How Big Does the Bid Get?

Sizing the Total Addressable Market for Bitcoin in Institutional Portfolios — Treasuries, Funds, Banks, Sovereigns, and Central Banks — and How Fast It Can Realistically Arrive

TON618 Capital Research · As of July 11, 2026 · Download PDF

A point-in-time sizing exercise, not a forecast with a date on it. Data: issuer disclosures (BlackRock, corporate treasury trackers), SEC 13F filings, the AIMA/PwC and Coinbase/EY-Parthenon institutional surveys, WTW's Global Pension Assets Study, Global SWF, the Basel Committee's cryptoasset standard, and primary reporting — every figure as-of dated. Fact ("Data shows…") is separated from opinion ("We believe…"). Prices and AUM figures move fast; treat each as of its stated date. Key claims were run through an adversarial verification pass before release; the Source & Verification Note documents what was confirmed and what remains hedged.

§0 The Argument

Why now. This question is being asked at a moment of maximum dissonance. Bitcoin trades near $63,800 (July 11, 2026), roughly 49% below its 52-week high of ~$126,000. The US spot-ETF complex has shrunk from a $147.5B peak (September 30, 2025) to ~$77B — much of that decline is price, not redemption: global Bitcoin ETF net flows were still positive (+$3.7B) in Q4 2025 even as the price fell 23%, and 13F-filing professional investors grew their ETF ownership 32% across full-year 2025. But the freshest data has turned, and honesty requires leading with it: CoinShares' Q1 2026 13F report shows professional holdings fell from 313k to 261k BTC (−17% in coins, −35% in dollars) in the first quarter, with hedge funds cutting −39% and brokerages −53% — 95% of the selling from the fast money — even as banks quietly more than doubled to 15,200 BTC and Abu Dhabi's sovereign funds kept adding. So the accurate picture is not "the bid is intact" but "the composition is rotating — patient balance sheets in, levered/tactical money out — while the aggregate professional coin count is falling in the drawdown." Corporates did buy a record 110,000 BTC in Q2 2026 (~2.6× that quarter's miner issuance). The institutional bid is being tested, and by the most recent measured data it is partially withdrawing at the fast-money margin — which is exactly why the base case below rests on whether the 2024–25 accumulation regime resumes, a question the current tape answers "not yet."

§1 The Question Has Changed (and the Famous Statistic Is Mis-Cited)

The 2020–2023 institutional debate was existential: is Bitcoin an investable asset at all? That debate effectively ended with three events — the January 2024 launch of US spot ETFs inside the world's largest asset managers; the March 2025 executive order creating a US Strategic Bitcoin Reserve; and the 2025 arrival of formal allocation guidance from BlackRock, VanEck, and Fidelity. The live debate now is sizing: 1% or 5%? Satellite or strategic?

First, we correct the statistic that launched this note. The widely-circulated pairing — "asset managers average ~7% of AUM in crypto, and 71% plan to increase" — is real but routinely mis-scoped. We traced it to its primary source: the 7th Annual Global Crypto Hedge Fund Report (AIMA/PwC, 2025; 122 managers, ~$982B combined AUM). The actual findings:

• 55% of traditional hedge funds have some digital-asset exposure in 2025, up from 47% in 2024;
• among funds already invested, the average allocation is 7% of AUM (up from 6%) — but the distribution is heavily skewed: 52% hold under 2%, 29% hold 2–10%, and 19% hold more than 10%;
• 71% of the invested funds plan to increase exposure within twelve months (67% in 2024, 46% in 2023).

So the honest reading is narrower and, for TAM purposes, more interesting: the 7%/71% figures describe crypto-invested hedge funds only — a self-selected minority of one channel. The broader allocator reality, from the same report's institutional sub-sample ($502B across family offices, funds-of-funds, pensions, endowments, sovereigns, insurers): two-thirds allocate something, over a third of those hold under 2%, and pensions and sovereign funds typically cap the "opportunistic" bucket that contains crypto at 2–5% of AUM, with pilot allocations below 1–2%. The Coinbase/EY-Parthenon surveys (352 institutions in January 2025; 351 in January 2026 — both pre-screened for crypto engagement) point the same way, though the level has cooled with the drawdown: in the January 2026 edition, the share planning to hold more than 5% of AUM in crypto rises from 18% today to 29%, ~three-quarters plan to increase allocations, and 81% prefer to hold via registered vehicles. Data shows intent running far ahead of position. We believe that gap — intent minus position — is the cleanest single description of the institutional TAM.

§2 What Institutions Already Own

Three wrappers now hold a measurable, auditable stock of Bitcoin (as of July 10–11, 2026, BTC ≈ $63.8k):

Together: ~3.08M BTC — 14.7% of terminal supply — worth ≈$197B. But wrapper totals overstate end-owner institutional holdings, because most ETF shares belong to households: 13F-filing institutions held ~28% of US spot-ETF assets in late 2025 (up from ~20% at launch), with registered investment advisors making up 50–60% of that professional slice. Counting only the institutional share of ETFs, plus corporate treasuries, government stockpiles, and the ~$2B of directly-disclosed sovereign/insurer positions, the true end-owner institutional stock is ≈$143B — 0.10% of the pools sized in §3.

Who, specifically, is in? The 13F record is the honest ledger — and it is genuinely two-sided. Buyers through the Q4 2025 drawdown: Mubadala (Abu Dhabi, +2.3k BTC-equivalent; position $565.6M by Q1 2026, adding in every quarter it has disclosed since Q4 2024), Millennium (+8.1k), Morgan Stanley (+1.9k), plus Al Warda (Abu Dhabi, 8,218,712 IBIT shares / ~$408M). Harvard and Brown both disclosed Bitcoin (and, for Harvard, Ether) ETF positions across 2025 — but by the Q1 2026 filing Harvard had cut its IBIT stake ~43% and fully exited its ETH-ETF position, the sharpest reversal in the endowment cohort. But the exits are just as instructive: Brevan Howard cut 17.7k BTC-equivalent, and the State of Wisconsin Investment Board — an early mover that disclosed $321M in Q4 2024 and doubled down — sold its entire IBIT stake in May 2025, ahead of the drawdown. That a pioneer pension fully round-tripped in and out is exactly what a maturing, two-sided institutional market looks like — adoption is not monotonic.

