There is a piece of software on the internet that has processed more than two trillion dollars in trades. It has no CEO, no customer service line, no headquarters, and no employees in the sense your bank would recognize. It cannot be switched off, because it does not run on any single company’s servers. It runs on thousands of computers scattered across the planet, owned by strangers who have never met. Its name is Uniswap, and it is a decentralized application, a “dapp.” Understanding why something like this exists, and why regulators and banks lose sleep over it, requires starting from zero. So let us start from zero.
The Ledger Problem
Strip away the marble lobbies and the mobile apps, and a bank is a ledger. A list. When your salary arrives, nobody moves physical cash. A clerk, or these days a database, edits a row: your number goes up, your employer’s number goes down. The entire financial system is a stack of these private ledgers, each maintained by an institution you are required to trust.
That trust carries a price. The institution can freeze your row. It can deny you a row in the first place, which is the daily reality for over a billion unbanked adults. It can fail, as Lehman Brothers did in 2008 and Silicon Valley Bank did in 2023. And when two institutions in different countries need to reconcile their ledgers, the process crawls through correspondent banks, each taking a fee and a day.
Decentralization asks a simple, almost impertinent question: what if the ledger belonged to no one, and therefore to everyone?
How a Blockchain Actually Works
A blockchain is that shared ledger. Instead of one bank holding one private copy, thousands of computers, called nodes, each hold an identical copy of the full transaction history. When you send money on this system, you broadcast the transaction to the network, and every node writes it into its copy.
This raises an obvious problem. If everyone keeps the books, who decides which transactions are valid? What stops me from writing myself a fortune? The answer comes in two parts, and both are worth understanding because they are the entire trick.
The first part is the “chain” in blockchain. Transactions get bundled into batches called blocks, roughly every few seconds or minutes depending on the network. Each block is run through a cryptographic hash function, a mathematical blender that converts the block’s contents into a unique fingerprint. Change a single character in the block, even one digit of one transaction, and the fingerprint changes completely. Here is the crucial design choice: every new block contains the fingerprint of the block before it. The blocks are chained together by their fingerprints. If a fraudster edits an old transaction, that block’s fingerprint changes, which breaks its link to the next block, which breaks the next, all the way to the present. To rewrite history, you would have to rebuild the whole chain faster than the honest network extends it. On a large network, that is economically absurd.
The second part is consensus, the mechanism by which thousands of strangers agree on which block comes next. Bitcoin uses proof of work: computers race to solve a brute-force puzzle, and the winner earns the right to add the block plus a reward in newly minted coins. Ethereum, the network where most dapps live, switched in 2022 to proof of stake: participants called validators lock up capital, currently a minimum of 32 ETH, as a security deposit. Validators take turns proposing blocks and voting on them. Behave honestly and you earn a yield. Cheat, and the protocol destroys part of your deposit, a punishment known as slashing. Honesty is not assumed. It is priced in.
The result is a ledger with three properties no bank database has ever offered. It is public, so anyone can audit every transaction ever made. It is permissionless, so anyone with an internet connection can use it without an application form. And it is immutable, so no manager, minister, or hacker can quietly edit the past.
Layers One and Two
The base networks that provide this security, Bitcoin, Ethereum, Solana, are called layer 1 blockchains. They are the ground floor, the final court of settlement. Their weakness is throughput. Ethereum’s base layer handles on the order of fifteen to thirty transactions per second, because every node on Earth must process every transaction. Visa’s network claims capacity in the tens of thousands. During busy periods, Ethereum users have paid fees of fifty dollars or more for a single transaction.
Layer 2 networks are the fix, and the analogy is the bar tab. You do not settle with the bartender after every drink. You run a tab and settle once at the end of the night. Layer 2s such as Arbitrum, Optimism, and Base process thousands of transactions cheaply on their own faster rails, then periodically compress the results, along with a cryptographic proof that the math was done honestly, and post that summary down to the layer 1. The base layer’s security anchors everything; the second layer provides the speed. A transaction that costs dollars on Ethereum itself costs a fraction of a cent on Base. This layered architecture is why decentralized finance has become plausible for everyday payment volumes rather than just large transfers.
From Ledger to Application
A ledger that only records payments would be interesting. What made Ethereum an industry is that its ledger also stores and executes code. These programs are called smart contracts, which is a misleading name; they are neither smart nor contracts in the legal sense. A smart contract is simply software published onto the blockchain, where it becomes as public and as tamper-proof as the transactions themselves. Once deployed, it runs exactly as written, for anyone, forever, and in most cases not even its author can alter it.
A decentralized application is a bundle of these smart contracts plus a website that lets humans talk to them. Compare it to the app on your phone. When you use your banking app, your request travels to the bank’s servers, where the bank’s staff and the bank’s rules decide what happens. When you use a dapp, your request goes straight to code on a public network. There is no company in the middle approving the transaction, holding your money, or maintaining an account for you. Your account is a cryptographic keypair in your own wallet. The dapp never has custody of anything.
This is the entire philosophical payload of decentralization, compressed: replace institutions you must trust with code you can verify.
