Sources & methodology

Holdings include third-party transfers. They are separate from deposits along the Payment Abstraction path and do not measure revenue or distributions to token holders. The history chart uses daily LINK balances; the source’s historical USD balances are repriced at its latest quote and are not historical valuations. Timezone-less source timestamps are interpreted as UTC.

These are gross deposits along one path, not all reserve receipts, net balance growth or all Chainlink income. Weekly bars use Monday–Sunday UTC weeks. Cumulative inflows reset at the selected period’s start; a rising cumulative chart alone does not show acceleration.

The dated funding study uses thirteen complete weeks. Its latest four-week totals are compared with the preceding four complete weeks. LINK amounts retain the raw 18-decimal token integers in the download. USD values come from DefiLlama’s daily series; they are not execution prices or current marked-to-market values. A missing day leaves the USD week unavailable.

The explorer page must extend before the study’s start or exhaust its history. This establishes a covered interval in the provider’s index, not independent chain completeness. Other senders and withdrawals are recorded separately. Transaction examples retain exact hashes, log indices and block hashes.

A conversion transaction can identify the input token and payment contract. It does not automatically identify the original customer, service or offchain payment. Amounts entering a shared balance are not assigned to later withdrawals solely because their values match.

Reserve design ↗ · Payment Abstraction architecture ↗

Volume is not fees, revenue or unique capital; the same capital may be transferred repeatedly. The first month is not inferred from zero. Missing months and negative corrections leave a monthly value unavailable. Coverage and filters are the provider’s, not an independent replay of the network.

Complete daily CCIP fees and their network-fee / execution-cost split are not covered. The separate GHO case study follows outbound Ethereum messages only. Returning origin addresses are not identified returning customers. Selected message fees cannot establish network totals.

The SVR charts use LlamaRisk’s Aave market records. They show gross shared proceeds before allocation, grouped by collateral and recorded UTC day. Missing dates remain unknown; the collection date’s partial day is excluded.

Daily histories end at different dates and do not fully reconcile with the separate financial summary. The API supplies no completeness watermark, and its Chainlink allocation does not separately identify builder costs. The page lists the gaps and exact coverage by market.

Reported allocations are not profit, verified LINK purchases or reserve receipts, and are not added to reserve funding or request payments. Collateral exposure to a proposed liquidation-fee change is historical exposure, not a forecast of lost income or proof the proposal was implemented.

Neither series measures all Data Feeds, CCIP or private enterprise activity. Batching means a receipt date can differ from when a service was used.

Aggregator receipts and reserve deposits can be successive stages of the same funds. They must not be added as separate revenue. Their ratio is not a conversion rate: timing, price, routing and other transfers can differ.

Request payment calculation ↗ · Aggregator calculation ↗

The supply feed has no observation timestamp or historical release series. Site history starts with its first collected daily reading, preserving gaps. A supply change does not identify an exact release time or a sale. Staking and reserve balances are not subtracted to estimate tradable supply.

Price is not a streaming quote. It is marked overdue after one hour from its source observation. The reserve’s USD valuation uses the official reserve feed’s own price and timestamp, which can differ from the headline quote.

Overview payment charts use complete Monday–Sunday weeks. The 4W, 13W and 52W selectors end on the latest complete Sunday. “All” begins with the later first observation shared by request payments and reserve funding. A missing day makes the week and its period total unavailable. The trailing average requires four complete consecutive weeks.

The latest reserve balance, latest complete CCIP calendar month, cumulative reported SVR allocation and dated case studies retain their own periods. They do not change with the overview’s payment-chart selector. A provider-reported zero is distinct from a missing observation.

Dashboard sources are checked on visits after six hours. Manual refresh is shared across visitors and available every ten minutes. Official reserve checks remain limited to once per six hours and monthly CCIP volume checks to once per day. Failed updates keep the last available data and display a warning.

Collection or source observations older than 48 hours are marked overdue, except monthly CCIP volume allows 45 days from month-end and price allows one hour from its observation. Case studies retain their own reviewed snapshots. A fresh source request does not update an old analysis date.