§3 The Denominator: $144 Trillion of End-Owner Pools

TAM math dies by double counting, so we size only end-owner pools — the entities whose balance sheet ultimately bears the exposure — and exclude conduits (mutual funds, hedge funds, ETFs) that intermediate the same dollars. Sources: WTW Thinking Ahead Institute (pensions from the GPAS 2026 edition; insurance/SWF/E&F from the 2025 study), Global SWF (2025), ARK (2025):

Two structural notes. First, the pool itself is growing: state-owned investors alone (SWFs + public pensions) hit a record $60T in 2025 and are projected near $80T by 2030 — the denominator compounds at ~4–5%/yr while the numerator compounds much faster. Second, the pools are not equally reachable: the US is 64.9% of world pension assets, so US regulatory plumbing (ERISA guidance, 401(k) access, wirehouse approval) governs most of the addressable base — and that plumbing moved decisively in 2025–26 (§4).

§4 Channel by Channel: Where the Next Dollar Comes From

A. Asset managers & the ETF wrapper — the on-ramp that already won

The registered fund is the institutional access vehicle, full stop: 66% of surveyed institutions access crypto via ETPs, and 81% prefer registered vehicles. The US complex took in $51.3B net in 30 months; IBIT alone gathered $60.3B of cumulative inflows (offset by $27.3B of GBTC bleed) and reached $46.8B of assets in 2.5 years — a scale GLD needed two decades to reach. The channel's constraint is no longer access but advice: most wirehouse platforms still hold Bitcoin ETFs to unsolicited orders. That is the 2026 swing factor (§4-G).

B. Digital-asset treasuries — big, concentrated, and no longer growing for free

Corporate treasuries hold 1.27M BTC (~$81B) — but Strategy alone is two-thirds of it, and the next names (Twenty One Capital 43.5k, Metaplanet 43.0k, MARA ~36k, Cantor Equity 30k) drop off steeply. The 2026 tape is double-edged: a record 110k BTC bought in Q2 2026 — corporates absorbed 2.1× miner issuance year-to-date — yet on July 6, 2026 Strategy sold 3,588 BTC ($216M), its first sale ever, and Standard Chartered's Geoffrey Kendrick now argues DAT buying is "seriously constrained" and unlikely to provide further price support. We believe the DAT channel matures from growth engine to cyclical amplifier: premium-to-NAV financing (issuing stock above asset value to buy more BTC) only works in bull tape. We carry corporate penetration from today's 1.2% of a $7T cash pool to just 2% in the base case — and note ARK's bull assumption (10%) looks unreachable this cycle. A distinct sub-class sits inside this line: Bitcoin miners (MARA ~36k BTC, plus Riot, CleanSpark and peers) that produce and often retain — accumulators whose "cost basis" is electricity. But even they are conditional holders: MARA revised its treasury policy in 2026 to permit sales and sold ~15,000 BTC to buy back convertible debt, cutting roughly a third of its peak stack. So the miner cohort's monthly HODL-vs-sell decision is itself a swing factor in net supply — and in a drawdown it can flip from sink to source.

C. Pensions, endowments & family offices — the slowest, biggest pool

Direct pension positions are pioneer-stage and, so far, reversible — Wisconsin's $321M pilot (Q4 2024) was fully sold by May 2025; only a handful of funds have live positions — and the AIMA allocator data says pilots sit below 1–2% inside 2–5% opportunistic buckets. The historical template is exact: alternatives went from 13% to 20% of large-market pension portfolios over twenty years — roughly 35bp of portfolio share per year, once consultants blessed the class. Crypto has no consultant infrastructure yet; it is where private equity was circa 1990. Endowments tend to move first (the Yale model diffused top-down): Harvard and Brown became filed holders in 2025, and endowments face the toughest forward-return math in traditional assets — though Harvard's Q1 2026 ~43% trim is a reminder that early endowment positions are tactical, not yet strategic. Base case: pensions reach 0.3% by 2030 — that alone is $250B, the largest single line in our model.

Family offices are the exception that proves the rule — and the group the AIMA data ranks first among crypto allocators (35% of the allocator sub-sample, versus 15% for pensions and 3% for sovereigns). Managing an estimated ~$6T with no board, no ERISA fiduciary overhang, no consultant gate and a multi-decade horizon, family offices can and do take 1–5% positions today. They are the fast money inside the slow pool — the leading indicator for where the fiduciary channels go once the infrastructure catches up. We fold them into the pension/endowment pool for sizing, but note they behave more like the sovereign-Gulf model: early, direct, and conviction-driven.

D. Sovereign wealth funds & governments — small dollars, huge signal

Disclosed SWF positions are ~$1B (Abu Dhabi complex) — 0.01% of a $13.7T pool — but their behavior is the tell: Mubadala has added to IBIT in every quarter it has disclosed since Q4 2024, straight through a 49% drawdown (14.72M shares / $565.6M by Q1 2026). Governments hold ~600k BTC, essentially all seized; the US Strategic Bitcoin Reserve (EO, March 6, 2025) prohibits selling but authorizes no buying, remains un-codified 16 months on (agencies still deciding whether Treasury or Commerce runs it; the NDAA in late 2026 is the likeliest legislative vehicle). A distinct sovereign acquisition mechanism deserves its own mention: Bhutan mines Bitcoin with state hydropower — accumulating without buying or seizing — and El Salvador built the original programmatic-purchase model — though its buying is now constrained by an IMF program that caps new public-sector purchases, with recent reserve growth attributed to wallet consolidation rather than fresh buying. These are proofs-of-concept (and, in El Salvador's case, a cautionary tale about external constraints) for the "budget-neutral" state accumulation that a codified US reserve with purchase authority would scale. Data shows the sovereign channel today is an option, not a flow. We believe it is the bull-case kicker, not the base case: VanEck's long-run framework has central banks at 2.5% of reserves by 2050; we carry 0.3% by 2030 in base, 2% in bull — and note that even 2% of ex-gold reserves is $286B, the sort of number that would re-rate everything else.

