Blockchains are good at agreeing about information already inside their own system. Every node can verify a signature, a balance, or the result of contract code. But a smart contract cannot simply open a news website and ask for today’s gold price.
That restriction is deliberate. If thousands of nodes fetched a web page independently, they might receive different answers because of timing, location, or a changed server response. The network would no longer have one deterministic result. Smart contracts therefore need data to be delivered to them in a form the blockchain can agree on.
An oracle is the bridge that supplies this outside information. It can publish asset prices, weather measurements, sports results, proof of reserves, or other facts. A DeFi lending market, for example, may use an oracle price to decide how much a borrower’s collateral is worth.
A single data source creates a dangerous point of failure. If it is wrong, delayed, hacked, or manipulated, the contract may execute perfectly using bad information. Decentralised oracle networks try to reduce this risk by collecting reports from multiple providers and combining them according to transparent rules.
Even then, the oracle problem does not disappear. Someone must decide which sources count, how often updates occur, what happens during extreme markets, and how disputes are handled. The security of an application is only as strong as the data and assumptions on which its contracts depend.
Oracle failures can be expensive because blockchain actions are difficult to reverse. Attackers have manipulated thin-market prices, borrowed against inflated collateral, and drained protocols before prices returned to normal. Sensible designs use several sources, time-weighted prices, update limits, and emergency controls.
Oracles make real-world financial products possible on-chain. One widely used product is a token designed to follow a familiar price, usually one US dollar. Tomorrow we will ask what actually keeps a stablecoin stable.