Who Keeps Buying? The Flows–Price Divergence, Dissected

The question

XRP trades at $1.37, down 27% on the year and 62.5% below its all-time high. Its exchange-traded funds just crossed $1.70 billion in cumulative net inflows — a record, set in the middle of a down year. The funds hold 1.07 billion XRP, about 1.7% of circulating supply. And the number that makes the whole thing strange: total net assets are about $1.45 billion, roughly a quarter-billion dollars below what investors contributed. Money came in; value did not stay.

This report takes the divergence apart: what the two ledgers say, who the marginal buyer actually is by the available evidence, what the network's own data adds, and what would confirm or falsify each explanation — starting with the Federal Reserve's September 16 decision.

Part 1 — The two ledgers

Every flow story needs its baseline, so here is the ETF complex's history in one table, all figures from SoSoValue as reported by 24/7 Wall St., Yahoo Finance, Gate and Finobird:

Read against the price — $1.90 to $1.06 to $1.70 to $1.37 over the same span — the ledger shows three distinct buyer behaviors: a launch-wave cohort that front-loaded, a momentum cohort that bought August's recovery, and a persistent residual that kept adding through the September fade. The third group is the interesting one, because it is the only one still transacting.

Part 2 — Who the marginal buyer is

The 13F and flow evidence lets us sketch the buyer rather than guess:

• Goldman Sachs is the complex's largest institutional holder, at ~$87.4M as of Q2 2026 filings (Gate, 13F data). Jane Street and Millennium are also reported among major holders.
• September's inflows were concentrated: the entire $12.29M on Sep 9 came through Bitwise and Grayscale funds; on Sep 8, only Franklin's XRPZ printed positive. When demand is this thin, single allocators move the daily print.
• 24/7 Wall St.'s reporting attributes the buying to institutional and RIA scheduled buying rather than retail conviction — rebalancing calendars, not momentum accounts.
• Canada's CDCC approved options on Evolve's and Purpose's XRP ETFs (Sep 10, US sales authorized) — the wrapper is beginning to serve covered-call and hedging strategies, which bring a different holder: the income desk.

Put together: the marginal buyer in September is a small set of institutional allocators running scheduled programs — the same profile that kept Bitcoin's complex alive through its own drawdowns, at one-thirty-third the scale. That is consistent with the concentration, consistent with the persistence through drawdowns, and consistent with the March–May precedent where the weak hands redeemed and the programs stayed. What it is not consistent with is a broad new channel opening: the flows are real, but five funds sharing ~$19M a week is a niche allocation, not a wave.

Part 3 — The other ledger: what the network says

The flows question has an on-chain twin, and the network's own numbers answer it in parallel. XRP Ledger data reported by crypto.news shows 2.4M daily transactions (+21% YoY) and DEX volume +79% YoY, on 40% fewer active accounts. The same report carries the stablecoin ledger: RLUSD at $2.32B, +1,278% YTD, with integrations at JPMorgan, Mastercard, Convera and Interactive Brokers — and the XRPL-side share of RLUSD supply climbing from 18.4% at the start of 2026 toward half (~$963M as of Aug 27).

Both stories are the same story wearing different clothes: the ecosystem's demand is consolidating into fewer, larger, programmatic participants — ETF allocators on the wrapper side, settlement institutions on the ledger side — while the broad retail base that used to define XRP thins out. The ETF divergence is what that looks like in the fund ledger. The transactions-up-accounts-down split is what it looks like on-chain. Neither is bullish or bearish by itself; both are directional for who owns the asset next.

And the newest wiring between the two ledgers appeared this week: Sentora approved FXRP as the first XRP collateral for a $280M RLUSD vault on Ethereum — XRP posted as collateral, RLUSD borrowed against it. It is the first credit-scale mechanism that needs both assets at once, and its usage data over Q4 is a cleaner test of institutional XRP demand than any flow print.

Part 4 — What would confirm or falsify each read

The value-trap row deserves honesty: for nine months, every institutional datapoint has improved while the token fell 27% from its January level. The base rates say wrapper adoption and network usage do not reprice an asset by themselves — Bitcoin's $55.6B ETF complex did not stop drawdowns either. What these flows do is change the holder base, and holder bases determine how an asset behaves when the macro cycle turns. That is the mechanism on the table, and it is slow.

Part 5 — Scenario map into the Fed decision

Probabilities, logged under the new scoring policy with sources to be checked in the next edition: 35% / 45% / 20%. The neutral scenario leads because XRP's September tape has tracked the broad risk complex — BTC near $78K, Fear & Greed at 69 — more than its own news, and the market does not price a large Fed surprise as the base case.

Risks to this whole framework

• Flow scale. A record XRP ETF month ($159M) equals roughly 0.19% of the token's market cap. At current size, the complex cannot outrun escrow releases and market supply. The floor thesis only compounds if flows compound.
• Concentration cuts both ways. Two or three funds carrying the entire bid means one allocator pausing shows up as a "streak break." The Sep 2 outflow was $7.2M — one desk's afternoon.
• Data timing. Flow figures settle with a lag, 13F positions are quarterly, and Sep 11–13 session flows were unpublished at press time. The divergence could look different on Monday's data.
• The stablecoin substitution question. RLUSD's growth is the ecosystem's best fundamental datapoint — and the hardest one to convert into token demand. If settlement migrates to RLUSD without the XRPL-side share climbing, the network thrives while the token's fee capture stays marginal.

Bottom line

Someone is still buying, and the evidence says who: a narrow set of institutions on schedules, not a crowd on momentum. They are underwater in aggregate and adding anyway — the single most informative behavioral fact about XRP's market right now. Whether that reads as conviction or as averaging-down will be decided by things neither the ETF complex nor the ledger controls: the Fed on Tuesday, and the macro cycle after it. The structural work — 1.7% of supply in custody, a $2.32B stablecoin, the first XRP-collateralized credit vault — is happening on calendars that do not care about the tape. The market has spent nine months separating those two clocks. The next quarter decides whether they re-connect.

FAQ

Related reading

• XRP & Ripple Weekly News Review — Week of September 7–13, 2026
• XRP Ledger Weekly Data Report — Week of September 7–13, 2026
• XRP Market Intelligence — Week of September 7–13, 2026