Bitcoin promised to eliminate banks. Blockchain promised to eliminate intermediaries. Web3 promised to return power to users.
Fifteen years after Bitcoin's whitepaper. Five years after DeFi summer. Three years after NFT mania.
What exists: new banks, new intermediaries, new platforms. Different branding. Same concentration.
The revolution got funded by venture capital. Then became what it opposed.
What Decentralization Promised
Satoshi Nakamoto's Bitcoin whitepaper (2008): "A purely peer-to-peer version of electronic cash would allow online payments to be sent directly from one party to another without going through a financial institution."
No trusted third party. No central authority. No banks.
The blockchain: distributed ledger across thousands of computers. No single point of control. No entity that can freeze accounts or reverse transactions.
Web3 followed: own your data, control your identity, capture value you create. The platforms—Facebook, Google, Twitter—extract value. You generate content. They collect revenue. Web3 promises to flip this.
The vision: disintermediate banks, corporations, governments. Distribute power to individuals. Eliminate the extractive middlemen.
Code is law. Math is truth. Decentralization is freedom.
That was the promise.
The Mining Concentration
Bitcoin mining: anyone with computer could mine. Early days, this was true. CPU mining, GPU mining. Individuals participated.
Then: ASICs (Application-Specific Integrated Circuits). Specialized hardware designed only for mining. Cost: $3,000-$10,000 per unit. Efficiency: orders of magnitude better than GPUs.
Individual mining became economically impossible. Can't compete with industrial operations running thousands of ASICs.
Mining pools emerged. Miners combine hash power, split rewards. Makes sense for individuals. Concentrates power.
Current state (2026):
- Foundry USA: ~30% of Bitcoin hashrate
- AntPool: ~20%
- F2Pool: ~15%
- ViaBTC: ~10%
- Remaining pools: ~25%
Four pools control 75% of Bitcoin mining. Two pools could coordinate 51% attack (theoretical ability to reverse transactions, double-spend).
Geographic concentration: Mining follows cheap electricity. China dominated until 2021 ban. Now: United States (35-40%), Kazakhstan (18%), Russia (11%).
The "distributed" network controlled by handful of entities in handful of locations. Distributed hardware. Concentrated control.
Individual participation: dead. Mining profitability requires:
- Industrial-scale operations (thousands of machines)
- Cheap electricity (under $0.04 per kWh)
- Cooling infrastructure
- Technical expertise
- Continuous hardware upgrades
Average person can't compete. The network runs on large operators. Decentralization didn't scale.
The Exchange Bottleneck
Most people don't hold their own Bitcoin. They use exchanges.
Coinbase: 108 million verified users. Binance: 150 million+. Kraken: 10 million+. The centralized platforms hold the crypto.
"Be your own bank" became "trust Coinbase to be your bank."
Why? Self-custody is hard. Lose private keys, lose funds permanently. No password reset. No customer service. No recourse.
Users choose convenience over sovereignty. Rational choice. Most people can't secure digital assets properly.
The exchanges:
- Require KYC/AML (Know Your Customer, Anti-Money Laundering)
- Can freeze accounts
- Can block transactions
- Report to governments
- Subject to regulations
This defeats crypto's original promise. No anonymity. No permissionless transactions. No censorship resistance. Just traditional financial system with crypto branding.
Sam Bankman-Fried's FTX: centralized exchange, customer funds misappropriated, $8 billion missing, fraud charges, bankruptcy. The "decentralized" revolution's second-largest exchange operated like traditional bank. Same fraud. Worse oversight.
The on-ramps and off-ramps (converting fiat to crypto, crypto to fiat): all centralized. Can't use crypto at scale without these chokepoints. The chokepoints are regulated, controlled, monitored.
Trading volume: 90%+ happens on centralized exchanges. Decentralized exchanges (DEXs) exist. Most users don't use them. Slower, more complex, less liquid.
The bottleneck is permanent. Users need easy way to buy, sell, store crypto. Centralized exchanges provide this. Decentralized alternatives can't compete on user experience.
The Venture Capital Capture
Web3 startups raised billions from venture capital.
