Bitcoin rose 39.0% between the closes of 29 June and 29 September 2026 [1]. Over roughly the same stretch, the 10-year U.S. Treasury real yield, the after-inflation return on government debt, climbed 70 basis points to 2.90%, its highest level since November 2008 [2]. By the usual macro playbook, Bitcoin should have struggled to win that fight.

Bitcoin remained sensitive to real yields even as it rallied. On a daily basis Bitcoin still falls when real yields rise [1][2]. The rally came from crypto's own plumbing. U.S. spot exchange-traded funds (ETFs) swung from $5.39 billion of net outflows in the first half to $6.51 billion of inflows in the third quarter [3]. A crowded short book was forced to cover. And the June low held well above the thresholds that marked earlier cycle failures [4][5].

The signals of a genuine liquidity trade are absent: a weaker dollar, growing stablecoin supply, accelerating money growth [6][7][8]. We respect the rally. We just don't trust it yet.

The Contradiction

The bond market spent the third quarter raising the price of time. Bitcoin rallied despite the rise in yields.

Consider what it rallied into. The 10-year real yield went from 2.20% at the end of June to 2.90% on 28 September [2]. The nominal 10-year reached 5.24% and the 30-year 5.56%, levels last seen in June 2007 and June 2004 [9][10]. On 16 September the Federal Reserve raised its policy range to 3.75–4.00%, its first hike since 2023 [11]. A three-month Treasury bill now pays 4.28% for doing nothing [12].

For an asset with no coupon, no dividend and no rent, that is an especially difficult environment.

Bitcoin gained 39% [1].

So the title holds, narrowly and literally. For one quarter, Bitcoin moved one way while the bond market pushed the other.

That quarterly divergence does not establish that Bitcoin has decoupled from real yields. Over the last 60 trading days, the correlation between Bitcoin's daily returns and daily changes in the 10-year real yield was −0.33 [1][2]. Rates still bite. On 23 September the 10-year jumped 15 basis points, its largest one-day rise since November 2024 [9], and Bitcoin fell 2.1% [1]. An asset that had genuinely escaped rate sensitivity would not behave like that. Bitcoin is still on the bond market's leash; for three months it simply pulled harder than the leash.

The distinction is the whole memo. Rallying through rates means another force overwhelmed a headwind that was still blowing. Decoupling from rates would mean the headwind had stopped mattering. The daily data supports the first and not the second.

It also helps to remember who needed the rally. Through 29 September the Nasdaq 100 was up 20.2% for the year; Bitcoin was down 4.4% [6][1]. The Chicago Fed National Financial Conditions Index, a weekly gauge of how easy it is to borrow and take risk, sat at −0.548 for the week to 25 September [13], looser than its long-run average. The backdrop was already supportive of risk-taking; Bitcoin was recovering from its own downturn. Much of the third quarter was Bitcoin clawing back a crypto-specific bear market while the bond market kept raising the cost of the climb.

Figure 1. Bitcoin rose while U.S. 10-year real yields repriced higher. Sources: Perplexity Finance BTCUSD; FRED DFII10; Axys calculations. The 30 September PCE/GDP marker falls outside the charted price/yield window.

What the Bond Market Is Actually Saying

"Yields up, risk down" is a bumper sticker. The implications for Bitcoin depend on what is driving yields higher.

The Treasury yield can be broken down in two ways. The breakeven is the inflation investors expect. The real yield is what is left after it, the genuine price of lending money. A second cut splits the yield into the expected path of policy rates, where investors think the Fed will hold short rates, and the term premium, the extra pay for tying money up long and bearing supply risk and uncertainty.

On the first cut, the quarter looks lopsided. The 10-year breakeven rose about 10 basis points, to 2.34% [14]. The real yield rose 70 [2]. Nobody is panicking about inflation. They are charging more for money.

The New York Fed's Adrian-Crump-Moench (ACM) model makes the second cut. Of the 72 basis points the 10-year added in the third quarter, the model assigns 44 to a higher expected policy path and 28 to term premium [15]. Fiscal supply is part of the story but the repricing of expected policy is most of it.

For Bitcoin, the distinction is whether higher yields come with any offsetting support.

