Exchange Flows Classification

What is Exchange Flows Classification?

The payments pipeline deliberately excludes exchange-touching transfers from the payment mix — intra-exchange activity is not observable on-chain, so its purpose cannot be inferred. stablecoins.intelligence.central_exchange_flows_classification makes those flows visible instead of invisible, by describing the counterparty: a business sending to an exchange typically signals liquidity management; a consumer, an on/off-ramp. Every stablecoin transfer leg with a labeled exchange on either side is classified — coverage reconciles exactly with the exchange-touching population upstream. These flows are never counted as payments, B2B, or real-world payment volume.

The classification matrix

Sender type always comes first. Wallet counterparties are typed behaviorally (Consumer / Business / Institutional / Unclassified); labeled counterparties get their own segments.

Category	Direction	Counterparty
`C2X` / `B2X` / `I2X` / `U2X`	deposit	Consumer / Business / Institutional / Unclassified wallet
`X2C` / `X2B` / `X2I` / `X2U`	withdrawal	wallet, typed as above
`DF2X` / `X2DF`	either	DeFi-labeled counterparty
`PS2X` / `X2PS`	either	payment services (incl. stablecoin issuers)
`IN2X` / `X2IN`	either	infrastructure (incl. chain, custody, oracle)
`BR2X` / `X2BR`	either	bridge
`GM2X` / `X2GM`	either	gambling
`TR2X` / `X2TR`	either	treasury
`X2X`	exchange_to_exchange	both sides labeled exchanges

Two refinements add color to every exchange leg:

exchange_flow_subtype splits X2X: intra_exchange (same entity — hot→cold consolidation, treasury sweeps), inter_exchange (genuine inter-venue settlement), unknown (entity attribution missing).
exchange_rank_tier sizes the venue by global stablecoin holdings: top_1_5, top_6_15, top_16_30, rest (ranked below 30), unranked (attributed but absent from the exchange ranking — may not be a real/active exchange), unknown (no entity attribution).

Grain matters. Rows are per transfer leg — exchange batch payouts pay many users in one transaction, and every leg is kept. Filter is_max_transfer = true for Visa-style per-transaction dedup; leave it unfiltered to analyze batch withdrawals.

Schema

Table: stablecoins.intelligence.central_exchange_flows_classification

Column Name	Data Type	Description
`chain`	VARCHAR	Blockchain network (33 chains)
`unique_id`	VARCHAR	Chain-scoped surrogate key for the transfer leg
`transaction_hash`	VARCHAR	On-chain transaction identifier
`block_timestamp` / `activity_date`	TIMESTAMP / DATE	Transfer time (UTC)
`block_number`	NUMBER	Chain-native block index
`token_address` / `token_symbol`	VARCHAR	Stablecoin identity
`product_id` / `stablecoin_type` / `currency`	VARCHAR	Registry product metadata
`amount` / `usd_amount`	NUMBER	Quantity / USD notional
`from_address` / `to_address`	VARCHAR	Transfer legs
`direction`	VARCHAR	deposit / withdrawal / exchange_to_exchange
`exchange_flow_category`	VARCHAR	The matrix above
`exchange_flow_subtype`	VARCHAR	intra_exchange / inter_exchange / unknown (X2X only)
`exchange_address`	VARCHAR	The exchange-labeled side
`exchange_rank_tier`	VARCHAR	Venue size tier by stablecoin holdings
`counterparty_address`	VARCHAR	The non-exchange side
`counterparty_segment`	VARCHAR	Unified taxonomy: consumer, business, institutional, unclassified, defi, payment_services, infrastructure, bridge, gambling, treasury, exchange
`counterparty_wallet_type`	VARCHAR	Behavioral type for wallet counterparties; null for labeled segments and X2X
`counterparty_balance_tier` / `counterparty_30d_volume`	VARCHAR / NUMBER	Counterparty wallet context
`is_max_transfer`	BOOLEAN	Largest leg of its (day, transaction) — per-transaction dedup gate
`is_mev_transfer` / `is_inorganic_sender` / `is_inorganic_receiver`	BOOLEAN	Kept and flagged — a bot depositing to an exchange is a real exchange flow

Example queries

Deposit vs withdrawal mix by counterparty segment (30d)

SELECT
  chain,
  exchange_flow_category,
  COUNT(*)        AS legs,
  SUM(usd_amount) AS usd
FROM stablecoins.intelligence.central_exchange_flows_classification
WHERE activity_date >= DATEADD('day', -30, CURRENT_DATE)
  AND is_max_transfer
GROUP BY 1, 2
ORDER BY usd DESC;

Hot→cold consolidation vs inter-venue settlement, by venue tier

SELECT
  exchange_rank_tier,
  exchange_flow_subtype,
  SUM(usd_amount) AS usd
FROM stablecoins.intelligence.central_exchange_flows_classification
WHERE exchange_flow_category = 'X2X'
  AND is_max_transfer
GROUP BY 1, 2
ORDER BY usd DESC;

Exchange-rail payouts (batch withdrawals to many recipients)

SELECT chain, activity_date, transaction_hash,
       COUNT(*) AS legs, SUM(usd_amount) AS usd
FROM stablecoins.intelligence.central_exchange_flows_classification
WHERE direction = 'withdrawal'
GROUP BY 1, 2, 3
HAVING COUNT(DISTINCT counterparty_address) >= 10
ORDER BY usd DESC;

Relationship to the behavioral flows pipeline. Payout runs executed from exchange accounts appear here as high-fan-out withdrawals — they are deliberately out of scope for merchant_payout_runs, which covers platform-wallet payouts. The two are adjacent and non-overlapping; union them knowingly for total on-chain disbursement.

Overview

Data Verticals

Solana

Hyperliquid

Lighter

Bitcoin Ecosystem

EVM

Cosmos Ecosystem

Move Ecosystem

Other Ecosystems

What is Exchange Flows Classification?

The classification matrix

Schema

Example queries

Deposit vs withdrawal mix by counterparty segment (30d)

Hot→cold consolidation vs inter-venue settlement, by venue tier

Exchange-rail payouts (batch withdrawals to many recipients)

​What is Exchange Flows Classification?

​The classification matrix

​Schema

​Example queries

​Deposit vs withdrawal mix by counterparty segment (30d)

​Hot→cold consolidation vs inter-venue settlement, by venue tier

​Exchange-rail payouts (batch withdrawals to many recipients)

What is Exchange Flows Classification?

The classification matrix

Schema

Example queries

Deposit vs withdrawal mix by counterparty segment (30d)

Hot→cold consolidation vs inter-venue settlement, by venue tier

Exchange-rail payouts (batch withdrawals to many recipients)