E. Insurance — the untouched $40T

The insurance general account is the emptiest large pool: MassMutual's $100M purchase (December 2020) remains the canonical direct position, and Northwestern Mutual's CIO publicly declines. What changed in January 2026: Delaware Life launched the first fixed-index annuity tied to a Bitcoin-containing index (a BlackRock 12%-vol balanced index) — the first time BTC exposure entered an insurance product chassis rather than the surplus account. Risk-based-capital treatment keeps general-account BTC near zero; product wrappers are the realistic path. We carry 0.05% by 2030 base — small percentage, $24B pool-weighted.

F. Banks: digital credit & structured products — the plumbing channel

Banks will intermediate Bitcoin long before they hold it. The Basel cryptoasset standard, in force January 2026, assigns unhedged BTC a 1,250% risk weight and caps Group-2 exposure near 1% of Tier 1 capital — a de-facto prohibition on balance-sheet holdings (a Basel "expedited targeted review" opened November 2025 after industry and US Senate pushback). But the credit and product businesses are compounding regardless: crypto-collateralized lending reached $67B (+49% y/y); Ledn priced the first investment-grade-rated Bitcoin-collateralized ABS ($188M); Strike lends against a $2.1B Tether facility at 7.5%+; and the OCC (March 2025) reaffirmed national-bank custody powers. Meanwhile the CFTC's pilot admitting BTC as derivatives collateral and CME's 2026 tokenized-collateral rollout make Bitcoin usable inside regulated market plumbing. This channel adds little direct holding to the TAM but it is the infrastructure that de-risks every other channel — collateral status is what turns an asset institutions may own into one they need to hold inventory of.

G. Wealth platforms & retirement rails — the 2026–27 unlock

Two rails matter more than everything else combined, because they sit upstream of household and pension money:

1. Wirehouse solicitation. Morgan Stanley is moving Bitcoin ETFs from unsolicited to advisor-recommended; Bank of America/Merrill is reported to have approved advisor-recommended crypto at 1–4% of client portfolios from January 2026; UBS and Wells still gate. The four wirehouses steward $10T+. Advised money does not self-direct into a new asset class — the difference between "your client may ask" and "your advisor may recommend" is the difference between a trickle and an allocation.
2. The 401(k) executive order (August 7, 2025) directs DOL to clear alternatives — explicitly including actively-managed crypto vehicles — for defined-contribution menus; EBSA's proposed fiduciary rule cleared OMB and was released March 24, 2026 (comment period through June 1). DC is where pension growth lives (59% of large-market pension assets, compounding 7.3%/yr vs 3.1% for DB). A default-menu sleeve is how 0.3% pension penetration actually happens mechanically.

H. Captive & structural demand — the "need to hold" bid

The user of this note asked the right question: when do institutions need Bitcoin? Three mechanisms already exist. Index inclusion: MSTR entered the Nasdaq-100 (December 2024), pulling every QQQ tracker into indirect BTC exposure; S&P 500 inclusion was passed over in late 2025 despite eligibility. The ratchet was live-tested in early 2026: MSCI consulted on excluding DAT companies (JPMorgan had estimated ~$2.8B of forced outflows if it did) but decided against it in January 2026 and reaffirmed that in its February review — so DATs stayed in the indices and the passive channel held. Inclusion can cut both ways, but so far the committees have kept the door open. Collateral status — the deepest "must hold" mechanism. This one deserves more than a clause. An asset moves from may own to must hold inventory of the moment it becomes accepted collateral in the wholesale plumbing: tri-party repo, prime-brokerage margin, futures-clearing waterfalls, and tokenized-collateral rails. The 2025–26 groundwork is already laid — the CFTC's pilot admitting BTC (and ETH/USDC) as derivatives collateral (announced December 2025), CME's tokenized-collateral rollout with Google Cloud, and OCC-blessed bank custody. Once a dealer must post or accept BTC to service a client, it must carry a book of it — a demand that is structural, price-insensitive, and entirely separate from any allocation decision. This is the truest answer to "when do institutions need Bitcoin": not when they want the return, but when their counterparties settle in it. Benchmark drift: if crypto enters standard multi-asset benchmarks at even 1%, tracking-error-averse managers become forced buyers. None of these is large today; all are ratchets — they only turn one way.

I. Stablecoin issuers & the credit channel — the price-insensitive buyers

Two holder-classes sit outside every allocation survey because they don't allocate — they accumulate mechanically.

Stablecoin-issuer reserves. Tether holds ~97,000 BTC visible on-chain (~$7B) — and, per its CEO, over 140,000 BTC once custodied and OTC positions are counted — enough to rank as a top-five holder globally, second only to Strategy among company-like holders — and buys more structurally, committing up to 15% of realized operating profits to Bitcoin each quarter (it added ~8,900 BTC in Q4 2025 alone, funded by >$10B of 2025 net profit on a ~$185B stablecoin float). This is a bid with no benchmark, no drawdown tolerance test, and no redemption committee: as the stablecoin economy grows, the issuer's earnings — and its programmatic BTC purchases — grow with it. It is the closest thing in the market to a central-bank-style balance-sheet accumulator that already exists.

The credit channel — how fixed income accesses Bitcoin. The single largest pool of institutional capital, fixed income, is barred by mandate from holding BTC or even equity — but it can hold debt. So a whole instrument stack has grown to route bond capital into Bitcoin: Strategy alone has ~$8.2B of convertible notes outstanding (proceeds used to buy BTC) plus a ladder of BTC-backed preferreds (STRK/STRF/STRD), and the market now includes sovereign "Volcano bonds" and early BTC-collateralized bond structures. A pension or insurer whose policy forbids spot BTC can often buy an investment-grade-rated, BTC-linked note — accessing the exposure through the one door its mandate leaves open. This channel adds little new coin at the system level (the converts fund DAT purchases already counted in §4-B), but it matters enormously for who can participate: it is the on-ramp for the ~$40T+ of credit capital that the equity and spot channels cannot reach.

