From Perfect Competition to Monopoly: The Natural Evolution of Dot-Com Markets

A new internet market opens. Entrepreneurs enter quickly, capital follows, and no one is certain which company – or technology – will prevail. Then the field narrows. Firms disappear or are acquired while a handful achieve meaningful scale. Search, social networks and digital marketplaces have all followed versions of this progression.

Online delivery provides a more recent example. Cheap capital and pandemic-era demand drew billions of dollars into companies promising ever-faster delivery. Funding for ultrafast delivery peaked at $7.8 billion across 46 deals in 2021, then fell 72 percent the following year as firms closed, retreated from markets or were acquired. More than $12 billion had been invested in the category overall, yet only a small number of operators achieved durable scale. Broad entry again produced enormous experimentation without producing equally broad maturity.

The explanation is not simply monopoly, market failure or superior management. Dot-com markets remain governed by supply, demand, marginal cost, capital allocation and barriers to entry. What changes is how quickly those forces compound. Conventional companies usually encounter physical friction as they grow. Internet-native firms can often expand customer relationships before reproducing comparable physical capacity in every market.

Additional users can improve the economics of the business itself. They strengthen networks, spread fixed costs and make intangible assets more productive across a larger customer base. Scale is therefore not merely the result of competition. It becomes a competitive asset.

Management may rationally sacrifice current margin for market share because greater scale can improve future economics. The same process can lower transaction costs and raise productivity beyond the firm. Later-integrating economies may also confront already-scaled global competitors rather than recreate the experimentation of the early commercial internet.

Dot-com markets often begin with broad competition and end with narrow concentration because scale changes the economics of participation. Silicon Valley provides the longest mature record of the pattern, but the underlying economics are global.

Early Dot-Com Market Formation and Survival

Sources: Kirsch & Goldfarb

The Economics of Dot-Com Scale

Perfect competition is useful as a benchmark even though few real markets satisfy its assumptions. New internet markets can nevertheless begin with unusually open experimentation because companies can test products without first constructing extensive physical distribution.

Scaling changes the equation. Digital systems can require substantial upfront investment while the incremental cost of serving another customer remains comparatively small. As fixed costs are spread across more users, average costs can fall, leaving smaller competitors to support comparable product and security functions across a much narrower base.

Increasing returns deepen the advantage. Intangible assets can be reused, while network effects can make the product more useful as participation rises. More riders attract more drivers; more buyers attract more sellers. Multi-sided platforms can also subsidize one group because attracting it raises the value of another.

Growth therefore strengthens competitive position as well as revenue. Market share becomes an input when additional users improve network usefulness or infrastructure efficiency. Distribution can reinforce that cycle. The Microsoft browser case showed how distribution supported an existing software ecosystem, while the Google search litigation later placed default distribution and rival access at the center of competition. Google had accounted for roughly 90 percent of U.S. searches for years before remedies addressed exclusive arrangements and rival access.

Cloud computing shows concentration and growth advancing together. Global enterprise cloud infrastructure revenue reached about $143 billion in the second quarter of 2026, up 43 percent in a year, while AWS, Microsoft and Google together controlled roughly 63 percent of the market. Quarterly revenue had been only about $8 billion a decade earlier. The market expanded nearly eighteenfold without dispersing leadership across a comparable number of firms.

None of this makes incumbents permanent. Netscape gave way to Internet Explorer, which later lost browser leadership. MySpace yielded to Facebook. Digital markets can reopen, but scale can still leave relatively few viable participants. The decisive question is whether a superior competitor can reach consequential scale.

Scale, Growth and Concentration in Digital Infrastructure Markets

Sources: Synergy Research Group

Managing the Dot-Com

For CEOs and CFOs, the difference is often sequencing. A conventional business may expand after establishing profitability at smaller scale. A dot-com may pursue scale first because management expects a larger business to have better economics.

Amazon’s early strategy illustrates the logic. Sales reached $147.8 million in 1997 after rising 838 percent in a year, while customer accounts expanded from roughly 180,000 to 1.51 million. Distribution capacity grew from about 50,000 to 285,000 square feet. Digital scaling did not eliminate physical investment; it allowed demand and geographic reach to expand before replicating a traditional retail footprint.