Andreessen Horowitz (a16z) crypto fund: $7.6 billion across multiple funds. Paradigm: $2.5 billion. Sequoia, Tiger Global, Pantera—every major VC has crypto fund.
The decentralization revolution funded by Silicon Valley's power structure.
Token distributions: 30-50% allocated to team, VCs, advisors before public. Vesting schedules: unlock over 2-4 years. Team and VCs get tokens at fractions of public price.
Example token distribution (typical):
- Team: 20%
- VCs/investors: 20%
- Foundation/treasury: 20%
- Community/ecosystem: 20%
- Public sale: 20%
"Community" tokens often controlled by foundation, which is controlled by team. "Ecosystem" grants go to aligned projects. The 40% marked for public access shrinks.
Public buys at launch. Insiders bought 12-24 months earlier at 90% discount. Vesting unlocks. Insiders sell. Price crashes. Public holds bags.
This happened with hundreds of tokens. The pattern is consistent.
"Decentralized" protocols governed by foundations. Foundations run by teams funded by VCs. The cap table determines governance, just like traditional startups. Tokenization doesn't change power dynamics. It obscures them.
The revolution needed funding to scale. Funding came from existing power structure. Power structure captured the revolution.
The Ethereum Staking Concentration
Ethereum switched from Proof of Work to Proof of Stake (2022). "The Merge."
Proof of Stake: validators lock up ETH, get selected to propose blocks, earn rewards. More ETH staked = more likely to be selected = more rewards.
Running validator requires: 32 ETH (currently worth $60,000-$100,000+), technical knowledge, reliable internet, computer running 24/7.
Most people can't afford 32 ETH. Solution: staking pools. Give your ETH to pool, pool runs validators, you get proportional rewards minus fees.
Current staking distribution:
- Lido: ~30% of all staked ETH
- Coinbase: ~15%
- Kraken: ~8%
- Binance: ~6%
- Others: ~41%
Top 4 entities control ~59% of staked ETH. Lido alone controls nearly one-third.
Lido is "decentralized" staking pool. Really: collection of node operators selected by Lido DAO. The DAO is controlled by LDO token holders. LDO token distribution: ~64% to team, VCs, and Lido DAO treasury at launch.
The "decentralized" staking pool controlled by insiders who allocated themselves majority of governance tokens.
Staking rewards compound. The rich get richer algorithmically. Those with most ETH earn most rewards. Use rewards to stake more. Earn more rewards. The concentration increases over time.
Proof of Stake is plutocracy by design. Wealth equals power. More wealth equals more power equals more wealth.
This was the upgrade that made Ethereum "more decentralized."
The NFT Platform Problem
NFTs promised: own your digital art. No platform can take it away. It's on blockchain. Permanent.
Reality: OpenSea dominated 80-90% of NFT trading volume at peak. Centralized platform. Can delist NFTs. Can ban users. Can change fees. Did all of these.
The NFT itself: token on blockchain containing URL or hash pointing to image. The actual image: stored somewhere else. Usually:
- Centralized servers (AWS)
- IPFS (InterPlanetary File System)
- Arweave (decentralized storage, but costs money to maintain)
IPFS sounds decentralized. Is distributed file system. But: files only available if someone "pins" them (keeps copy running). Most NFT projects use IPFS gateway services (Pinata, NFT.Storage). These are centralized.
Gateway goes down, file unavailable. Project stops paying pinning service, file disappears. The blockchain has permanent record of token pointing to URL. The URL returns 404.
Bored Ape Yacht Club briefly went offline (2022). Yuga Labs' server issues. $1 billion+ in NFT value displayed as broken links. Problem fixed after hours. But demonstrated: "ownership" depended on company maintaining server.
True decentralized storage exists (Arweave, Filecoin). Most projects don't use it. Costs more. More complex. Centralized solutions are easier.
You "own" the NFT. Meaning: you own pointer on blockchain. The thing pointer points to: stored on someone else's server. Server goes down, your $100,000 jpeg is gone.
"Own your art" meant "own a receipt."
The DAO Governance Illusion
DAOs (Decentralized Autonomous Organizations): crypto promised new governance model. Token holders vote on proposals. Democratic. Transparent. Decentralized.
