by historical standards, remarkably permissive. The Communications Decency Act of 1996, despite its controversial content provisions, included Section 230, which protected internet platforms from liability for user-generated content. That legal protection made it possible for platforms like eBay, YouTube, and eventually Twitter to exist without employing armies of lawyers to review every post and listing. The Digital Millennium Copyright Act of 1998 created safe harbors for platforms hosting copyrighted content, which made the economics of user-generated platforms viable.

These weren’t neutral technical standards. They were deliberate policy choices that shaped which business models could be profitable, and the internet economy would have developed very differently, or perhaps not developed at all in its current form, without them.

The space economy faces an analogous regulatory moment. The Outer Space Treaty of 1967 established that no nation can claim sovereignty over celestial bodies, which creates legal uncertainty around resource extraction and property rights. The U.S. Commercial Space Launch Competitiveness Act of 2015 took a step toward resolving this by affirming that American citizens can own resources they extract from asteroids and the Moon, even if they can’t own the territory itself. Luxembourg passed similar legislation in 2017.

But orbital spectrum allocation, debris liability, and traffic management in low Earth orbit remain contested and largely unresolved. The International Telecommunication Union allocates orbital slots and radio frequencies through a first-come, first-served registration process that was designed for a world with a handful of satellites, not the 40,000-plus constellation that SpaceX has licensed for Starlink. Whether regulators can adapt fast enough to support continued commercial growth, without either stifling it with red tape or allowing it to create a debris field that makes low Earth orbit unusable, is genuinely uncertain. This is actually the space economy’s biggest structural risk, and it gets far less attention than rocket development schedules or funding rounds.

The Broadband Parallel: When Access Transforms Markets

In 1995, most home internet connections ran at 28.8 or 56 kilobits per second over telephone lines. Video streaming was impossible. Downloading a song took 30 minutes. The commercial internet was, as a result, a text-and-static-image medium. It was valuable for information retrieval and communication, but it couldn’t support media, commerce at scale, or real-time interaction. That wasn’t a permanent feature of the internet. It was a temporary constraint created by infrastructure limitations.

Broadband changed everything. When cable and DSL connections began reaching U.S. homes in significant numbers around 1999 to 2002, and when penetration crossed 50% of U.S. households around 2007, entirely new categories of business became viable. Netflix launched its streaming service in 2007, the same year U.S. broadband penetration crossed that threshold. That’s not coincidence. The service was waiting for the infrastructure.

In the space economy, the equivalent of broadband is launch cost reduction. When launch costs fall below a certain threshold, categories of commercial activity that were previously impossible become viable. The question isn’t whether this will happen; it’s already happening. The question is which specific services represent the “streaming video” of the space economy, the category that was always theoretically possible but required cheap and reliable access before it could be economically built.

Earth observation analytics is a strong candidate. In-space manufacturing of materials impossible to produce in gravity is another. Direct-to-device satellite connectivity, which AST SpaceMobile is building with its BlueBird constellation, could be the most disruptive of all, because it eliminates the infrastructure gap between terrestrial cellular networks and global coverage without requiring users to buy any new equipment.

The Data Economy Dimension

The internet’s most valuable output turned out not to be the services it enabled but the data those services generated. Google processes roughly 8.5 billion searches per day, and each search produces a data point. Those data points, aggregated and analyzed, create a real-time picture of human intent at planetary scale. Google’s ability to sell advertising against that signal generated $237.9 billion in revenue in 2023.

This wasn’t obvious in 1998 when Google launched, or in 2000 when the dot-com bust was underway. It became obvious only in retrospect, after the data flywheel had been spinning long enough to produce something unambiguous.

