Chainlink solved a real problem. EVM smart contracts execute in a deterministic sandbox. They can't reach outside their execution environment to fetch data. If a DeFi protocol needs a price feed, it needs something to push that data onto the chain. Enter oracle networks.
The architecture made sense for the constraints of the time. A decentralised network of nodes aggregates data, reaches consensus, and writes the result to the chain. The smart contract reads from the chain. The loop closes.
But those constraints were always specific to EVM. And the world is expanding beyond EVM.
A Discreet Log Contract is a Bitcoin-native smart contract. Two counterparties lock funds in a multi-sig output and agree in advance: if the oracle attests outcome X, counterparty A can claim the funds; if it attests Y, counterparty B can. The outcome is determined by a cryptographic signature from a pre-committed oracle — not by an on-chain price feed.
The oracle doesn't write anything on-chain. It just signs a message. The signature is the attestation. Anyone can verify it against the oracle's published public key. The contract resolves without the oracle touching the chain at all.
This changes the oracle requirement entirely. A Bitcoin DLC oracle doesn't need to be a blockchain node. It doesn't need to participate in a consensus network. It doesn't need a token. It just needs to sign messages — Schnorr signatures over secp256k1, committed in advance via R-point announcements, revealed at attestation time.
An HTTP server with a private key is sufficient. And a sovereign HTTP oracle with no token, no governance mechanism, and no shared infrastructure is actually better — because it has an independent failure mode.
An AI agent making a financial decision needs signed data it can verify. It doesn't need that data to come from a blockchain. It needs a cryptographic proof that the data is what the provider claims it is.
Ed25519 and secp256k1 signatures over a canonical string achieve this. The agent fetches a price, receives a signature, checks the signature against a published public key, and knows the data hasn't been tampered with. No chain required. No consensus mechanism required. No token required.
What the agent does need is a payment mechanism that doesn't require a human account holder. That's where L402 (Lightning) and x402 (USDC on Base) come in — per-query micropayments the agent can execute autonomously, with no subscription, no API key, and no human in the loop.
A sovereign HTTP oracle is exactly what it sounds like: an HTTP server that signs its data and publishes its public key. No token. No DAO. No governance mechanism. No shared infrastructure with any other oracle.
The sovereignty matters for the same reason independence matters in any verification system. If your primary oracle and your backup oracle share infrastructure, governance, or exchange relationships, they share attack surface. A compromise that affects one may affect both. An oracle with independent failure modes closes attack vectors that shared-infrastructure solutions leave open.
The architecture we've built at Mycelia Signal:
Two independent GCP nodes — US (Iowa) and Asia (Taiwan) — running identical stacks with per-instance keypairs. A compromise of one node doesn't compromise the other's signing key.
Ed25519 signing on every response — the canonical string is SHA-256 hashed and signed. Anyone can verify the signature offline against the published public key. No trusted intermediary required.
DLC support via Schnorr — R-point announcements published in advance, attestations published at event resolution. Bitcoin-native oracle support without touching the Bitcoin chain.
Pay-per-query via L402 and x402 — Lightning invoices or USDC on Base. No account, no subscription, no API key. An agent can pay autonomously.
To be clear: EVM smart contracts still need on-chain oracles. Chainlink still serves a real market. This isn't a prediction that oracle networks will disappear.
It's an observation that the next wave of programmable money — Bitcoin DLCs, AI agent commerce, agent-to-agent financial settlement — has different requirements. Requirements that a sovereign HTTP oracle serves better than a blockchain-native oracle network.
The constraint that created the oracle problem was always EVM-specific. The solution that emerges for non-EVM environments will look different. It already does.
63 signed endpoints. DLC support. Pay per query via Lightning or USDC. No account required.
Mycelia Signal is a sovereign cryptographic oracle — 63 signed endpoints across crypto, stablecoin pegs, volatility indices, FX, economic indicators, and commodities. Payable by AI agents via Lightning (L402) or USDC on Base (x402). myceliasignal.com