Reality: token-weighted voting. 1 token = 1 vote. More tokens = more votes. Plutocracy formalized in code.
Typical distribution: top 10 token holders own 50-70% of supply. These holders: often founders, VCs, early investors. The "decentralized" organization controlled by those who allocated tokens to themselves.
Voter participation: typically 5-15% of token holders vote. Of those voting: large holders dominate. Small holders' votes are rounding errors.
Proposals that benefit large holders pass. Proposals that benefit small holders fail. This is predictable math, not conspiracy.
Example: Uniswap governance proposal to create Uniswap Foundation, allocate $74 million in UNI tokens. Passed with strong support. Who benefited? The insiders running the foundation. Who paid? UNI holders through dilution.
Many proposals are "governance theater." The outcome is predetermined by large holders' preferences. Voting happens for appearance of legitimacy.
Some DAOs use "conviction voting" or "quadratic voting" to reduce whale dominance. These are marginal improvements. The fundamental problem remains: tokens are distributed unequally, voting power follows token holdings, those with most tokens control governance.
"Decentralized" in name. Plutocratic in practice. Same power dynamics as corporations, just with tokens and blockchain branding.
The Infrastructure Dependency
Most "decentralized apps" (dApps) don't run their own Ethereum nodes. Running full node costs $1,000+ monthly in infrastructure. Requires technical expertise. Requires maintenance.
Instead: use RPC (Remote Procedure Call) providers. Infura, Alchemy, QuickNode. These are centralized API services. You send transaction to their servers. They broadcast to Ethereum network.
MetaMask (most popular Ethereum wallet, ~30 million users): defaults to Infura. Most users never change this. Their "decentralized" transactions route through Infura's servers.
Infura outage (November 2020): parts of Ethereum ecosystem stopped working. MetaMask, many dApps, unable to interact with blockchain. The "decentralized" network depended on centralized service.
Infura is owned by Consensys. Consensys founded by Joe Lubin (Ethereum co-founder). Centralized company runs critical infrastructure for "decentralized" network.
Alchemy raised $200+ million in venture funding. Sequoia, Coatue, Andreessen Horowitz. Built infrastructure company for Web3. That infrastructure is centralized.
Can Infura or Alchemy block transactions? Yes. Can they monitor transactions? Yes. Can they identify users? Yes. Did Infura block certain addresses (Tornado Cash, after OFAC sanctions)? Yes.
The "decentralized" web runs on centralized servers. Because running infrastructure is hard and expensive. Centralized providers make it easy. Users and developers choose easy.
The Stablecoin Centralization
Crypto volatility makes it unusable as money. Bitcoin dropped from $69,000 to $16,000 (2021-2022). Can't price products in asset that fluctuates 70%.
Solution: stablecoins. Tokens pegged to dollar.
USDC (USD Coin): issued by Circle. Centralized company. Claims 1:1 backing with dollar reserves in banks. Circle can freeze USDC addresses. Has done so on blacklisted addresses.
USDT (Tether): issued by Tether Limited. Centralized company. Claims 1:1 backing. History of opacity about reserves. Fined by CFTC for misleading statements about backing. Still largest stablecoin by market cap ($100+ billion).
DAI: "decentralized" stablecoin by MakerDAO. Backed by collateral. Initially backed by ETH and other crypto. Volatile collateral requires over-collateralization. To scale, MakerDAO added USDC as collateral. Now ~50%+ of DAI backing is USDC.
"Decentralized" stablecoin backed by centralized stablecoin. Circle freezes USDC, DAI becomes partially unbacked.
Why do stablecoins recentralize? Because true algorithmic stablecoins (backed by nothing but code) fail. UST/LUNA: $60 billion market cap, collapsed to near-zero in 48 hours (May 2022). Death spiral: UST loses peg, LUNA minted to restore peg, LUNA supply inflates, LUNA price crashes, UST can't maintain peg, system collapses.
Algorithmic stablecoins are bank runs waiting to happen. Fiat-backed stablecoins are centralized by necessity. Need entity holding reserves, issuing tokens, guaranteeing redemptions.
DeFi (Decentralized Finance) built on centralized stablecoins. The money layer is centralized. Everything built on top inherits this centralization.
