Spot bitcoin ETFs began July on a record outflow streak and ended it taking money again, and bitcoin came out the other side within about 5% of where it went in.
On 5 July, The Block reported a record eighth straight week of net outflows from US spot bitcoin ETFs, about $527 million shed in a holiday-shortened week, with spot ether ETFs tying their own worst stretch on record [1]. Seventeen days later the same complex had logged six consecutive days of net inflows totalling roughly $779 million since 13 July, with bitcoin above $65,000 [2]. Bitcoin traded at $61,201.59 on 2 July [3] and closed the month at $64,274 on 31 July [4].
A number that can travel from a record outflow streak to a six-day inflow run without moving the asset more than a few thousand dollars is not the number setting the price. That is July's tell, and it reframes everything else the month did.
The round trip, in order
- June closed with $4.5 billion of net outflows from US spot bitcoin ETFs, the largest monthly outflow since their January 2024 launch, against $51.15 billion of cumulative inflows still sitting in the products [5].
- 3 July — $222 million of inflows snapped a ten-day losing streak, following a weak jobs report and softer Fed signals [6].
- 7 July — $265.7 million, the biggest single-day inflow since 5 May, with bitcoin around $63,000 [7].
- 11 July — the first positive weekly inflow since May at $197.4 million, characterised in the coverage as modest and routine rather than a catalyst [8].
- 22 July — six straight inflow days, roughly $779 million since 13 July, bitcoin holding above $65,000 [2].
- 31 July — $233 million into bitcoin funds and $13.3 million into ether funds, bitcoin up 0.35% on the day to $64,274 [4].
And the counterweight, from a holder rather than a wire. On 24 July, Michael Saylor's own market snapshot put bitcoin at $64,131, 49% below its all-time high, with net ETF outflows of $1.931 billion over the trailing 30 days and 30-day historical volatility at 27% [9].
Both readings are true at once. A month can contain a six-day inflow run and a negative trailing-30-day net simultaneously — that is what a round trip looks like from the inside, and it is why a single monthly flow headline was the wrong instrument for July.
The wrapper went quiet as a venue
Flows are one measure of an ETF complex; turnover is another, and it moved one way all month.
Glassnode reported on 14 July that US spot bitcoin ETF trading volume had fallen 78% from its recent peak across ARKB, GBTC, FBTC and IBIT, reading it as muted institutional participation [10]. By 25 July the products had traded $8.05 billion across a week — the lowest five-session total since October 2024 — while spot ether ETFs took $103.9 million of net inflows against $33.8 million into bitcoin funds, more than three times as much [11].
That ether-over-bitcoin split was not a single session. It showed up on 9 July, when bitcoin funds shed $85 million and ether funds took $70.5 million [12], and Coin Bureau was still pointing at ether ETF inflows on 30 July [13].
So the wrapper was not just flow-neutral in July. It was quieter as a trading venue than at any point since October 2024, and the flows it did take were increasingly not bitcoin's.
What actually arrived
Set the flow tape aside and July's institutional news has a shape.
- Visa announced an internal stablecoin platform to expand use of USD-pegged tokens among the financial institutions and merchants it services, launching with the Open Standard stablecoin OUSD [14]. Cointelegraph put the reach at more than 200 million merchants worldwide [15].
- PayPal added native issuance of PYUSD on Polygon [16].
- Stripe was reported to have moved $53 billion into stablecoins [17].
- DTCC began a limited-production trial of tokenized stocks and US Treasurys involving nearly 40 firms, including JPMorgan, Microsoft and Circle [18].
- JPMorgan tokenized the Invesco QQQ Trust [19].
- Circle received final OCC approval for First National Digital Currency Bank — custody for Circle affiliates first, the USDC reserve later [20] — and closed the month with a limited-purpose trust charter from NYDFS [21]. It also acquired nearly 1,000 IBM blockchain patents, which Cointelegraph reported makes it the biggest US holder [22].
- Sony Bank took conditional OCC approval for a national trust bank, Connectia Trust, planning a dollar-backed stablecoin next year [23].
- The US and UK jointly published a policy line on folding regulated stablecoins into cross-border finance [24].
Underneath the announcements, the balance sheet grew. Securitize passed $5 billion of tokenized assets with the global tokenization market at an all-time high of $34 billion [25], and Tether reported $1.5 billion of Q2 net operating profit on USDT supply of $184.6 billion, over 60% of the stablecoin market [26].
Every item in that list moves dollars or existing securities across crypto rails. None of them is a new way to take crypto price exposure. That is the sense in which July's institutional money arrived: on the settlement side of the business, not the ownership side.
The other side of it
The ownership side did not empty out, and it would be wrong to read July that way.
- Morgan Stanley listed ether and solana exchange-traded products on the NYSE at a 0.14% sponsor fee, described as the lowest in the US, with staking rewards attached [27].
- T. Rowe Price debuted an actively managed multi-token spot crypto ETP on NYSE Arca, rotating among bitcoin, ether, XRP, solana and Hyperliquid [28].
- Public companies added 110,000 BTC in the second quarter, a 1.8x jump [29].
- Long-term holders absorbed supply straight through the outflows. Glassnode flagged accumulation while bitcoin sat below $60,000 on 1 July [30], and The Block tied the reclaim of $61,000 the next day to the same behaviour [3].
So the wrapper business kept building while the wrapper's flows went sideways. Read one way, that is product arriving ahead of the demand for it. Read the other, it is issuers positioning for a cycle the flow tape has not started yet. July does not settle which.
The macro that did not decide it
July was not a quiet month for the price of money, which makes the price's stillness more interesting rather than less.
June CPI came in at 3.5% year-over-year against a 3.8% forecast [31] — a 0.4% monthly decline, the steepest since 2020, which lifted bitcoin above $63,700 and later over $64,000 [32]. The Federal Reserve then left rates unchanged on 29 July [33], and the dollar posted its sharpest drop in two weeks [34]. Pulling the other way, the 30-year US Treasury yield reached its highest level since 2007 [35], and US publicly held debt passed 100% of GDP for the first time since WWII [36]. Bitcoin held near $65,000 into month-end [37].
A soft inflation print, a central-bank hold, a falling dollar, a long yield at 2007 levels and a debt ratio at wartime levels — and the asset finished a few thousand dollars from where it started. The macro tape was loud and the price was not listening to it either.
Where the attention went
Across the sources we track, the month's rotation followed the news rather than the tape. Among July's ten most-discussed sectors, the exchange and venue layer gained the most share of the conversation and the ETF wrapper lost the most, with infrastructure and layer-1 protocols also gaining ground. Talk about "the crypto market" as a single object fell sharply from June's level. The curated voices moved to the plumbing before the plumbing became the headline.
What would settle it
Two claims, both falsifiable, and neither requires the other to be false.
1. The flow variable is on pause. If August's ETF flows resume moving the price — a large net week landing with a proportionate move in bitcoin — then July was a lull in a working transmission channel, and this piece describes a fortnight rather than a handover.
2. The rails took over. If the settlement-side launches keep shipping while flows stay noise — another card network, another trust charter, another custodian trial — then the ownership wrapper has become one product line among several rather than the market's main event.
The informative outcome is the divergence. If both resolve true at once, the honest reading is that crypto acquired a second demand curve during 2026 — one that buys settlement rather than exposure — and that the first one never went away, it simply stopped being the one that explained the tape.