DeFi, in the Wild
Decentralized finance, DeFi, is what happens when people rebuild the products of a bank out of these contracts. The flagship examples have been running for years, in public, through several market collapses.
Uniswap is an exchange with no order book and no market makers on salary. Liquidity providers deposit pairs of tokens into shared pools; a formula sets the price automatically based on the ratio in the pool; traders swap against the pool and pay a small fee that flows back to the depositors. Aave and Compound are lending markets. Depositors supply assets and earn interest; borrowers post collateral worth more than their loan and draw against it; if the collateral’s value falls below a threshold, the contract liquidates it automatically, no collections department required. MakerDAO, now rebranded as Sky, issues DAI, a dollar-pegged stablecoin generated against locked collateral, effectively a decentralized central bank governed by tokenholder vote. Lido lets small holders pool funds to run Ethereum validators and share the staking yield. None of these platforms closes for the weekend, and every rate, reserve, and rule is inspectable on-chain by anyone with a browser.
What the Incumbents Fear
Regulators and banks look at this and see three genuine threats tangled together.
The first is the compliance gap. The modern anti-money-laundering regime, from the Bank Secrecy Act to the FATF’s Travel Rule, is built on regulated intermediaries who identify their customers and report suspicious activity. A dapp has no intermediary to deputize. The US Treasury made the sharpest version of this argument in 2022 when it sanctioned Tornado Cash, a privacy dapp, effectively blacklisting a piece of autonomous code. Courts pushed back, ruling in 2024 that immutable smart contracts were not sanctionable property, and the sanctions were lifted in 2025, but the underlying question, who answers for software that no one controls, remains unresolved in most of the world.
The second is jurisdictional. Securities and banking law assumes an issuer or an operator you can subpoena. The SEC spent years arguing that many DeFi tokens are unregistered securities and pursued Uniswap’s developer before dropping the effort in 2025 as US policy shifted with the GENIUS Act’s stablecoin framework and a friendlier administration. Europe’s MiCA regulation, meanwhile, largely exempts fully decentralized protocols while capturing anything with an identifiable operator, which has turned “how decentralized is it, really” into a legal question worth billions.
The third threat is the oldest one: disintermediation. Deposits, payments, FX spreads, and custody fees are the income of banking. A system where users custody their own assets and settle peer to peer competes with all of it. The banks’ response has been twofold: lobby for restrictive rules, and quietly adopt the technology, as JPMorgan has with its Kinexys tokenized settlement platform. Debanking of crypto-adjacent businesses, the practice at the center of the US “Chokepoint 2.0” controversy, showed how effectively traditional finance can squeeze the on-ramps and off-ramps even when it cannot touch the protocols themselves.
Every one of these fights concedes the same point: the software works well enough to be worth fighting over.
Five Use Cases That Hold Up
Set aside speculation, and the strongest arguments for dapps are practical.
1. Cross-Border Payments and Remittances
A migrant worker sending money home through legacy remittance channels pays global average fees above six percent and waits days. A stablecoin transfer over a layer 2 settles in seconds for pennies, at any hour. Stablecoins now move trillions of dollars a year in settlement volume, and the busiest adoption is in corridors the banks serve worst.
2. Banking the Excluded
In Argentina, with inflation that has run at triple digits, and in Nigeria and Turkey with chronically weak currencies, ordinary people hold dollar stablecoins in self-custody wallets because a hard-currency bank account is unavailable or untrustworthy. The dapp does not ask for a minimum balance, a credit history, or a passport.
3. Transparent, Automated Credit
On Aave, the collateral rules, interest model, and reserves are public code. There is no discretionary loan officer and no hidden balance sheet. During the 2022 collapses, it was the centralized lenders, Celsius and BlockFi, that froze withdrawals and filed for bankruptcy; the major DeFi lending contracts kept liquidating and repaying exactly as written.
4. Tokenized Real-World Assets
US Treasury yields are now sold on-chain through funds like BlackRock’s BUIDL, giving a business in Lagos or Karachi access to T-bill returns with a wallet instead of a Wall Street relationship. Tokenized Treasuries grew from roughly one hundred million dollars in early 2023 to several billion by 2025.
5. Censorship-Resistant Fundraising and Settlement
When Canadian authorities froze protest-linked bank accounts in 2022, and when NGOs needed to move funds into Ukraine within days of the invasion, crypto rails carried value that the banking system either blocked or could not deliver in time. Whatever one thinks of any specific cause, infrastructure that cannot be selectively switched off is a different kind of public good, and occasionally a lifeline.
The Uncomfortable Conclusion
Dapps are not a finished product. Users lose funds to phishing, bad code, and their own mistakes, and there is no fraud department to call. But the direction of travel is hard to miss. The world’s largest asset managers are tokenizing funds, banks are building on the same rails they lobbied against, and a generation of users in weak-currency countries has already voted with its savings. The bank with no building is still standing. The question the incumbents must now answer is not whether decentralized finance is legitimate. It is why, after every crackdown and collapse, the ledger that belongs to no one keeps growing.