The Layer 2 Concentration
Ethereum transaction fees: $50-$200 per transaction during congestion. Makes network unusable for average user. Can't pay $100 to send $50.
Solution: Layer 2 networks. Process transactions off-chain, settle on Ethereum. Arbitrum, Optimism, Base, Polygon, zkSync.
Each Layer 2 run by centralized sequencer. Single entity (or small set of entities) ordering transactions. Can censor transactions. Can reorder transactions for profit (MEV: Maximal Extractable Value).
Sequencer sees pending transactions. Can front-run (place own transaction before yours). Can reorder to maximize profit. Can exclude transactions entirely.
Most Layer 2s promise "progressive decentralization." Will decentralize sequencer eventually. Eventually never comes. Centralized sequencer is more efficient, more profitable for team running it.
Users forced to Layer 2s for usability. Ethereum base layer too expensive. Layer 2s are faster, cheaper. They're also centralized.
The trade-off: affordability requires centralization. Users accept this. No alternative that's both cheap and decentralized.
"Decentralized" Ethereum. Centralized execution layers. Most users on centralized layers.
Why Decentralization Concentrates
The pattern repeats. Every decentralized technology recentralizes. The forces are predictable.
Economic forces:
Network effects. Users go where other users are. Liquidity attracts more liquidity. Coinbase gets users because it has users. OpenSea gets listings because it has buyers. The platform with most activity attracts more activity.
Economies of scale. Mining: industrial operations are 10x-100x more efficient than individual miners. Staking: large pools get better returns than individuals. Infrastructure: running at scale is cheaper per unit than running small.
Capital requirements. Running competitive mining operation requires millions. Running validators requires tens of thousands. Running infrastructure requires technical expertise and ongoing costs. This excludes most people.
User behavior:
Users choose convenience over sovereignty. Coinbase custody is easier than hardware wallet. Centralized exchange is faster than DEX. Infura API is simpler than running node.
Users choose speed over decentralization. Centralized exchanges process transactions instantly. DEXs require blockchain confirmations. Centralized platforms have better UX.
Users don't verify. Don't run nodes. Don't check blockchain. Trust platforms. This is rational. Verification costs time and expertise. Trust is easier.
Regulatory pressure:
KYC/AML requirements force centralized chokepoints. Can't have truly permissionless system with identity requirements. Identity requires entities that verify identity. Those entities are centralized.
Securities laws make decentralized issuance risky. SEC goes after projects for unregistered securities offerings. Centralized entities with lawyers can navigate this. Truly decentralized projects can't.
Governments can regulate centralized entities. Harder to regulate code or pseudonymous developers. The projects that survive regulatory pressure are ones with legal entities, compliance teams, identifiable operators. These are centralized.
Technical reality:
Running infrastructure is expensive. Most people can't afford it. Most developers don't want to deal with it. Delegation to specialists makes sense. Specialists concentrate power.
Coordination is difficult. DAOs struggle to make decisions. Centralized entities decide fast. Markets reward speed. Centralization wins.
Decentralization requires continuous effort. Centralization is default state. Water flows downhill. Power concentrates.
The Pattern Across Technologies
This isn't unique to crypto. Every "decentralization" movement follows same path.
Email (1970s-1980s):
Started decentralized. SMTP protocol. Anyone could run mail server. Early internet: universities, companies, individuals ran own servers.
Now: Gmail (1.8 billion users), Outlook (400 million+), Yahoo (200 million+). Three providers control majority of email.
Running own mail server in 2026: deliverability problems. Gmail marks your emails as spam. Other providers block your server. You need domain reputation. Reputation requires being established provider. Established providers won't let you become established.
Technical capability exists. Economic reality prevents it. Email recentralized.
The Web (1990s):
Open protocol. HTML, HTTP. Anyone could host website. Early web: individuals, universities, small companies hosted own servers.
Now: AWS, Microsoft Azure, Google Cloud host majority of websites. Three companies provide infrastructure for "decentralized" web.
Google and Facebook capture majority of traffic. Most websites depend on traffic from these platforms. The platforms control visibility. The platforms control reach.
The open web became walled gardens and centralized infrastructure. Technical decentralization. Practical centralization.