Channels considered but not TAM-additive

For transparency, three vectors that look like demand channels but do not add to an ownership TAM, and are excluded from the model: (1) the CME futures / cash-and-carry basis trade — an enormous institutional flow, but a hedge (long spot/ETF, short future) that is net-neutral on holdings; it is an access and price-discovery mechanism, not a holder; (2) ETF authorized participants and seed capital — market-making plumbing, not end-owners; and (3) crypto philanthropy / donor-advised funds (e.g. Fidelity Charitable accepting BTC gifts) — real, but immaterial to the aggregate.

§5 What History Says About the Pace

The gold-ETF precedent is the best-documented institutional adoption curve in existence — and it is a double-edged one. GLD (launched November 2004): $1B in three days, $5B in 15 months, $10B in three years, a ~$77B peak in year seven (2011). Bitcoin's ETF ramp beat the entry: US spot Bitcoin ETFs hit $147.5B in 21 months — GLD's peak-era level in under two years, at tighter fees (0.25% vs GLD's 0.40%). But GLD's post-peak path is the cautionary half the bull case must confront: after 2011, gold-backed ETFs shed ~$24B+ of outflows in 2013 alone (552 tonnes, with only 17 days of net inflows the entire year) and the category fell roughly a third from its peak, then spent close to a decade recovering — GLD did not durably reclaim its 2011 high for years. The read-across cuts both ways: on the compressed clock (GLD's twenty-year AUM reached in two), Bitcoin's round-trip from $147.5B (Sep 2025) to ~$77B now (−48%, tracking the price) may be the analogue of GLD's 2013 bleed, not a pause before the next leg. If the 2024 launch released a finite stock of pent-up demand — GBTC conversions, DAT premium-financing (which we call spent in §4-B) — then "the wave was the trade" is a live scenario, not a throwaway bear. The precedent supports both the ramp and the plateau; intellectual honesty requires showing the whole curve.

The second precedent sets the ceiling rather than the pace: alternatives inside pension portfolios went 13% → 20% over 2004–2024, and US public pensions ran ~35% in alternatives by 2022 (up from 27% in 2019), with large endowments above 50%. That climb required a generation of consultant infrastructure, benchmark inclusion, and staffing — and it happened for an asset class (private equity) with worse liquidity than Bitcoin and, per the academic literature, an allocation size driven almost entirely by perceived excess return. The gating factor for crypto is not liquidity or custody anymore; it is the perception question — and 74% of surveyed institutions expect crypto prices to rise over the next year even in this drawdown.

The third precedent is a warning about reflexivity: BlackRock's own sizing paper argues the 1–2% window exists precisely because adoption is incomplete — "broad adoption would itself dull further price upside." TAM analyses that project both mass adoption and early-adopter returns are double-counting. Ours does not: the implied-price grid below prices the stock, not a return narrative.

§6 The Model (Lens One): Three Penetration Worlds

This is the portfolio-allocation lens — Bitcoin as a satellite sleeve inside institutional portfolios, sized against the allocator anchors. It is the conservative, defensible base for a TAM. §7 then layers a second, larger lens on top: Bitcoin as a store of value competing for the world's monetary premium.

Method: for each end-owner pool, grow the pool at its trend rate to 2030, apply a steady-state penetration (set from the §4 channel work, disciplined by the allocator anchors — BlackRock 1–2%, VanEck 1–3%, ARK's 2.5% base institutional penetration, pension pilot reality of <1%), and sum. The output is a dollar stock of institutionally-held BTC in 2030, converted to implied price only through an explicit assumption about what share of supply institutions end up holding.

Reading the scenarios through precedent. The base case (+35%/yr) asks institutional Bitcoin to compound at roughly the pace pension alternatives grew dollar holdings through their 2000s adoption decade, and takes blended penetration to just ~0.4% — still below every published allocator anchor. The bull case (+80%/yr) is, almost exactly, GLD's first-seven-years AUM CAGR — it has a precedent, but the precedent required a decade-long gold bull market. The bear case is stagnation: the 2024–25 wave proves to have been the whole trade, DATs unwind at the margin, and the 401(k)/wirehouse rails stay shut.

From dollars to price — honestly. A dollar stock is not a market cap; converting one to the other requires assuming what fraction of Bitcoin's ~20.5M-coin 2030 supply institutions end up holding (they hold ~15% of wrappers' worth today; ~40% of supply is effectively vaulted per ARK's liveliness work, so the free float is far smaller):

Implied BTC price = scenario dollar stock ÷ (20.5M coins × share held). † Greyed cells are infeasible for adoption scenarios: institutions already hold ~11% of 2030 supply, so shares below that imply net selling (§8); shares above ~25% are unreachable — net purchases would exceed the entire available float. Shaded: base-case anchor (~$181k at 29% float absorption). Red: outcomes below the July 2026 spot price of ~$64k. Within each row, a tighter float means lower absorption and a HIGHER clearing price — see §8 for the supply-adjusted bands.

The base-case feasible band — institutions hold $556B at between their current ~11% and 15% of supply — lands at $181k–$248k, squarely bracketing Bernstein's $200k end-2027 target and sitting on the path to Standard Chartered's $500k-by-2030. The bull row's feasible cells span $491–898k — the full band, since §8 shows the float is elastic (long-term holders release coins into strength), so the sub-$500k cells are what an elastic float produces, not impossibilities. We take the rough convergence of a bottom-up penetration model with top-down sell-side targets as a sign the shape of the estimate is right, while stressing every input is an assumption, not a measurement. But note the ceiling this lens imposes: even the bull case tops out near ~$900k because a portfolio allocation is inherently bounded at a few percent. To get materially above that, Bitcoin has to stop being an allocation and start being money — which is the subject of §7.