When yields rise mainly on inflation expectations, investors go looking for scarce assets, and the "digital gold" argument gets attractive. When yields rise mainly on growth, higher rates arrive with stronger earnings and more risk appetite, and a high-beta asset can ride along. Neither is painless, but both hand Bitcoin something in return.

A policy-led rise in real yields offers Bitcoin no comparable support. The central bank is deliberately making money more expensive after inflation, which lifts the hurdle every asset has to clear and does nothing to boost the scarcity story. It also tends to support the dollar. For an asset whose entire return is price appreciation, that is the hardest version of the headwind: the cost of waiting goes up and nothing offsets it.

That is the version the third quarter delivered. The Fed did not mince words on 16 September: a 12–0 vote, "economic activity is expanding at a solid pace," "inflation remains elevated" [11]. The median projection in the Summary of Economic Projections (SEP) puts the policy rate at 4.1% by year-end, which is one more hike [16]. The data gave cover. S&P Global's flash composite purchasing managers' index (PMI) printed 58.4 in September, the strongest since July 2021 [17], and energy prices were up 16.3% on the year to August [18]. There is a growth element, but it is the growth that invites hikes, not the growth that makes them unnecessary.

A policy-driven rise in real yields hits a non-yielding asset head on, and it quietly guts the trades that keep capital parked in crypto. Deribit's dated futures priced an annualized basis of roughly 5.0–5.3% [19], barely a point over the 4.28% on a Treasury bill [12]. When the carry trade pays a point more than cash for a great deal more risk, the carry crowd goes elsewhere.

Then 30 September muddied the picture, in Bitcoin's favour. August core personal consumption expenditures (PCE) inflation, the Fed's preferred gauge excluding food and energy, came in at 3.0% year on year and 0.2% on the month, after an annual revision reaching back to 2021 [20]. Second-quarter real gross domestic product (GDP) growth was revised up from 1.5% to 2.2% [21]. Real consumer spending rose 0.6% in August, even as real disposable income went nowhere [20]. By that morning, CME FedWatch put the odds of an October hike at 34.9%, down from 70.9% a week earlier [22]. New York Fed President John Williams had already told an audience, as Reuters reported, that there was "no need for urgency" [24].

So both markets may be partly right. Growth is firmer, inflation was revised softer, and the policy impulse behind real yields may be topping out.

Topping out is a long way from reversing. The policy rate has not moved, and the SEP still pencils in another hike [16]. The latest official print of the 10-year real yield is still 2.90% [25]. And the long end did not exactly throw a party: the iShares 20+ Year Treasury Bond ETF (TLT) slipped in the half hour after the release [23]. A softer inflation print lowers the odds of the next hike. It does not refund the last 70 basis points though.

Why Bitcoin Recovered Anyway

If macro didn't lift Bitcoin, follow the money that did.

The ETF bid came back with force. On Farside Investors' aggregation of fund-level data, U.S. spot Bitcoin ETFs bled $5.39 billion in the first half of 2026, then took in $6.51 billion in the third quarter [3]. The week to 25 September was the largest weekly inflow of the year [3]. Converted at daily prices, the quarter's buying comes to roughly 83,000 BTC (Our calculation). Since the April 2024 halving, the network issues 3.125 BTC per block, about 450 BTC a day [26], or roughly 41,000 BTC across the quarter. The funds bought about twice what the miners produced. In the third quarter, the sellers only showed up higher.

The comparison with issuance matters because miners must sell to fund their operations, so new supply is the sell flow the market faces every day. A buyer that absorbs all of it, and as much again, has to pry the rest from existing holders, who sell at prices of their choosing. That is the mechanical core of the rally, and its weak point: remove the ETF bid and the miner supply is still there, with no one absorbing it.

Figure 2. Estimated BTC absorbed by U.S. spot ETF flows exceeded estimated newly mined supply in Q3. Sources: Farside Investors; Perplexity Finance BTCUSD; CoinGecko; Axys calculations. ETF conversion uses daily net flows divided by the BTC daily close; 450 BTC/day is an average issuance assumption.