Social Media (2000s):
Started: blogs, RSS feeds, personal websites. Open standards. Decentralized publishing. Anyone could create content, host it, distribute it.
Now: Facebook, Instagram, Twitter, TikTok. Four platforms dominate social media. Blogs declined. RSS readers died. Personal websites irrelevant.
Why? Network effects. Your friends are on Facebook. Your audience is on Twitter. Your reach comes from platforms' algorithms. Can't compete with zero-network personal blog.
The blog ecosystem died. Platforms captured. Same pattern.
Crypto (2010s-2020s):
Started: Bitcoin. Decentralized peer-to-peer money. No intermediaries.
Now: Centralized exchanges, mining pools, staking pools, infrastructure providers, VC-controlled protocols.
The pattern repeats across decades and domains. Technology enables decentralization. Economics drives recentralization.
The Philosophical Problem
Decentralization assumes coordination without hierarchy is efficient.
Reality: hierarchy emerges because it's more efficient. Faster decisions. Clearer accountability. Economies of scale. Easier for users.
Markets consolidate naturally:
- Winner-take-most dynamics (network effects)
- Users flock to liquidity and activity
- Capital concentrates where returns are highest
- Successful entities grow, unsuccessful die
This is not conspiracy. This is competition. Competition creates winners. Winners dominate markets.
The dream of decentralization fights economic reality. Economic reality wins.
Users reveal preferences through actions:
- Prefer Coinbase to hardware wallet (revealed preference: convenience over sovereignty)
- Prefer centralized exchange to DEX (revealed preference: speed over decentralization)
- Prefer trusting Infura to running node (revealed preference: ease over verification)
These preferences are rational. Self-custody is hard. Running infrastructure is expensive. Coordination is difficult.
Decentralization requires sustained effort. User education. Technical sophistication. Ongoing costs. Continuous vigilance.
Centralization requires nothing. Platform makes it easy. User clicks button. Everything handled.
The path of least resistance is centralization. Most users take path of least resistance.
Decentralization advocates say users should care more about sovereignty. Users say they care more about convenience. Users' actions reveal truth.
The Tokenization Grift
Web3 didn't solve centralization. It financialized it.
Every protocol has token. Token serves functions:
- Governance (in theory)
- Value capture (in practice)
- Speculation (always)
Token distribution: 30-50% to insiders. Public buys remainder hoping for appreciation. Insiders dump on public through vesting unlocks.
ICO boom (2017): $5.6 billion raised. Thousands of projects. Promises of decentralized future. Token sales to fund development.
2026: 90%+ of ICO projects dead or abandoned. The tokens worthless. The development stopped. The promises unfulfilled.
Why? Tokens aren't products. They're financial instruments. Product development is hard. Token speculation is easy. Projects raised millions, founders got rich, development stalled.
The "decentralized" protocol became new way for VCs to exit. Traditional startup: wait for IPO or acquisition. Can take 10+ years. Crypto startup: launch token after 2 years. VCs dump on public. Exit achieved.
This was marketed as "democratizing access to early-stage investing." Reality: retail buying tokens at high prices, VCs selling tokens they got at low prices.
The technology promised decentralization. The tokenomics created extraction mechanism. Same power dynamics as traditional finance. Worse consumer protections. Better marketing.
What Remains
Fifteen years after Bitcoin whitepaper. Five years after DeFi summer. Three years after NFT mania.
What exists:
- Centralized exchanges controlling on/off ramps
- Centralized stablecoins providing money layer
- Centralized infrastructure providers enabling dApps
- Concentrated mining pools controlling Bitcoin
- Concentrated staking pools controlling Ethereum
- VC-controlled protocols with token governance
- Centralized Layer 2s providing usability
What doesn't exist:
- Decentralized finance at scale (DeFi is niche, centralized platforms dominate)
- Banking the unbanked (crypto users are wealthy tech-savvy individuals in developed countries)
- Power to the people (power concentrated in exchanges, pools, protocols)
- Individual sovereignty (users depend on platforms)
- Censorship resistance (platforms comply with regulations, block addresses)
The promise: eliminate intermediaries. The reality: new intermediaries with worse oversight.
The promise: return power to users. The reality: concentrate power in exchanges and protocols.