§7 The Second Lens: From Risk Asset to Store of Value

Everything to this point priced Bitcoin as a satellite risk-asset allocation — the 1–2% sleeve BlackRock frames. That is the conservative lens, and it is the one the current market actually uses. But it may be pricing the wrong thing. If Bitcoin's monetary premium matures — if the market comes to hold it not as a high-beta bet on liquidity but as a durable, non-sovereign store of value, uncorrelated with equities and with a supply schedule no government can dilute — then the relevant denominator is no longer "a few percent of portfolios." It is a share of the world's monetary premium itself: the value humanity parks in gold, in government bonds, in trophy real estate, and in cash precisely because it wants a place to store wealth. Those pools dwarf the allocation pools. This section sizes that regime — and is explicitly more speculative than the rest of the note.

The condition the whole thesis rests on

Data shows Bitcoin today does not behave as a store of value — and, more damaging, the thesis just failed its own best test. The load-bearing condition (below) is that Bitcoin's correlation to risk assets falls toward zero. The falsifier we would set is clean: if, through a full debasement-regime stress test, Bitcoin keeps trading as a leveraged Nasdaq while the monetary premium flows to gold, the thesis is wrong. That test just ran, and the thesis lost it. 2025–26 delivered precisely the regime the store-of-value case needs — fiscal stress, debasement fear, gold reaching record highs (~$5,600/oz, up ~80% from early 2025 and passing US Treasuries as a share of official reserves). In that exact window, Bitcoin fell ~49% from its October 2025 high and stayed tightly correlated to equities, while gold absorbed the monetary-premium bid. The natural experiment the thesis was built for arrived, and the premium went to the metal. The one genuine mark in Bitcoin's favor is narrower than we previously claimed: its realized volatility has fallen to multi-year lows near 27%, and the BTC/gold volatility ratio has compressed to ~1.5× (JPMorgan) — but part of that is gold's volatility rising in its own bull market, not Bitcoin maturing, and volatility was never the load-bearing objection anyway. We believe the honest status is this: the store-of-value transition remains an unproven, single-observation option whose most recent and most favorable test cut against it. Everything below is conditional on a decorrelation that has not yet occurred in any cycle, including this one. Gold itself took decades and a specific macro regime — the end of Bretton Woods, then stagflation, then 2008, then the 2022+ debasement era — to cement its premium; Bitcoin's is, at best, still forming, and on current evidence the burden sits with the bull.

The monetary pools — a different order of magnitude

The store-of-value question is not "what % of portfolios" but "what share of the value the world stores." Four pools hold that value (2026 figures; sources in the verification note):

• Gold — ~$29T above-ground (~$13–15T of it investment and official-reserve gold). This is the directly substitutable pool: gold and Bitcoin serve the identical function — non-sovereign, non-yielding, hard-capped stores of value. The tell is already flashing, but pointing at gold: central banks bought over 1,000 tonnes a year in 2022, 2023 and 2024 (moderating to ~850t in 2025), and gold reached 27% of global official reserves at end-2025 — passing US Treasuries (22%) for the first time since 1996. A monetary-premium migration is unmistakably underway; the only question is whether any of it redirects to the digital version.
• Global bonds — ~$143T. The incumbent "risk-free store of value." Sovereign debt is where the world is forced to park savings when it wants safety and yield.
• Global real estate — ~$393T (Savills), the single largest store of wealth on Earth, and a large slice of it is held as a store of value — investment property, land-banking, empty Tier-1 units — not for use.
• Broad money — ~$99T of global M2, the cash-and-deposit layer itself.

Against these, Bitcoin's entire $1.27T market cap is a rounding error — ~4% of gold, under 1% of bonds, ~0.3% of real estate. That gap is the store-of-value TAM. It is not additive to the §6 allocation TAM; it is a different way of asking the same question, and a far larger one.

What disruption looks like, pool by pool

• Gold is the first domino, and the cleanest. Because the two assets are functional twins, "digital gold" is less a metaphor than a substitution schedule. If Bitcoin captures even half of gold's investment role, that alone is a ~$6–7T market cap (~$300–340k/BTC); parity with all above-ground gold is ~$1.4M. The US Strategic Bitcoin Reserve is the institutional template for exactly this migration — a sovereign choosing the digital reserve asset alongside the metal.
• Bonds are the deepest disruption. Government debt is the default savings vehicle only because there has been no credible non-sovereign alternative at scale. In a world of financial repression and negative real yields — the world our own Debt Endgame and Private Credit notes describe — a hard-capped asset that no treasury can inflate competes directly with sovereign bonds for the "store my savings" mandate. It need not replace them; if it captures even 1–3% of the $143T bond pool as a reserve alternative ($1.4–4.3T), the second-order effect is that it raises the marginal cost of sovereign funding — savers have somewhere else to go. That is the mechanism by which a Bitcoin store-of-value regime would actually disrupt the debt markets: not by crashing them, but by ending their monopoly on safe savings.
• Real estate is the largest and least-discussed prize. A meaningful fraction of the $393T is a monetary holding — wealth stored in bricks because bricks hold value better than depreciating cash, especially in high-inflation or capital-controlled economies. Bitcoin competes for precisely that portion: it is portable, divisible, borderless, non-confiscatable, and carries no maintenance, tax, or tenant. It does not threaten housing-as-shelter; it threatens housing-as-savings-account. Even a 0.5–2% rotation of the store-of-value premium in global real estate is $2–8T of demand — the biggest single number in this entire note.
• Money is the endgame. If Bitcoin becomes a unit in which institutions hold savings — a slice of the $99T broad-money layer — it disrupts the money-market and cash-management complex itself. This is "hyperbitcoinization-lite," and it is the least likely, but it is the logical terminus of the other three.

Bull / base / bear under the store-of-value thesis

These are regime scenarios, not dated forecasts — the monetization of an asset class plays out over a decade-plus, not to a 2030 line. They price a total market cap (contrast §6, which priced an institutional stock), converted at ~20.5M coins.