Growth that improves future economics can be investment. Growth that merely enlarges an uneconomic business remains a bad business. A discounted ride may earn little today but acquire users and increase network density, strengthening tomorrow’s service.

The CFO question is what future economic position the expenditure purchased.

Capital can become competitive strategy because it finances the time and scale required to strengthen market position. First, second and fourth place may therefore have very different economics if smaller networks fall below the scale needed to reinforce themselves.

Acquisition accelerates the process. An FTC study identified 616 transactions of at least $1 million by Alphabet, Amazon, Apple, Facebook and Microsoft between 2010 and 2019 that fell outside ordinary HSR premerger reporting; 65 percent were valued between $1 million and $25 million. Acquisition can deliver rational returns while the market still consolidates.

Amazon and Walmart ultimately demonstrate the distinction. Both became infrastructure-intensive giants, but their routes to scale differed. By maturity, the next entrant confronts a market the original winner never faced.

The barrier can build itself.

Scaling Economics in Practice

*Among transactions for which target age was available.

Economic Multipliers, Prosperity and Regional Integration

Scaled internet businesses affect far more than their own revenue. Mobile technologies and services generated about $6.5 trillion of economic value in 2024, equivalent to 5.8 percent of global GDP, with more than $4 trillion arising from broader productivity effects. Much of the benefit therefore appeared outside the firms operating the networks.

Prosperity can coexist with dependency. A mobility platform may give a driver access to customers and payments while leaving that worker reliant on very few intermediaries.

Regional timing changes the path. Early digital economies experienced the full sequence from experimentation through consolidation and regulation. Later-integrating economies inherit mature technologies and already-scaled firms, compressing the competitive phase. Developing economies attracted about $531 billion in announced digital-economy greenfield investment over the five years through 2024, yet nearly 80 percent of Global South digital projects went to only ten economies.

The infrastructure gap is sharper still. Developing economies received only about $9 billion for core ICT infrastructure in 2024 against an estimated annual requirement of roughly $62 billion. Digital adoption can therefore advance much faster than local ownership of the capital-intensive systems underneath it.

Leapfrogging can accelerate both prosperity and concentration, but interoperability offers another path. Brazil’s Pix reached 114 million users within fifteen months, equal to 67 percent of adults, while merchant charges averaged about 0.22 percent compared with roughly 2.2 percent for credit cards. India’s UPI provides corroboration at larger scale, processing about 23.2 billion transactions across 720 banks in May 2026.

Shared infrastructure allows network effects to accumulate without requiring every participant to build a proprietary network. Regional integration therefore turns on timing, mode of integration and ownership of the underlying infrastructure.

Technology travels globally, but business models localize. A country may gain productivity and financial access while much of the platform profit and infrastructure value accumulates elsewhere. The development question becomes not only who connects, but who captures the value created by connectivity.

Digital Prosperity, Investment and Infrastructure

Sources: GSMA Intelligence, UN Trade and Development

Governance and the Economics of Contestability

Once platforms become economically important, governments must decide how concentrated markets should be managed. Europe has moved further toward ex ante gatekeeper obligations, while the United States has historically relied more heavily on competition followed by enforcement. Developing economies may place greater immediate weight on investment and connectivity, while China combines commercial competition with broader state objectives.

These are tendencies rather than clean ideological categories. Behind them sits a common economic test: can a materially better competitor still enter and reach enough scale to matter?

That is contestability.

Concentration can coexist with genuine efficiencies if challengers can reach customers and obtain the inputs needed to expand. Concern becomes sharper when accumulated advantages in distribution or infrastructure prevent superior entrants from scaling.

The policy question is therefore not simply how large a dot-com should become. It is whether the benefits of scale coexist with a credible path for displacement.

The Next Cycle

Artificial intelligence is recreating the same tension. The cost of querying a model capable of roughly GPT-3.5-level performance fell from about $20 per million tokens in late 2022 to $0.07 by late 2024. Capability became dramatically cheaper to access.

