DeFi Without Jargon: Lending, Exchanges and the Risks in Between

Imagine a currency exchange office without staff, without opening hours and without a bank behind it. You hand over one currency, a program calculates the rate and returns another, and anyone can read the rules beforehand. That, in simplified form, is the idea behind decentralized finance, usually shortened to DeFi.

DeFi bundles financial services that run as software on public blockchains, most of them on Ethereum and the networks built around it. There is no account opening, no credit check and no customer service. What replaces them is code that executes automatically and is visible to everyone, which is both the strength and the weakness of the model.

This article explains the two building blocks that carry most of DeFi, decentralized exchanges and lending markets. It shows where the advertised yields come from and walks through the risks that tend to get lost between the promises. It does not recommend any product or strategy.

What decentralized finance actually is

The core of DeFi is the smart contract: a program stored on a blockchain that holds funds and moves them according to fixed rules. Once deployed, it runs as long as the network runs. Users interact with it through a wallet, a piece of software that manages their keys and signs transactions.

Because the contracts are public, they can be combined. A token received from one protocol can be deposited in a second, which in turn uses a third as its price source. Developers call this composability, and it explains why DeFi grew so quickly once it gained traction around 2020. It also explains why a failure in one place can spread to others.

Decentralized does not automatically mean ownerless. Many protocols are run by small teams in their early years and keep administrator keys that can change important parameters. How decentralized a service really is has to be checked case by case.

Exchanges without an order book

On a traditional exchange, buyers and sellers place orders and a matching engine pairs them. Most decentralized exchanges work differently. They use an automated market maker, or AMM: a smart contract that holds two tokens in a liquidity pool and sets the price based on the ratio between them.

The best known version follows a simple rule: the product of the two token quantities in the pool stays constant. If someone buys token A with token B, the pool ends up holding less A and more B, so A becomes more expensive. Large trades in small pools move the price noticeably, an effect known as slippage.

The tokens in the pool come from liquidity providers, private individuals or professional firms who deposit both assets. In return, they receive a share of the trading fees, so their income depends on how much trading actually takes place.

Lending against collateral

DeFi lending markets bring together depositors who want interest and borrowers who want liquidity. Since there is no credit check and borrowers are often anonymous, the system relies entirely on collateral. A borrower who wants the equivalent of 100 dollars in stablecoins might, for example, have to deposit crypto assets worth 150 dollars or more.

This overcollateralization sounds inefficient, and from a borrower's point of view it often is. People use it anyway, for instance to obtain liquidity without selling an asset they want to keep, or to build leveraged positions. Interest rates usually adjust automatically: the larger the share of a pool that is lent out, the higher the rate.

If the value of the collateral falls below a defined threshold, the protocol triggers a liquidation. Outside parties may repay part of the loan and receive the collateral at a discount. This keeps the lending pool solvent, but in sharp market declines many positions can be liquidated at once, which adds further pressure on prices.

Where the yield comes from

Many DeFi interfaces display annual percentage yields prominently. Behind every such number there should be a source that can be named. In practice, there are three common ones, sometimes combined.

• Trading fees: liquidity providers receive part of what traders pay for each swap. The return rises and falls with trading activity.
• Borrowing interest: depositors in lending markets receive the interest that borrowers pay. It depends on demand for loans and can change from one block to the next.
• Token rewards: protocols distribute their own tokens to attract deposits. Their value depends on the market and often declines once many recipients sell.

A fourth source is less visible: risk. Some returns are high because depositors carry risks they cannot see, such as untested code, thin collateral or a business model that pays existing users with fresh deposits. The collapse of the Terra ecosystem in May 2022, when the algorithmic stablecoin UST lost its peg to the dollar, showed how quickly seemingly safe high interest can vanish.

A yield without a source you can name is not a reward. It is usually the price of a risk that has not become visible yet.

The risks between the promises

DeFi removes some familiar risks, such as a bank that freezes an account, but it adds others. Most of them are technical or economic and hard to judge from the outside.

Smart contract risk

Code can contain errors, and on a blockchain errors are expensive because transactions are generally final. As early as 2016, an attack on The DAO, an investment vehicle built on Ethereum, drained a substantial share of its funds and led to a split of the network. Security audits reduce this risk, but they do not eliminate it.

Oracle risk

Smart contracts cannot see what happens outside the blockchain. To know the price of an asset, they depend on oracles, services that feed external data into the chain. If an oracle reports a wrong or manipulated price, a lending protocol can liquidate healthy positions or lend against collateral that is worth far less than assumed.

Impermanent loss

If the prices of the two tokens in a pool drift apart, the pricing formula shifts a liquidity provider's holdings toward the token that has lost value. Compared with simply holding both tokens, this can mean a loss. It is called impermanent because it shrinks if prices return to where they were, but it becomes permanent once the provider withdraws.

On top of this come bridges, contracts that move tokens between blockchains and have been the target of several of the largest thefts in crypto, particularly in 2022. Fake websites that imitate familiar interfaces are another common trap.

Governance tokens and who is really in charge

Many protocols issue governance tokens. Holders can vote on proposals, for example on fees, accepted collateral or the use of a treasury. In theory, this hands control to the community. In practice, a large share of the tokens is often held by founders, early investors and a few large holders.

Turnout in such votes tends to be low, and anyone who controls enough tokens can push decisions through. There have been cases in which attackers borrowed voting power at short notice to pass harmful proposals. A governance token is therefore no guarantee of a democratic structure, and its market value says little about the quality of the protocol behind it.

Regulation and a sober outlook

In the European Union, the Markets in Crypto-Assets Regulation, known as MiCA, has applied in full since December 30, 2024. It covers crypto-asset service providers and stablecoin issuers but largely leaves out services that are provided in a fully decentralized way without an intermediary. Where exactly that line runs is still a matter of debate.

For users, this means that the usual protections of regulated finance often do not apply. There is no deposit insurance, no ombudsman and usually no one to call when something goes wrong. Anyone trying to judge a service can start with a few sober questions: how long has it existed, who has reviewed its code, who can change its rules and where exactly the return comes from.

DeFi has shown that basic financial functions can run as open software. Whether it becomes a lasting part of the financial system depends less on new yield promises than on unglamorous qualities: secure code, reliable price data, clear responsibility and rules that protect users without stifling the technology.