• Bear — regime shift fails · ~$1.4T · ~$71k/BTC. The correlation never breaks; volatility stays disqualifying; Bitcoin remains a high-beta risk asset and a minor "digital-gold curio" capturing ~5% of gold and nothing else. The monetary premium is a story, not a balance-sheet line — and the §6 portfolio lens is the whole truth. Note this sits below today's price: a failed monetization is not neutral, it is a de-rating.
• Base — digital gold realized · ~$10.5T · ~$512k/BTC. Bitcoin matures into "gold 2.0": it captures roughly 30% of gold's monetary role plus a whisper of the bond, real-estate, and money pools (≤0.5% each). It decorrelates from equities specifically in debasement regimes — behaving as a hedge when it matters — without yet threatening bonds or property at scale. This is the scenario in which the second lens is real but bounded.
• Bull — broad monetization begins · ~$27T · ~$1.32M/BTC. Bitcoin is recognized as a non-sovereign reserve asset: ~70% of gold's monetary role, ~2% of the bond pool, ~0.5% of real estate's monetary premium, ~2% of broad money. Central banks hold it alongside gold; it is a standard treasury reserve line. This is where ARK's $1.5M 2030 bull lives — and it requires the correlation transition to be not just real but complete. (Note this sits above §8's supply-adjusted allocation bull of ~$490–900k, and deliberately so: that number prices institutions accumulating coins against an elastic float, while this one prices the entire supply re-rating as money — the rung above it on the ladder laid out at the end of §8.)

Beyond the bull: what full monetization actually computes to

**The bull case is not the ceiling. The floor of the terminal calculation is full gold parity: $29T over 20.5M coins ≈ $1.4M per coin. But stopping there — or adding token slivers of the other pools — quietly assumes the marginal buyer disappears the moment gold is conquered, which is backwards: the non-gold pools (bonds $143T, real estate $393T, broad money $99T — together $635T) are twenty-two times gold's size. The right way to size them is not raw bulk (most of a bond is yield backing a liability; most of a house is shelter) but the store-of-value premium buried inside them — the portion held purely to warehouse wealth. Our illustrative decomposition: 20–30% of bonds, 10–20% of real estate, 30–50% of broad money ≈ $100–170T of competable monetary premium beyond gold — three to six times the entire gold market. Adding gold, the total prize is ~$163T today.** What follows sizes how much of it Bitcoin takes, and how fast.

The benchmark argument — VanEck's — and why it can be beaten

The most rigorous published long-horizon framework is VanEck's: $2.9M per coin by 2050 (base), built on Bitcoin settling 5–10% of international trade and reaching 2.5% of central-bank reserves, with a hyper-monetization tail at ~$53M. Credit where due — it is explicit, falsifiable, and time-stamped. But its load-bearing channel is the hardest one: trade settlement requires coordination — exporters, importers, and their banks must all move, against the full network-effect gravity of the invoicing dollar, and reserve adoption requires politics. There is a cleaner channel, and it is the one actually observable in the data since 2024: savings. A saver — a household, a fund, a treasurer — needs no one's permission and no counterparty migration to store wealth in Bitcoin; adoption is unilateral, and the ETF/DAT/13F record of §2–§4 is precisely this channel operating. An argument built on savings should therefore run ahead of one built on settlement. Three questions decide how far ahead.

Question one: how fast does the prize itself grow?

The static tiers above freeze the pools at 2026 size — a large understatement, because store-of-value pools compound at roughly the debasement rate. Global M2 has grown ~6–7% a year across cycles (the note's own global-liquidity work); global public-debt ratios are still rising (bond stock ≈ nominal GDP+); property tracks nominal GDP; even gold's above-ground stock dilutes ~1.7%/yr from mining while its dollar value compounds with the price of everything. We assume the blended premium pool grows 5.5%/yr — deliberately below trend M2. The consequence is the single most under-appreciated fact in the terminal math: the ~$163T prize doubles by ~2039 and reaches ~$590T by 2050. The same force that drives savers out of fiat — debasement — grows the pool Bitcoin is capturing. The TAM is a treadmill moving toward the thesis.

Question two: why does Bitcoin take share — pool by pool?

Substitution needs a superiority argument per incumbent, for the specific job "store wealth for decades against sovereign debasement":

• vs gold: equal scarcity credibility with a hardening edge — gold's stock dilutes ~1.7%/yr forever; Bitcoin's issuance halves toward zero (by 2050, gold's stock will have grown ~50%, Bitcoin's ~4–5%). Add superior portability, divisibility, auditability, and custody cost. Gold's remaining moats — 5,000 years of Lindy, central-bank incumbency, no key risk — are real but eroding (the vol gap is at a record-low 1.5×; a US reserve already holds BTC alongside bullion).
• vs bonds: the incumbent is failing at the job. In a financial-repression regime — the world of our Debt Endgame note — sovereign bonds deliver negative real yields by design; the "risk-free store" quietly taxes its holders. Bitcoin carries no issuer, no duration, and no inflation pass-through. The bond pool doesn't need to collapse; it needs only to keep repressing.
• vs real estate: property-as-savings pays 1–2%/yr in taxes, maintenance and friction, cannot move, and sits exposed to local politics and capital controls — the exact vulnerabilities BTC inverts. The bid competes for housing-as-savings-account, not housing-as-shelter, and it is strongest where property premiums are most monetary (EM, capital-control economies).
• vs cash: pure debasement exposure; stablecoins are already migrating cash onto crypto rails, one custody decision away from the harder asset.