The size required to reach useful performance fell as well. In 2022, the smallest model exceeding a 60 percent score on the MMLU benchmark contained roughly 540 billion parameters. By 2024, a model with about 3.8 billion parameters could cross the same threshold, a reduction of more than 140-fold.

At the frontier, however, infrastructure and capital are becoming more consequential. Industry produced more than 90 percent of notable AI models in 2025, while private investment expanded rapidly. The infrastructure beneath AI is concentrating even as the cost of building useful applications falls.

The contradiction is familiar: the cost of starting can fall while the cost of becoming consequential rises.

New developers and open systems can reopen competition, while capital-intensive infrastructure can reinforce concentration beneath them. The outcome remains unsettled.

The internet created an environment in which familiar economic forces operate faster, reach farther and reward scale more strongly. A new market may still begin with dozens of firms and leave only a few at sustainable scale without any single decision producing that structure.

The consequential question for the next generation of dot-com markets is whether the next superior company can still become large enough to matter.

AI Entry Costs and Frontier Scale

Sources: Stanford HAI

TL;DR Summary

• Dot-com markets can begin with broad entry but narrow as scale changes the economics of competition.
• Increasing returns and network effects can make market share a competitive input.
• Capital and acquisition can reinforce firms already moving toward viable scale.
• Concentration can coexist with productivity gains and wider economic prosperity.
• Regional timing and infrastructure ownership influence whether scale becomes concentrated or interoperable.
• Contestability is a more useful policy test than concentration alone.
• AI lowers entry costs while increasing the importance of capital-intensive infrastructure.
• Technological disruption can reopen markets, but entrants still have to become large enough to matter.

From Open Competition to Concentration

• Institute of Internet Economics; Dot-Com Rules; – Link
• Kirsch & Goldfarb; Small Ideas, Big Ideas, Bad Ideas, Good Ideas: Get Big Fast and Dot Com Venture Creation; – Link
• Goldfarb, Pfarrer & Kirsch; Searching for Ghosts: Business Survival, Unmeasured Entrepreneurial Activity and Private Equity Investment in the Dot-Com Era; – Link

The Economics of Dot-Com Scale

• Harvard Business Review; Increasing Returns and the New World of Business; – Link
• American Economic Association; Systems Competition and Network Effects; – Link
• Journal of the European Economic Association; Platform Competition in Two-Sided Markets; – Link
• Synergy Research Group; Q2 Cloud Market Passes $143 Billion, Highest Growth Rate in Eight Years; – Link

Managing the Dot-Com

• Amazon; Amazon’s Original 1997 Letter to Shareholders; – Link
• Federal Trade Commission; Non-HSR Reported Acquisitions by Select Technology Platforms, 2010–2019; – Link
• Walmart; Fiscal 1997 Form 10-K; – Link
• OECD; Acquisitions and Their Effect on Start-Up Innovation: Stifling or Scaling?; – Link

Economic Multipliers, Prosperity and Regional Integration

• GSMA Intelligence; The Mobile Economy 2025; – Link
• UN Trade and Development; World Investment Report 2025: International Investment in the Digital Economy; – Link
• Bank for International Settlements; Digital Payments as a Boon to Financial Inclusion; – Link
• National Payments Corporation of India; Unified Payments Interface Product Statistics; – Link

Governance and the Economics of Contestability

• European Commission; About the Digital Markets Act; – Link
• U.S. Department of Justice; Department of Justice Wins Significant Remedies Against Google; – Link
• U.S. Department of Justice; U.S. v. Microsoft Corporation, Browser and Middleware Findings of Fact; – Link
• National Development and Reform Commission of China; Opinions on Promoting the Standardized, Healthy and Sustainable Development of the Platform Economy; – Link

The Next Cycle

• Stanford Institute for Human-Centered Artificial Intelligence; AI Index 2025: State of AI in 10 Charts; – Link
• Stanford Institute for Human-Centered Artificial Intelligence; The 2026 AI Index Report; – Link
• Epoch AI; LLM Inference Prices Have Fallen Rapidly but Unequally Across Tasks; – Link
• International Energy Agency; Energy and AI; – Link

Keywords: Internet Economics, Digital Markets, Market Concentration, Network Economics, Increasing Returns, Platform Contestability, Digital Market Structure