The space economy is beginning to generate its own data flywheel, and the scale is worth understanding. Planet Labs captures an image of the entire Earth’s landmass every day. That’s not a one-time snapshot. It’s a continuous, daily record of land use, agricultural conditions, construction activity, maritime shipping, and human settlement patterns globally. The economic value of that data, properly analyzed, is orders of magnitude larger than the value of the raw images. Spire Global collects GPS radio occultation data from its constellation of over 100 nanosatellites to generate atmospheric measurements that improve weather forecasting, and it sells that data to national meteorological agencies and commercial weather services. HawkEye 360 geolocates RF signals from space, identifying ship transponder anomalies, illegal fishing activity, and spectrum interference events. These are data businesses built on space infrastructure, and they’re early versions of what eventually becomes, if the internet analogy holds, the dominant commercial model of the space economy.

The “Killer App” Question

The internet had several watershed applications that drove mass adoption: email in the late 1980s and early 1990s, the web browser starting in 1994, online shopping with Amazon‘s launch in 1995 and eBay‘s launch in 1995, and eventually social networking with Facebook‘s launch in 2004. Each of these applications didn’t just use the internet. They redefined what it was.

The space economy hasn’t had its killer app yet, though Starlink comes closest. Providing high-speed broadband to rural and maritime users, including the roughly 3.5 billion people who live outside reliable terrestrial connectivity, is a genuinely mass-market value proposition. Starlink’s rapid subscriber growth, from zero to 4 million in roughly four years, suggests it has found real demand.

But Starlink is connectivity, not an application. The killer app of the space economy, the thing ordinary people will interact with daily without necessarily knowing it involves satellites, might be something nobody’s building yet. Or it might be something already being built that hasn’t reached scale. The best guess, and this is a genuine guess rather than a certain prediction, is that precision agriculture enabled by daily satellite imagery and AI analysis will be one of the first space economy applications to reach hundreds of millions of indirect users. The combination of Planet Labs‘ daily imaging and machine learning tools capable of detecting crop stress, predicting yields, and identifying disease outbreaks before they spread could deliver economic value to smallholder farmers in sub-Saharan Africa and South Asia that has no terrestrial alternative.

Comparative Development Timelines

The table below compares key development phases of the internet economy to their space economy analogs, illustrating how the timelines and dynamics map across both sectors.

The Investment Cycle: Familiar Patterns, Different Stakes

Venture capital investment in space companies totaled approximately $272 million globally in 2012. By 2021, that number had reached $14.5 billion, according to data compiled by Space Capital. The growth curve is visually indistinguishable from internet venture investment between 1994 and 1999, and that should trigger both excitement and caution in roughly equal measure.

The dot-com era produced genuine category-defining companies alongside hundreds of failed bets. The ratio of winners to losers in that era was roughly one to ten, maybe worse. The same ratio probably applies to the current space investment cycle. This doesn’t mean the money is being wasted. Most of the companies that fail will have built something, tested something, or hired engineers who then take that knowledge to the survivors. That’s how technological development actually works: inefficiently and wastefully by any individual company’s standards, but productively at the ecosystem level.

Andreessen Horowitz, which manages over $35 billion in assets, launched a dedicated American Dynamism fund in 2022 that explicitly includes space as an investment category. Bessemer Venture Partners has backed Satellogic, a high-resolution Earth observation company, among other space investments. The presence of top-tier generalist venture firms in space investing represents a maturation of the sector’s investment ecosystem that has no real precedent before about 2018.

Books That Shaped the Mental Models

Several books have influenced how analysts and investors think about technology economic development in ways that apply directly to the space economy.

Clayton Christensen’s The Innovator’s Dilemma describes how established companies, optimizing for their best customers with existing products, consistently miss disruptive entrants who start at the low end of a market. SpaceX, which began by targeting the small satellite launch market that established providers like Arianespace and United Launch Alliance weren’t serving cost-effectively, followed the Christensen playbook almost exactly, and the incumbents were slow to respond for exactly the reasons the book predicts.

