The oracle "problem" is about to be solved very differently than most people think. Not because anyone out-competed Chainlink, but because the technical constraint that made Chainlink necessary doesn't apply to what's being built next.
Chainlink and the EVM oracle ecosystem emerged because of a specific technical constraint: smart contracts are sandboxed. They can't make HTTP calls. The only way to get external data on-chain was to push it there — through a consensus mechanism backed by a token economy.
That architecture wasn't a design choice. It was a workaround. But it became big and standardized, and now it's what most people think of when they hear the word "oracle."
The existing EVM oracle world is complex, expensive, and governance-heavy infrastructure. It once made sense given the limitations it was working around. But those limitations don't apply to the next wave of applications.
Bitcoin DLCs don't need on-chain consensus. They use Schnorr attestations — an oracle signs a value off-chain and counterparties use it to unlock the correct contract execution transaction. No token. No committee. No consensus round. The oracle makes an HTTP call, signs a number, and the cryptography handles the rest.
AI agents don't need on-chain consensus either. They already speak HTTP. They already have wallets. They need price data the same way a human developer does — with a fetch call — except they can pay for it automatically.
The oracle problem was never really about trustless data. It was about getting data into an environment that couldn't reach out and get it. Once the environment can make an HTTP call, the problem dissolves.
Before these protocols, the only viable oracle business models were subscription APIs (centralized, account-based, trust-dependent) or on-chain token systems (decentralized, but complex and expensive). There was no middle ground.
L402 and x402 create that middle ground. Pay per query, in Lightning sats or USDC on Base, with no accounts and no API keys. The payment is the authentication. You either pay or you don't. The response is cryptographically signed. You either verify it or you don't.
The Chainlink model will persist in the EVM world. Too much entrenched infrastructure and capital to unwind quickly. For applications that genuinely need on-chain data — DeFi protocols on Ethereum, Arbitrum, and similar — Chainlink and its competitors remain the right tool.
But new entrants building Bitcoin-native financial contracts and AI agent infrastructure won't reach for that toolbox. They'll reach for the simpler one based on HTTP. The reason is the same reason no one builds their own consensus mechanism when a database will do — it's the right level of complexity for the problem.
Sovereign HTTP oracles — pay-per-query, cryptographically signed, no governance token, no committee — are the natural oracle model for this next wave. An agent makes a request. It pays atomically. It receives a signed attestation it can verify independently against a published public key. No prior relationship required. No account. No trust in the transport.
That infrastructure barely exists today. Most data APIs were designed for humans — sign up, get an API key, pay monthly whether you used it or not. None of that works for machines that need to pay per call, autonomously, at any hour, for exactly the queries they make.
We're building it.
Mycelia Signal is a sovereign price oracle: 56 endpoints across crypto, FX, economic indicators, and commodities. Cryptographically signed attestations over Lightning (L402) and USDC on Base (x402). No accounts. No API keys. myceliasignal.com