There is a difference, though, between a strong marginal buyer and a broad buyer base. A strong marginal buyer is one large, price-insensitive participant whose purchases set the price at the margin. A broad base is many buyers with different motives and funding sources: corporate treasuries, crypto-native capital arriving through stablecoins, leveraged traders, long-term savers. It keeps working when any one of them steps back.

A marginal buyer moves a market faster. A broad base keeps it moving longer. The third-quarter rally was driven primarily by the ETF bid.

It was also, overwhelmingly, one buyer. BlackRock's iShares Bitcoin Trust (IBIT) took about 65% of September's net inflows [3] and held $66.8 billion in net assets on 28 September [27]. A market leaning that hard on one fund family's distribution is driving down a single-lane road. Concentrated flows tend to move together, because the platforms and allocation committees that route through the same product add and trim together. They are also volatile. They turned from two weeks of outflows to the year's largest weekly inflow within a month [3], and they can turn back just as fast. A liquidity regime builds up gradually and fades gradually. An ETF flow can reverse in a single session.

Then came the squeeze, and the timing was almost too neat. On 15 September, a procedural vote to advance the CLARITY Act, the crypto market-structure bill, failed 49 to 50 in a Senate that needed 60 [28]. The Fed hiked the next day [11]. Bitcoin closed 15 September at $75,584 [1]. Two bad headlines in two days is exactly when shorts get comfortable.

They did not stay comfortable for long. ETF inflows resumed on 17 September and hit $999 million on 21 September [3]. That same day, CoinGlass data reported by The Block counted about $648 million of crypto shorts liquidated over 24 hours [29]. Funding on Deribit's perpetual contracts, the fee longs pay shorts to keep leveraged bets open, spiked to an annualized 21.4% [30].

The more interesting part came after. By 26–30 September, funding had collapsed to near zero [30], yet price held in the low-to-mid $80,000s [1]. Pure short-covering rallies tend to give it all back once the forced buyers are gone. This one hasn't. Not yet.

Beneath the flows, the floor held. Bitcoin closed 30 June at $58,524 and touched $57,718 intraday on 1 July [1]. Glassnode notes that price never closed below the realized price, the average cost basis of every coin in circulation, during this bear market [4]. The peak-to-trough drawdown was about 54% [1]; the previous completed cycles fell between 77% and 85% [5]. By Bitcoin's own ugly standards, this was a shallow bear.

Bitcoin's technical position has also improved. Price reclaimed the 200-day moving average on 19 August [1]. The 50-day crossed above the 200-day on 8 September, the golden cross chartists love to announce [1]. On 21 September Bitcoin closed at $86,595, clearing the May swing high of $82,814 and marking its best close since 28 January [1].

Now it has walked into a crowd of holders waiting to break even. Bitcoin sits inside the largest cluster of long-term-holder cost basis, $84,000 to $85,000, where most coins held for six months or more last changed hands [4]. Support is near Glassnode's True Market Mean of about $77,000. Resistance is near its mean market-value-to-realized-value (MVRV) price of about $96,700 [4].

The technical improvement does not yet establish that the recovery can last.

Why This Is Not Yet a Liquidity Trade

"Liquidity trade" is the phrase crypto reaches for whenever prices rise, because it sounds structural and implies the rally will keep going. A real one needs money getting cheaper and more plentiful, lifting assets that have no business moving together.

The dollar hasn't cooperated. The U.S. Dollar Index (DXY) closed at 101.37 on 29 September, up 3.1% on the year [6]. Easing global liquidity usually shows up as a softer dollar first. It hasn't shown up.

Stablecoins aren't growing. The dollar-pegged tokens that serve as crypto's working cash totalled $311.2 billion on 30 September, 3.0% below their 17 May peak [7]. Fresh money usually passes through here on its way in.

Global money growth is fading. Our proxy, combining the M2 money supply of the United States, China, the euro area and Japan in dollars, hit a record of about $103.5 trillion in August, but its three-month growth rate slid from 4.3% in February to 1.1% [8]. The stock is large. The incremental flow is fading.

The corporate buyer has gone quiet. Listed companies bought about 5,900 BTC in the three months to 14 September, against 89,000 BTC in July 2025 alone [31]. The balance-sheet buyers who powered much of 2025 have largely stopped buying.