Chris Anderson’s The Long Tail argued that internet economics enables profitable service of niche markets that physical distribution economics made impossible. The same logic applies to the space economy: when launch costs fall far enough, it becomes economically viable to build and operate satellites for markets too small to justify dedicated assets at 1990s launch prices. That’s the economic logic behind the entire small satellite industry, and it took an internet-era framework to articulate it clearly.

The Geography of the Space Economy

The internet economy’s geography is instructive. Early concentration in Silicon Valley gave way to secondary hubs in Seattle, Austin, New York, and eventually to a genuinely global distribution of technology centers. The factors that created Silicon Valley’s dominance, proximity to Stanford University research, venture capital availability, and a talent network built over decades, weren’t replicable on demand elsewhere, but they were partially replicable, and the partial replication was enough to create a genuinely global ecosystem.

The space economy is following a similar geographic evolution, and faster. SpaceX and Rocket Lab are U.S.-headquartered, but Rocket Lab operates launch facilities in New Zealand. OneWeb, a satellite broadband constellation company that went through bankruptcy in 2020 and was acquired by a consortium led by the UK government and Bharti Global, is British-headquartered. Europe’s Arianespace operates from the Guiana Space Centre in French Guiana. China’s CASC and CASIC, alongside private entrants, are building a parallel ecosystem largely decoupled from Western capital and regulatory frameworks.

This geographic distribution changes the regulatory and competitive dynamics significantly. The internet’s global expansion happened into a pre-existing framework of bilateral telecommunications agreements and national telecom regulators. The space economy is expanding into a framework that was designed for a handful of government operators and is straining under the weight of thousands of commercial satellites.

The Talent and Culture Dimension

The internet economy’s success depended on concentrating a specific kind of talent in a specific kind of culture. The engineers who built the early web worked in an environment that rewarded risk-taking, tolerated failure, and moved faster than established institutions. That culture was partly deliberate and partly emergent from the compensation structures, particularly equity-heavy packages, that startups used to attract talent away from established companies.

The space economy is replicating this culture more successfully than most observers expected. SpaceX famously operates on an iterative failure-and-learn model that more closely resembles a software startup than traditional aerospace. The controlled explosions of Starship prototypes during 2020 and 2021 were, from SpaceX’s perspective, exactly what the process was supposed to produce: fast learning at low cost relative to building one perfect vehicle and hoping it worked.

This cultural shift has diffused across the new space sector. Rocket Lab founder Peter Beck built the company around the principle that simplicity and reliability matter more than theoretical optimality. Relativity Space is attempting to 3D-print entire rocket structures, not because it’s the easiest path to orbit but because it believes the manufacturing innovation will compound over time in ways conventional approaches can’t match. These are software-industry thinking patterns applied to hardware problems, and the transfer is happening faster than skeptics predicted in 2012.

The Consumer Layer That Doesn’t Exist Yet

The internet economy has a consumer layer, a business-to-consumer layer, and an infrastructure layer, and they operate at very different margins and scales. The consumer layer, represented by companies like Netflix, Spotify, and Airbnb, generates relatively thin margins at very large scale. The infrastructure layer, represented by AWS, Google Cloud, and Microsoft Azure, generates thick margins at medium scale. The B2B software layer in between, represented by Salesforce, Workday, and ServiceNow, generates very thick margins at substantial scale.

The space economy currently has almost no consumer layer. Starlink is the closest thing to a direct-to-consumer space product, and it’s really an infrastructure service that consumers happen to buy directly. Space tourism through Blue Originexists but hasn’t demonstrated it can operate at commercially viable scale for most people.

The consumer layer of the space economy will exist eventually, and the most likely form it takes isn’t humans in space but satellite services integrated so deeply into consumer products that most users don’t know they’re interacting with space infrastructure. Global real-time navigation with GPS is already this: billions of people use it daily without thinking of it as a space product. Direct-to-device satellite messaging, which Apple integrated into the iPhone 14 in 2022 through a partnership with Globalstar, is the next step in that direction. AST SpaceMobile is taking it further still, building a constellation designed to provide full broadband connectivity directly to standard smartphones without any special hardware.

