OCB Agentic Share of Onchain Capital
How much tokenized-dollar activity onchain looks machine-driven rather than human. A coverage-aware measure of the agentic share of stablecoin transfer activity — market-wide and per instrument. A leading indicator, not a hard classifier.
Market-wide, 44.4% of stablecoin transfer activity carries a machine-driven fingerprint (-0.5 points over the trailing 30 days). The reading is measured on 98.1% of volume — the share with machine-counterparty detection — across 35 covered instruments. Agent-driven activity means the counterparty is a smart contract, not a human wallet; the amounts are sub-dollar or non-round; and the timing runs off human hours.
Which stablecoins look most agent-driven?
Every stablecoin OCB collects transfer history for, ranked by its latest agentic-share score (0–100), most machine-like first. This score is specific to this page — a composite of the five behavioral sub-signals below — and is not the OCB quality rating. A dash in the contract column means OCB could not classify the counterparty on that instrument’s chains: not a zero, not missing activity. Click a column to sort, or filter by ticker.
| Stablecoin | Agentic score▼ | Contract | Sub-dollar | Non-round | Off-hours |
|---|---|---|---|---|---|
| FDUSD | 55.60% | — | 48.0% | 98.0% | 100.0% |
| U | 51.45% | 97.6% | 3.4% | 97.9% | 76.6% |
| USDS | 50.40% | — | 32.7% | 97.5% | 100.0% |
| USDe | 50.02% | — | 39.2% | 84.8% | 100.0% |
| USDC | 49.63% | — | 39.3% | 82.8% | 100.0% |
| USDT | 48.18% | — | 33.4% | 85.3% | 100.0% |
| USD1 | 45.28% | — | 17.3% | 97.6% | 100.0% |
| USD0 | 45.10% | 98.1% | 6.2% | 98.1% | 28.7% |
| EURC | 45.00% | 96.3% | 5.7% | 88.2% | 41.7% |
| FRAX | 44.29% | 99.7% | 2.5% | 96.9% | 28.1% |
| DAI | 42.88% | 96.0% | 12.8% | 93.9% | 10.5% |
| GHO | 42.32% | 96.1% | 4.5% | 91.1% | 23.4% |
| PYUSD | 41.84% | — | 15.3% | 83.7% | 100.0% |
| EURW | 40.00% | 100.0% | 0.0% | 81.0% | 19.0% |
| USDG | 38.71% | — | 5.7% | 84.1% | 100.0% |
| MXNB_ARBITRUM | 38.40% | 56.0% | 28.0% | 16.0% | 100.0% |
| XSGD | 38.13% | 93.8% | 6.3% | 75.0% | 12.5% |
| EURe | 37.25% | 98.3% | 6.6% | 66.1% | 7.6% |
| RLUSD | 37.04% | 77.0% | 11.1% | 59.7% | 40.4% |
| BRLA_POLYGON | 36.77% | 67.0% | 7.6% | 52.2% | 68.6% |
| USDP | 36.70% | 73.8% | 12.0% | 86.4% | 15.2% |
| IDRX_POLYGON | 36.67% | 46.7% | 20.0% | 100.0% | 33.3% |
| USDD | 33.91% | 96.8% | 3.4% | 28.2% | 30.8% |
| EURCV | 31.90% | 82.8% | 1.9% | 49.2% | 22.4% |
| QCAD | 30.00% | 0.0% | 0.0% | 100.0% | 100.0% |
Score 0–100, weighted 25 / 25 / 15 / 15 / 20 across the five sub-signals; any unmeasurable sub-signal is dropped and the rest re-normalized. Latest day per instrument.
A leading indicator, not a verdict.
OCB does not label any single transfer as agentic. The score is a weighted blend of behavioral signals that, in aggregate, track how machine-like an instrument’s activity is. Two of the five sub-signals are deliberately conservative today, so the number reads as a lower bound.
| Sub-signal | Weight | What it reads |
|---|---|---|
| Contract counterparty | 25.0% | share of transfers with a smart contract, not a human wallet, on the other side |
| Sub-dollar amounts | 25.0% | share of transfers below $1 — atypical for human payments |
| Non-round values | 15.0% | share of amounts a human would not hand-type |
| Off-hours timing | 15.0% | share of activity outside human working hours (a coarse UTC proxy today) |
| Sub-second cadence | 20.0% | not computed at v1.0 — contributes zero, so the score is understated, not overstated |
This is a measure of tokenized-dollar activity, not all onchain capital. The market number is a volume-weighted blend across the covered stablecoin universe, so a large-volume instrument moves it more than a small one. The denominator is the tokenized dollars OCB collects transfers for — a stated scope, not an implied one. Contract share is measured only over transfers OCB could classify; on a chain without machine-counterparty detection it reads as a dash, never a false zero. The sub-second sub-signal is not yet computed, so the agentic share is a lower bound. The scope widens as coverage grows.
As of Sep 8, 2026