Credit is starting to grumble. The high-yield option-adjusted spread (HY OAS), the premium investors demand to hold junk bonds, widened from 2.60% on 28 August to 3.02% on 28 September [32]. Conditions remain loose, but the marginal impulse is less supportive.

And holders are cashing out. By 28 September Glassnode reported realized profit-taking and short-term holders' unrealized gains "well above their bands," with nearly three-quarters of supply in profit [33]. Profit-taking is harmless while the marginal buyer is growing. That buyer is not growing: daily ETF inflows fell from $999 million on 21 September to $31 million on 28 September and $66 million on 29 September [3].

Meanwhile, rates still land daily punches. The −0.33 correlation with real-yield changes describes the market as it trades today [1][2], and the 23 September session showed what that looks like in practice: yields spiked, Bitcoin fell [9][1].

A firm dollar, flat stablecoin supply, and a missing corporate bid. As liquidity trades go, it is remarkably short of liquidity.

What the Market Is Pricing

It is tempting to frame the quarter as a duel in which one market must be wrong. We think that misreads both.

The bond market is pricing a higher real cost of capital. A 2.90% real yield [2], a policy range of 3.75–4.00% [11], and an SEP that still pencils in another hike [16] describe a world in which money stays expensive after inflation for some time. The 30 September data softened the edges without changing the shape. Growth was revised firmer [21], inflation was revised softer [20], and October hike odds fell [22], but the long end did not rally [23]. Bonds are saying the economy is strong enough to carry expensive money and that the Fed is not yet done making it expensive.

Bitcoin is pricing something narrower. It is not forecasting a liquidity boom; the dollar, stablecoin supply and money growth would look very different if it were [6][7][8]. It is pricing the possibility that ETF demand stays large for long enough that the macro headwind stops getting worse before the buyer gets tired. That is a bet about persistence and timing, not about the direction of global money.

Those claims run on different clocks. The bond market is pricing policy and real rates over years. Bitcoin's third-quarter move was a quarter's worth of flow meeting a quarter's worth of supply. Real yields can stay elevated for a long time while a determined buyer absorbs every mined coin for a few more months. Then both markets were right about what each was pricing.

Where the two genuinely collide is at the margin. Every extra basis point of real yield makes the ETF buyer's job harder; every softer inflation print, like 30 September's, buys that buyer time. The daily correlation of −0.33 [1][2] is that collision in miniature.

Bitcoin's outlook therefore depends on ETF demand remaining strong until the rate backdrop stops deteriorating.

The market has given both sides evidence. It has not yet given either side permission to be complacent.

Our Take

The rally is real. We won't wave it away because the macro backdrop is hostile, and we won't crown it a new liquidity cycle because the chart finally looks respectable.

Bitcoin went up because a large buyer came back through the ETFs and absorbed about twice the network's new supply, because traders who bet on a fall were forced to buy back in, and because the June low held above the levels that marked deeper failures in earlier cycles.

That earns tactical respect for now but it doesn't earn our strategic trust.

Almost everything holding the price up runs through one channel: U.S. ETF demand, concentrated in one fund family, and already slowing. The bond market, meanwhile, is still charging a 2.90% real yield, and cash still pays more than 4% [25][12]. Bitcoin has rallied despite a rate backdrop that remains restrictive.

The next few months come down to a race between two lines. If ETF buying continues to exceed the roughly 450 bitcoin mined each day while real yields stop climbing, Bitcoin could break above its recent range. Softer inflation and firmer growth make that a plausible scenario. If real yields push higher while the ETF bid fades, Bitcoin has to stand on its own, and the first honest test comes around $77,000 [4].

We will watch those two lines. The narrative can look after itself.

Outbid, Not Decoupled

Bitcoin won the quarter on price, but it has not shown that rates no longer matter. Real yields sit near a seventeen-year high, the dollar is firm, and nearly all the buying has come through one channel. What passes for independence is one very large bid.

The bond market hasn't been proven wrong. It has been outbid, and bids disappear a lot faster than yields fall.