The Question of Market Size

The internet economy’s market size was dramatically underestimated in every analysis produced before 2005. McKinsey, Goldman Sachs, and virtually every credible research institution produced projections in the mid-1990s that turned out to be orders of magnitude too small. The reason wasn’t analytical failure. It was structural: the services that would generate most of the internet’s economic value hadn’t been invented yet, and they couldn’t be invented until the infrastructure was built.

The same structural limitation applies to space economy forecasts. Morgan Stanley projected the space economy could reach $1.1 trillion by 2040, up from roughly $370 billion in 2022. Bank of America projected $1.4 trillion by 2030. These numbers are likely wrong, but probably in the same direction that internet economy forecasts were wrong in 1997: too low, not too high. The services that will generate most of the space economy’s value by 2040 probably don’t exist yet in recognizable form, for the same reason that Uber, WhatsApp, and cloud storage didn’t exist in 1997. They couldn’t exist because the infrastructure wasn’t there, and the infrastructure is now being built at scale for the first time.

The Failure Rate Is Not the Story

The space economy will produce hundreds of failed companies over the next decade. Some have already failed: Virgin Orbit ceased operations in April 2023 after failing to complete a successful orbital launch from British soil. Momentusstruggled after regulatory and technical problems following its 2021 SPAC merger. Astra Space retired its Rocket 3 launch vehicle in 2022 after repeated failures to reach orbit reliably.

None of these failures indicate the space economy isn’t developing. They indicate it’s developing exactly as expected, with a high failure rate during the infrastructure-building phase that produces learning at the ecosystem level. Every failed internet company of the 1990s left behind engineers, code, business models, and market data that the survivors used. The same is true of failed space companies.

The analogy that matters here isn’t the dot-com boom and bust. It’s the railroad boom and bust of the 1840s and 1870s. American railroad companies went bankrupt repeatedly, and the railroads stayed. The land that had been granted to railroad companies, and the tracks themselves, were acquired in bankruptcy at low prices and operated profitably for decades. The infrastructure outlasted every company that built it, and the same pattern will hold in space.

Summary

The internet economy and the space economy are running the same playbook on different timescales, and the differences in timescale matter less than the structural similarities. Infrastructure precedes commerce. Commerce precedes platforms. Platforms precede the data economy. That sequence took roughly thirty years to play out on the internet, from ARPANET’s 1969 origins to Google’s 2004 IPO, and the space economy is compressing that timeline because the engineers building it can observe what happened with the internet and make better-informed bets.

What the internet analogy doesn’t capture is the physical constraint of space. The internet is infinitely expandable: adding a server adds capacity without displacing anything. Low Earth orbit is finite. The debris problem is real. The orbital slot allocation problem is real. These physical constraints have no internet equivalent, and they may force the space economy to evolve governance mechanisms that have no precedent in internet history, potentially slowing development in ways that no historical comparison fully prepares investors and policymakers to anticipate.

The companies building picks-and-shovels infrastructure today are the Ciscos and Akamais of the space economy, and history suggests they deserve far more attention than they currently receive. The loudest bets in any infrastructure boom tend to be placed on the applications layer, on the things that will be built once the pipes are in place. But the pipes themselves have a way of generating the most durable returns.

Appendix: Top 10 Questions Answered in This Article

What is the internet economy and how did it evolve from its origins?

The internet economy evolved from ARPANET, a U.S. Defense Department research network launched in 1969, into a global commercial infrastructure after the National Science Foundation lifted its acceptable use policy in 1991. Commercial activity accelerated rapidly between 1995 and 2000, collapsed in the dot-com bust of 2000-2002, and rebuilt into the platform-driven economy that generated trillions of dollars in value by the 2010s. Each phase created infrastructure that the next phase exploited.

