Source: data/out/daily.csv, column l1_staked: the staked balance (pool get_accounts) of the 73 Rule A accounts at the first block after 00:00 UTC each day. Dots: data/issuer_snapshots.csv, compared with the next day's 00:00 row. Built by nfi/build.py.
What the flow tells us
| What has to be true | Reading now |
|---|
Product and holder evidence status
Verdict: large position established; persistent incremental buying not proven. The full thesis, with the evidence against it, two pre-registered predictions and frozen signposts that can prove it wrong, is in THESIS.md. Research, not financial advice.
Is NRR adding or releasing NEAR faster than supply changes?
Cumulative signed NRR custody flow is shown against cumulative NEAR minted and the buyback-fund balance change. Each line is rebased to zero at the selected range start. These are balance movements, not market purchases or sales.
The long-term test, 2026 to 2029
One number decides it: the absorption ratio, wrapper net custody flow divided by NEAR minted in the same 30-day window. +1 means the wrapper took in all new NEAR; −1 means it released as much again. Rules were frozen on 14 Nov 2026, before the first window (SIGNPOSTS.md, SHA-256 and timestamp proof).
| Window | Activity covered (UTC) | Status | Absorption | Buyback ratio |
|---|
Issuance is scheduled to fall, not falling
What each scenario would mean for the wrapper's share of supply
Test calendar, indicators, review routine and build order: LONG_TERM_PLAN.md.
Follow the NEAR
Pick a creation or a redemption and follow it account by account. Every step links to the transaction on NearBlocks. Each step says how sure we are: "observed" is a transaction by that account; past the custody hub, NEAR is pooled, so a link is at best "probable" (the only matching outflow within 72 hours) and never certain.
Day by day
One row per day at 00:00 UTC. Select a row to see the transactions behind it and whether each one completed its sequence. Rewards are measured from delegators who made no pool transaction that day; whatever is left is shown as unexplained, never hidden. A value we could not read is shown as "missing", never as zero.
| Date | Into custody | Out of custody | Staking rewards | Unexplained | Staked by NRR | Bitwise, prior day | Gap | Data |
|---|
Fund details
From Bitwise's NRR page, dated. Shares outstanding are implied: NEAR in Trust divided by NEAR per share.
What this proves, and what it doesn't
Shown directly by the chain
- The NEAR staked by Rule A accounts attributed to NRR, every day.
- Every creation and redemption transaction, with amounts.
- Each step a redeemed amount took, as far as the custody hub.
- NEAR's total supply, the NEAR minted each epoch, and total stake.
- The buyback fund's balance, including its wNEAR inside NEAR Intents.
Not shown by the chain
- Whether NEAR moving into custody was bought on the market, or delivered from existing holdings.
- Where NEAR went after the custody hub: the hub pools balances, so those links are at best probable.
- Whether redeemed NEAR was sold. Trades inside an exchange are not on-chain.
- NEAR's total float. Ordinary staking outside the wrapper moves several times more NEAR a day.
- Legal ownership of the accounts. Our attribution is rule-based and checked against Bitwise's figures.
- Executed buyback purchases or burns. We show the buyback fund's balance only.
- Other products are not fully reconstructed; SVRN is shown only as a partial one-pool observation.
Check it yourself
Everything on this page is rebuilt from cached public-chain data. The full raw-to-output rebuild is offline. No API key, wallet or paid service.
git clone https://github.com/Gosho1992/near-flow-intelligence cd near-flow-intelligence make rebuild # raw cached responses -> work datasets -> outputs and site make test # offline invariant and reproducibility checks make serve # this page on http://127.0.0.1:8000
Method, attribution rules and formulas: METHOD.md. Investment thesis: THESIS.md. Long-term plan: LONG_TERM_PLAN.md.