Disclaimer

This material is provided for information purposes only and does not constitute investment advice, a recommendation, or an offer or solicitation to buy or sell any financial product or service in any jurisdiction.

The views expressed are those of the author as at the date of publication and are subject to change without notice. While reasonable care has been taken, no representation or warranty, express or implied, is made as to the accuracy, completeness, or reliability of the information contained herein, and no reliance should be placed upon it.

This material has not been reviewed or approved by the Dubai Financial Services Authority, the British Virgin Islands Financial Services Commission, or the Cayman Islands Monetary Authority, or any other regulatory authority.

This material is not intended for distribution to, or use by, any person in any jurisdiction where such distribution or use would be contrary to local law or regulation. In particular, this material does not constitute a public offer of securities in the British Virgin Islands or the Cayman Islands.

Any investment or service referred to herein will be available only to persons who meet the relevant eligibility criteria under applicable laws and regulations, including, where applicable, Professional Clients (as defined by the Dubai Financial Services Authority) or other equivalent categories of sophisticated or professional investors under applicable law.

Investments involve risk, including the possible loss of capital. Past performance is not a reliable indicator of future results.

“Axys” is a trading name used by a network of independent businesses. Axys does not operate as, and is not, a separate legal entity in any jurisdiction and has no distinct legal personality. Each business operating under the Axys name is independently owned and operated.

Axys Capital  Ltd is authorised and regulated by the Dubai Financial Services Authority. Axys Investment Management Ltd is authorised and regulated by the British Virgin Islands Financial Services Commission.

References

1. Perplexity Finance, BTCUSD daily price history (open, high, low, close), data through 29 September 2026 close; 60-day correlation and moving averages are Axys calculations. https://www.perplexity.ai/finance/BTCUSD

2. Federal Reserve Bank of St. Louis (FRED), 10-Year Treasury Inflation-Indexed Security, Constant Maturity (DFII10), data through 28 September 2026. https://fred.stlouisfed.org/series/DFII10

3. Farside Investors, Bitcoin ETF Flow — All Data (US$m), data through 29 September 2026; BTC-equivalent conversion is an Axys calculation. https://farside.co.uk/bitcoin-etf-flow-all-data/

4. Glassnode, "The Week On-chain, Week 38 2026: Escape Velocity," 23 September 2026 (data to 21 September). https://research.glassnode.com/the-week-onchain-week-38-2026/

5. Galaxy Research (Alex Thorn), "Bitcoin Halving Cycle: Where BTC Could Bottom," 12 June 2026. https://www.galaxy.com/insights/research/bitcoin-four-year-cycle-where-is-the-bottom

6. Perplexity Finance, daily closes for the Nasdaq 100 (^NDX) and U.S. Dollar Index (DX-Y.NYB), data through 29 September 2026. https://www.perplexity.ai/finance/

7. DefiLlama, Stablecoin Charts (total circulating, USD), data through 30 September 2026. https://defillama.com/stablecoins

8. Axys calculation of four-economy M2 in U.S. dollars from People's Bank of China, European Central Bank, Bank of Japan and Federal Reserve series (via Perplexity Finance macro data), converted using FRED H.10 exchange rates DEXCHUS, DEXJPUS and DEXUSEU; data through August 2026. https://fred.stlouisfed.org/series/DEXUSEU

9. FRED, 10-Year Treasury Constant Maturity (DGS10), data through 28 September 2026. https://fred.stlouisfed.org/series/DGS10

10. FRED, 30-Year Treasury Constant Maturity (DGS30), data through 28 September 2026. https://fred.stlouisfed.org/series/DGS30

11. Board of Governors of the Federal Reserve System, FOMC Statement and Implementation Note, 16 September 2026. https://www.federalreserve.gov/monetarypolicy/files/monetary20260916a1.pdf

12. FRED, 3-Month Treasury Constant Maturity (DGS3MO), data through 28 September 2026. https://fred.stlouisfed.org/series/DGS3MO

13. FRED / Federal Reserve Bank of Chicago, National Financial Conditions Index (NFCI), week ending 25 September 2026. https://fred.stlouisfed.org/series/NFCI