Question three: how deep, how fast? The S-curve, anchored to measured history

Monetary substitutions follow logistic curves over decades, not linear drifts. We model Bitcoin's share of the (growing) premium pool as a logistic: p(t) = K / (1 + ((K−p₀)/p₀)·e^(−rt)), with every parameter argued:

• p₀ = 0.78% — measured, not assumed: today's $1.27T market cap over the $163T pool.
• r ≈ 0.20–0.30 — anchored to measured history: Bitcoin's share of the pool rose from roughly 0.15% (estimated, 2020) to 0.78% (measured, 2026) — ~27%/yr early-phase; we haircut it, since S-curves decelerate as they scale.
• K (terminal share) — the depth question: shallow 5% (a permanent niche, about a quarter of gold's current ~18% share of the pool — the "digital curiosity" endgame); base 25% (Bitcoin ends meaningfully above gold's current ~18% share — the claim being that a superior instrument doing the same job ultimately out-earns the incumbent's premium); deep 50% (Bitcoin becomes the primary non-sovereign store of value — the full §7 thesis).

Run the machine — logistic share × compounding pool ÷ halving-stepped supply — and the paths are:

Why we believe this argument beats the benchmark

What the model is — and, honestly, is not. The dynamic path's 2030 waypoint (~$198k) is robust, because at four years the logistic has barely moved off its measured starting point (all three K scenarios land near $180–250k), so it agrees with the §6/§8 bottom-up build almost mechanically. That agreement is reassuring for 2030 but carries little weight for 2050 — the far end is driven entirely by the two parameters the data cannot pin down (the terminal share K and rate r), not by the cross-check. We flag this plainly rather than sell "two independent methods, one answer" as validation it is not.

Three things an adversarial reading gets right, and we concede:

• The 2050 figure is mostly denominator debasement, not share-winning. Freeze the pool at today's $163T and the base 2050 price is ~$1.8M, not $6.5M — so ~72% of the headline number is simply the measuring-stick (the debasing dollar) shrinking, exactly the force the thesis relies on. In today's purchasing power the base path is roughly $580k (2035), $1.4M (2040), and $3.4M (2050) at ~2.75% CPI — or about half those figures deflated at the pool's own 5.5% rate. Read the nominal numbers as what they are: 24-year compounds in debased dollars.
• The competable pool is smaller than the gross premium. Much of the $100–170T "premium" is not actually contestable by Bitcoin — sovereign bonds are held for LCR/HQLA, collateral, and liability-matching that a 1,250%-risk-weight asset cannot satisfy (§4-F); investment real estate is largely levered and non-transferable. Haircut the pool to a defensible ~$50–70T and the base 2050 falls to ~$2.4M nominal (~$1M real) — below VanEck, not above it. Our "beats VanEck" claim survives only on the un-haircut pool; on the conservative pool it does not.
• K = 25% is a choice, not a measurement. No voluntary monetary substitution in history took a new store of value above gold's ~18% pool share in 25 years (sterling→dollar took 30–50 years and happened via sterling's collapse; silver and gold demonetizations were statutory). We keep K = 25% as the labeled base scenario but concede it is an assumption with no direct precedent, not a forecast.

So the defensible statement is narrower than the headline. On like-for-like nominal terms our base 2050 (~$6.5M) is ~2.2× VanEck's $2.9M, and we'd argue the savings channel is cleaner than VanEck's trade-settlement mechanism (unilateral, and measurably underway in §2–§4). But once the pool is conservatively haircut, the honest range straddles VanEck rather than clearing it — Bitcoin's 2050 price is on the order of $2–6.5M nominal / ~$1–3.4M in 2026 purchasing power, comparable to VanEck, not a clean beat. What we hold to is the shape: the 2030 waypoint is solid, the savings mechanism is sounder than settlement, and even the shallow path reaches seven figures by 2050 on denominator growth alone. The precision past ~2035 is scenario sketch, not analysis — and the VanEck $53M hyper-tail one rung up is incoherent as a construct (it double-counts the same coins as both settlement medium and dormant savings, and $53M × 21M ≈ $1.1 quadrillion exceeds all world wealth), which is why we quote it only as a boundary marker, not a scenario.

And the honesty notes, at altitude. These are parameter choices, argued but not provable — the table is built so a reader can substitute their own K and r. The history anchoring r is one asset's single adoption run (n = 1). This is a transfer, not creation: every dollar arriving is monetary premium leaving bonds, property, and cash — the disruption thesis, priced. And at $4–13M per coin Bitcoin is a $90–280T asset; dollar targets bend under their own weight at that scale, and the meaningful claim becomes Bitcoin's share of world wealth — which is why the table's last column is the one that matters.

Reference points, single-pool. Because the blend above is a judgment, the honest cross-check is per pool: Bitcoin at 50% of investment-grade gold ≈ $350k; at 1% of the bond market ≈ $70k; at 1% of real estate's total value ≈ $190k; at 2% of broad money ≈ $97k. That any single one of these fully-monetized fractions lands in the mid-six figures is the whole point — the store-of-value lens and the portfolio lens converge on the same conclusion from opposite directions: **the current ~$64k price embeds almost none of the monetary premium, and a large fraction of the upside case is the market simply changing its mind about what Bitcoin is.**

§8 The Other Blade: A Tight but Elastic Float Meets the Bid

Every section to this point has sized demand. But a TAM for a fixed-supply asset is only half an argument — the other half is what happens when that demand meets a supply that cannot expand, and whose available portion is actively shrinking. This is the mechanism that turns the demand dollars of §6 and §7 into price. It is also the single most important reason the numbers in this note are expressed as implied prices rather than absorbed dollars: with Bitcoin, demand cannot be met with more coins, only with higher ones.

Supply is fixed, and issuance is halving

Bitcoin's supply schedule is not a forecast; it is code. Of the 21M maximum, ~20.05M is already mined, and new issuance — since the April 2024 halving — runs at just 3.125 BTC per block, ≈450 a day, ≈164,000 a year (worth ~$10.5B at today's price). At the April 2028 halving it drops to ~82,000 a year. Across this note's entire mid-2026-to-2030 horizon, the network will mint only about half a million new coins — total. That is the whole of new supply available to satisfy every buyer in every channel of §4, combined.

