Where the Stack Sits: A Supply Map of XRP Going Into the Fed

After a +28% August, a record ≈$159.2M ETF month, and a $2B stablecoin milestone, the question that actually prices September is blunt: who holds the stack, where did they buy it, and how much of it can move? This report maps the answer with sourced, dated figures — some series carry older snapshot dates than others, and we print the dates rather than refresh unverifiable numbers. Written into decision week; the September 16 Fed decision still lies ahead as of the September 14 revision.

Part 1 — Two supplies, one token

Every supply map starts with an embarrassment: the two most-published circulation figures disagree by roughly 5 billion XRP.

The gap is methodology, not error: aggregators apply Ripple's distribution-API methodology to exclude certain holdings (escrow-adjacent, treasury-adjacent) from "circulating," while the ledger's raw activated count is what the chain itself says. Market caps — including the ~$86B August print and the $85.8B September baseline — are computed against the estimate. My take: treat the ledger number as the physical inventory and the market-data number as the investable float estimate; the 4.8B XRP between them is where supply debates go to hide.

Part 2 — The cost-basis map: the $1.31–$1.38 zone

On-chain analysis flagged a whale demand zone at $1.31–$1.38 where roughly 4.8 billion XRP was previously accumulated — about 7–8% of circulating supply on either basis (Ali Martinez, September coverage). In the week's tape the zone did its job: the early-September push toward $1.65 faded through $1.44, $1.38 and $1.37, and buyers defended the $1.38 line with the zone underneath (Yahoo Finance / ICO Announcement chart data; the Sep 7–13 editions cover the sessions).

What a cost-basis map can and cannot tell you. It can: where large holders' average entries sit — the prices at which they have historically defended, added, or capitulated. It cannot: tell you their intent, their stop levels, or whether this cycle's cluster behaves like the last one. My take: the zone is best read as a referendum line. If it holds through a hawkish surprise, the accumulation thesis gains its strongest price-side evidence yet; if it breaks, the same 4.8B XRP becomes overhead supply for months.

Part 3 — Institutional custody: 1.07B XRP and counting

The spot ETF complex now custodies roughly 1.07 billion XRP — about 1.7% of circulating supply (SoSoValue via 24/7 Wall St., Sep 10; ~1% of total supply on the tracker's August reading of ~1.11B). The custody grew through both halves of an ugly year: $1.70B contributed against ~$1.45B of net assets — a ~$250M mark-to-market gap absorbed without net redemptions (SoSoValue; calculated).

(Q2 2026 13F filings via James Seyffart / Yellow.com, Aug 31; retail ~84% of complex inflows.) Scale check: Bitcoin's spot complex holds ~$55.6B cumulative since January 2024 and took ~$3.52B in August alone — roughly 22 times XRP's record month (SoSoValue, September tally). XRP's custody stack is small at market scale and historic at complex scale. My take: 1.7% of float sounds trivial until you remember it is the fastest-compounding line on the map — the only holder class that grew every single week through a 27% YTD drawdown.

Part 4 — Escrow, burn, and the standing seller

The ledger's own mechanics remain the largest single force on the supply side: 32.445B XRP in escrow releasing ~1B monthly, typically 60–80% re-escrowed — roughly a 300M net monthly drip (XRPScan, Jul 26; Yellow.com). Against it: 14.37M XRP permanently burned via fees since inception, and 8.02M activated accounts ledger-wide (XRPScan, Jul 26). My take: escrow is best modeled as a standing limit seller whose size is known and whose schedule is published — the rare supply overhang that everyone can see. August's fund bid (~$159.2M) did not out-absorb it in unit terms; it out-signaled it, which is a different and more fragile achievement.

Part 5 — The non-speculative stack: RLUSD

The newest holder class does not hold XRP at all — it holds the dollar side of the ledger. RLUSD reached $2.02B at the August 25 crossing and $2.32B in September's compilation (RWA.xyz via Crypto Times, Aug 25; via crypto.news), with the XRPL side above $1B — over 90% of the ledger's stablecoin supply (Ripple via COINOTAG, Aug 28). Its holder profile is the wholesale mirror of the ETF complex's: 84,630 holders, $11.81B monthly transfer volume (+17% MoM) on 8,252 active addresses (−25.2% MoM) (RWA.xyz via Crypto Times, Aug 25). My take: every RLUSD dollar on XRPL is a dollar of stablecoin demand that competes with XRP for the ledger's trust and throughput but not for its price. It is the stack that grows without adding sell pressure — the map's only free lunch, and the reason the network-usage line (2.4M daily transactions, +21% YoY; DEX +79% YoY — crypto.news, September compilation) can rise while active accounts fall 40%.

Confirming and falsifying the accumulation thesis

Risks to the map

• Snapshot dates are not uniform. The ledger series carry July 26 dates; the market series run into September. The map mixes vintages by necessity and labels each — anyone averaging across vintages without dates will get it wrong.
• Custody ≠ sentiment. ETF holdings are legally distinct from directional bets; some share is market-making and hedged inventory, invisible in the 13F lens.
• The zone is history, not physics. Cost-basis clusters describe what holders paid; a macro shock can move all of them at once.
• Escrow's schedule is published; its re-escrow rate is not guaranteed. A shift from 80% re-escrow toward 50% would double the net drip without any headline.

Bottom line

The stack, mapped: ~62.7B investable (by estimate), 4.8B of it with cost basis at $1.31–$1.38, 1.07B locked in fund custody and compounding weekly, 32.445B standing in escrow on a published schedule, 14.37M burned out of existence, and $2.32B of dollar demand parked on the ledger in RLUSD. It is a market where the biggest buyer runs on a calendar, the biggest seller runs on a schedule, and everything in between is negotiation. The September 16 decision does not change the map — it reveals which parts of it were load-bearing.

FAQ

Related reading

• Who Keeps Buying? The Flows–Price Divergence, Dissected
• XRP Market Report — August 2026
• XRP Ledger Weekly Data Report — Week of September 7–13, 2026
• The September Decision Stack: How XRP Markets Are Positioned for FOMC Week