The promise: financial inclusion. The reality: speculative casino for wealthy.
Crypto didn't decentralize finance. It created new financial oligarchy with tokens.
Web3 won't decentralize the internet. It's recentralizing with new branding.
The Libertarian Dream Meets Venture Capital
The sequence is predictable:
Phase 1: Libertarian dream. Technology enables freedom. "Code is law." "Be your own bank." True believers build.
Phase 2: Venture capital sees opportunity. "This could be big." Funding flows. Millions, then billions.
Phase 3: User adoption. Makes sense for most users to use platforms (easier, faster, better UX). Centralization emerges.
Phase 4: Concentration. Network effects, economies of scale, winner-take-most dynamics. Handful of entities dominate.
Phase 5: Extraction. Dominant platforms have power. Use power to extract value. Fees increase. Alternatives die.
This is the cycle. Bitcoin → exchanges. Ethereum → Infura. NFTs → OpenSea. The pattern is consistent.
The revolution starts decentralized. Gets funded by capital. Capital requires returns. Returns require scale. Scale requires centralization. Centralization enables extraction.
Then it becomes what it opposed. Just with different branding.
Why This Matters
Decentralization advocates argue about technical details. Block size. Consensus mechanisms. Governance models.
The problem isn't technical. It's economic.
Can't have decentralized system that requires expensive infrastructure. The infrastructure will concentrate where capital is.
Can't have decentralized system where convenience matters. Users will choose convenient centralized platforms.
Can't have decentralized finance that interfaces with traditional finance. The interface points (exchanges) will be centralized and regulated.
Can't have decentralized governance with unequal token distribution. Governance will be controlled by largest holders.
The technical capabilities exist. The economic incentives prevent realization.
Every attempt at decentralization faces same forces:
- Users prefer convenience
- Capital requires returns
- Regulations favor compliance
- Coordination is costly
- Hierarchy is efficient
These forces don't go away. They're fundamental to markets and human behavior.
Advocates claim "next version will be different." Next consensus mechanism. Next Layer 2 solution. Next governance model.
The pattern suggests otherwise. Economic forces are stronger than technical solutions.
What Decentralization Proved
Not that decentralization is impossible. That it's uncompetitive.
Decentralized systems can work technically. Bitcoin still runs. Ethereum still processes transactions. The blockchain still maintains consensus across thousands of nodes.
But users don't use decentralized features. They use centralized platforms built on top.
The base layer is decentralized. The user experience is centralized.
This reveals the truth: decentralization is technical capability, not economic inevitability.
Markets optimize for efficiency and convenience. Decentralization is neither. Centralization wins in free market competition.
The irony: libertarian technology that empowers individuals gets beaten by market forces that concentrate power.
The free market chose centralization.
The Permanent Intermediaries
Every technological wave promises to eliminate intermediaries. Every wave creates new intermediaries.
Email: eliminate postal service → Gmail, Outlook become intermediaries
E-commerce: eliminate retail stores → Amazon becomes intermediary
Social media: eliminate traditional media → Facebook, Twitter become intermediaries
Crypto: eliminate banks → Coinbase, Binance become intermediaries
The platforms change. The pattern stays same. Because intermediaries serve function. They aggregate. They provide liquidity. They make things easy for users.
Users need these functions. Someone will provide them. Whoever provides them captures value.
Decentralization tries to eliminate intermediaries by eliminating function they serve. But users still need that function. New intermediaries emerge to provide it.
The intermediaries are permanent. Only their names change.
What This Means
Decentralization failed not through conspiracy but through economics.
Network effects favor concentration. Economies of scale favor large operators. Users favor convenience. Capital favors returns. Regulations favor compliance.
These forces exist regardless of technology. They're fundamental to markets.
Blockchain didn't change this. Tokens didn't change this. Smart contracts didn't change this.
The technology enabled new possibilities. The economics drove same outcomes.
Power concentrates. Intermediaries emerge. Users accept this. The cycle continues.
Decentralization as narrative, not reality. As aspiration, not achievement. As marketing copy, not technical description.
The revolution got funded by venture capital. Then became what it opposed.
Same power structures. Different branding. With tokens.