What caused the dot-com bust and which companies survived it?

The dot-com bust was triggered by a combination of rising interest rates, exhausted venture capital, and the collapse of companies with no path to profitability. Amazon, eBay, and Google survived because they had built real services on top of shared infrastructure. The fiber optic networks and data centers built during the boom were acquired at low prices in bankruptcy and became the backbone of the modern internet economy.

How does SpaceX’s launch cost reduction parallel the internet’s bandwidth price collapse?

SpaceX reduced launch costs from roughly $54,500 per kilogram with the Space Shuttle to under $3,000 per kilogram with Falcon 9, a decline that mirrors the 99% collapse in bandwidth prices between 1998 and 2003. Both cost compressions created the economic conditions for entirely new categories of commercial services that had been technically possible but financially unviable at previous price points, unlocking markets that couldn’t exist before.

What is the space economy’s platform layer and why hasn’t it been fully built yet?

The internet’s platform layer, services like AWS and Google’s advertising system built on top of shared infrastructure, created the most valuable internet companies. The space economy’s equivalent, a service that turns satellite assets into developer-accessible infrastructure, is only beginning to emerge in the mid-2020s because the underlying launch and satellite infrastructure wasn’t affordable enough to build on until recently. This gap represents the sector’s largest near-term commercial opportunity.

Which space companies are most analogous to early internet picks-and-shovels businesses?

Rocket Lab (launch access and satellite manufacturing), Maxar Technologies (satellite buses and geospatial data), Iridium Communications (satellite connectivity infrastructure), and Planet Labs (Earth observation data) are the space economy’s equivalents of Cisco, Akamai, and Sun Microsystems. They generate infrastructure revenue regardless of which end-service companies ultimately win their respective markets.

How is government procurement shaping the space economy’s commercial development?

NASA’s Commercial Crew Program paid SpaceX and Boeing to develop crewed spacecraft under commercial contracts, transforming government R&D into a commercial market. The Commercial Lunar Payload Services program contracted with Intuitive Machines and Astrobotic Technology for lunar delivery services. This anchor-customer model mirrors how CIA and federal cloud contracts helped validate Amazon Web Services as enterprise-grade infrastructure in the early 2010s.

What role do network effects play in the current space economy?

The internet’s leading companies were built on demand-side network effects where value increased as user numbers grew, creating winner-take-most dynamics. The space economy hasn’t yet produced a business with comparably strong network effects. The most likely candidate for the space economy’s first network-effects business is a data aggregation and analytics platform combining imagery and signals from multiple satellite operators, similar in structure to Google’s early search data flywheel.

Why are space economy market size forecasts likely to be underestimating eventual scale?

Morgan Stanley projected the space economy could reach $1.1 trillion by 2040. Mid-1990s internet economy projections were similarly ambitious but turned out to be dramatically lower than reality because the most valuable future services, including cloud computing, social networks, and streaming, hadn’t been invented yet. The space economy’s most valuable services by 2040 likely haven’t been built yet either, for the same structural reason.

What is the significance of geographic diversification in the space economy?

The space economy is developing a global geographic distribution similar to the internet economy’s evolution from Silicon Valley concentration to a worldwide ecosystem. Companies including OneWeb (UK), Arianespace (France), and emerging Chinese private launch providers are creating multiple independent development centers. This diversification changes regulatory dynamics and reduces the risk that any single government’s policy decisions can constrain the sector’s overall growth trajectory.

How does the space economy’s failure rate compare to what happened during the dot-com era?

Virgin Orbit, Astra Space, and Momentus are among notable space economy failures since 2021, consistent with the dot-com era’s approximately one-in-ten success ratio for early-stage companies in a new infrastructure sector. These failures follow the pattern of 1840s and 1870s American railroad booms, where company-level bankruptcies didn’t prevent infrastructure-level success because the assets built by failing companies were acquired and operated profitably by successors.