14. FRED, 10-Year Breakeven Inflation Rate (T10YIE), data through 29 September 2026. https://fred.stlouisfed.org/series/T10YIE

15. Federal Reserve Bank of New York, ACM Treasury Term Premia (daily), data through 28 September 2026. https://www.newyorkfed.org/research/data_indicators/term-premia-tabs

16. Board of Governors of the Federal Reserve System, Summary of Economic Projections, 16 September 2026. https://www.federalreserve.gov/monetarypolicy/fomcprojtabl20260916.htm

17. S&P Global Market Intelligence, "US flash PMI signals fastest growth for over five years in September," 23 September 2026. https://www.spglobal.com/market-intelligence/en/news-insights/research/2026/09/us-flash-pmi-signals-fastest-growth-for-over-five-years-in-september

18. U.S. Bureau of Labor Statistics, Consumer Price Index — August 2026, released 11 September 2026. https://www.bls.gov/news.release/archives/cpi_09112026.htm

19. Deribit, public API book summary for BTC futures, snapshot 30 September 2026, 12:50 UTC; annualized basis is an Axys calculation. https://www.deribit.com/api/v2/public/get_book_summary_by_currency?currency=BTC&kind=future

20. U.S. Bureau of Economic Analysis, Personal Income and Outlays, August 2026, released 30 September 2026. https://www.bea.gov/news/2026/personal-income-and-outlays-august-2026

21. U.S. Bureau of Economic Analysis, Gross Domestic Product, Second Quarter 2026 (Third Estimate), released 30 September 2026. https://www.bea.gov/taxonomy/term/451

22. CME Group, CME FedWatch Tool, October 2026 meeting probabilities, data as of 30 September 2026, 08:10:58 CT. https://www.cmegroup.com/markets/interest-rates/cme-fedwatch-tool.html

23. Perplexity Finance, intraday five-minute bars for TLT, 30 September 2026, 12:25–13:10 UTC. https://www.perplexity.ai/finance/TLT

24. Reuters, "Fed's Williams sees no urgency for next Fed rate hike," 29 September 2026. https://www.reuters.com/business/feds-williams-sees-no-urgency-next-fed-rate-hike-2026-09-29/

25. Board of Governors of the Federal Reserve System, H.15 Selected Interest Rates (Daily), values for 28 September 2026. https://www.federalreserve.gov/releases/h15/

26. CoinGecko, Bitcoin Halving Countdown (20 April 2024 halving at block 840,000; 3.125 BTC block reward). https://www.coingecko.com/en/coins/bitcoin/bitcoin-halving

27. BlackRock iShares, iShares Bitcoin Trust ETF (IBIT), net assets as of 28 September 2026. https://www.ishares.com/us/products/333011/ishares-bitcoin-trust-etf

28. United States Senate, Roll Call Vote 119th Congress, 2nd Session, Vote Number 234, 15 September 2026. https://www.senate.gov/legislative/LIS/roll_call_votes/vote1192/vote_119_2_00234.htm

29. The Block, "Bitcoin taps $85,000 for first time since January as crypto short liquidations surge" (CoinGlass data), 21 September 2026. https://www.theblock.co/news/markets/2026-09-21-bitcoin-taps-85000-for-first-time-since-january-as-crypto-short-liquidations-surge-415905

30. Deribit, public API funding-rate history for BTC-PERPETUAL, 30 August–30 September 2026; annualization is an Axys calculation. https://www.deribit.com/api/v2/public/get_funding_rate_history?instrument_name=BTC-PERPETUAL&start_timestamp=1756684800000&end_timestamp=1790812800000

31. Glassnode, "The Week On-chain, Week 37 2026: Breakdown into Thin Support," 16 September 2026. https://research.glassnode.com/the-week-onchain-week-37-2026/

32. FRED / ICE Data Indices, ICE BofA US High Yield Index Option-Adjusted Spread (BAMLH0A0HYM2), data through 28 September 2026. https://fred.stlouisfed.org/series/BAMLH0A0HYM2

33. Glassnode, "BTC Market Pulse: Week 40," 28 September 2026. https://research.glassnode.com/btc-market-pulse-week-40-2026/