The available float is far smaller than the supply

The ~20M figure badly overstates what can actually be bought. Strip it down (2026 data; illustrative decomposition):

• ~3.1M lost or dormant — provably unspendable (early wallets, lost keys, Satoshi's untouched ~1.1M). Gone.
• ~11.4M long-term illiquid — self-custodied by holders whose coins rarely move. Taken together with lost coins, Glassnode classifies ~14.4M (roughly 72–74% of supply) as "illiquid," a share that has held in the low-to-mid 70s and near record highs.
• ~3.1M institutionally locked — the ETF (~1.21M), corporate-treasury (~1.27M) and government (~0.6M) coins from §2, which leave the tradable market when they enter these wrappers.
• ~2.4M genuinely liquid — and this is the number that matters. Exchange reserves have fallen to ~2.2–2.4M BTC, a 7-to-9-year low, down from ~3.2M in 2023. Roughly 12% of mined supply is actually available; the rest is spoken for.

The flow imbalance: demand runs ~9× new supply

Now put the two blades together. New issuance is ~164,000 BTC a year. The §6 base case implies ~$92B of net new institutional demand per year — which, at today's price, is ~1.44M BTC a year, or roughly 9× everything the network mints. The bull case is ~40×.

New coins cannot come close to filling that bid. The difference — 1.3 to 6+ million coins a year of wanted Bitcoin that the network will not create — can only be sourced from existing holders' ~2.4M-coin liquid float. And here is the crux: at today's price, base-case demand alone would consume the entire liquid float in under two years; the bull case would exhaust it in months. That cannot happen at a constant price. Long-term holders do not sell into a bid at $64k when demand is running ten times issuance — they sell higher, or not at all. The only variable that clears the market is price.

Why this makes price the release valve — and validates the grids

This is the answer to the reflexivity objection raised in §10, made concrete — and it changes the note's numbers, so we work it through carefully in stock terms (counting coins, not dollars, since most of a scenario's dollar growth is appreciation of coins already held, not fresh purchases). Institutions hold ~2.24M coins today — about 10.9% of the 20.5M coins that will exist in 2030. The most they can add is the absorbable float: ~2.9M coins. So their 2030 share of supply is hard-bounded between ~11% and ~25% — and each destination inside that corridor has a price:

• Reaching 15% of supply requires prying loose ~0.83M coins — 29% of everything available (float plus every newly mined coin for 4.5 years).
• Reaching 20% requires ~1.86M coins — 64% of everything available.
• Reaching the 25% ceiling means absorbing 100% of it: every liquid coin, every mined coin, nothing left over for anyone else.

Because supply is inelastic, a demand shock cannot be absorbed in quantity — it is absorbed in price, which rises until enough long-term holders are tempted to part with coins. The mechanism cuts one way within each scenario: the less the float yields, the fewer coins the same dollars can hold, and the higher the price must be. That is precisely why §6 and §7 are expressed as implied-price grids — and why a fixed-supply asset re-rates in discrete, violent steps rather than smoothly.

So does the supply constraint change the answer? It sets the transmission, not the direction — and an earlier draft got the sign wrong

Applying the corridor to the §6 scenarios re-prices them (dollar stock ÷ feasible coin holdings):

• Base ($556B stock): $108k–$248k, from full-float absorption up to a float that never lets go. The middle of the road — ~29% absorption — is ~$181k. The constraint's main effect is to invalidate the naive grid's sub-11% cells (institutions cannot hold less than today's ~11% in an adoption scenario), leaving the honest base band at ~$180–250k.
• Bull ($2.01T stock): $390k–$898k. The naive grid's ~$490k bull cell requires institutions to hold ~20% of supply — capturing ~64% of today's available float. Whether that is reachable is the whole question, and it turns on a fact an earlier version of this section got backwards.

The correction — the float is elastic, not fixed. A prior draft argued the tight float pushes the bull up, toward $650–900k, on the reasoning that illiquid coins "won't surrender." That is wrong, and this note's own §5 is the refutation: long-term holders distribute precisely into strength. On-chain, LTH supply fell ~1.8M coins in the 2021 rally and holders released roughly 2M more into the post-ETF 2024–25 advance — each cycle surrendering more coins than the entire "liquid float" this section counts, and doing so because price rose. The 2.9M "absorbable" figure is therefore a snapshot of behavior at ~$64k, not a ceiling at $200k. As price climbs, dormant supply liquefies, the absorbable float grows, and the same $2.01T bull dollar-stock buys more coins at a lower implied price — which pushes the bull back down into the ~$490–650k cells, not up. The Germany precedent makes the elasticity concrete: the state sold ~49,858 BTC in three weeks in mid-2024 (~5× daily issuance), and the market absorbed it with a transient ~15% dip and full recovery within days. A float that clears a forced sovereign sale that fast is not the inelastic wall the earlier draft assumed.

So the honest supply-adjusted bull is the full ~$490–900k band, not a $650–900k concentration — the low-mid cells are what an elastic float produces, not impossibilities. What the tight float genuinely does is set the transmission: because demand cannot be met in quantity at today's price, it is met in price, and a fixed-issuance asset re-rates in violent steps — in both directions. That symmetry is the real lesson: the same thin float that would amplify a demand surge amplified the downside in June 2026, when a comparatively modest $4.4B, 13-session ETF redemption streak coincided with the drawdown's final leg. A tight float is a volatility multiplier and a price-transmission mechanism — it is not, by itself, a directional bull driver.

Two transparency notes on this section's own history: (i) v1.1 replaced a flow-based absorption estimate (which double-counted appreciation of coins already held) with the stock-consistent method above — a genuine improvement; but (ii) v1.1 also over-claimed that the constraint "raises the top," which an adversarial review correctly flagged as the sign error corrected here. The supply blade is real; it cuts both ways.

Reading the two lenses together: the full price ladder

A careful reader will notice the note carries two different "bull" prices — ~$490–900k here, $1.32M in §7 — and ask which is right. They are not the same scenario, and the distinction is the note's whole architecture. The §6/§8 numbers price an accumulation trade: institutions build a position while the market still treats Bitcoin as an allocation, so the constraint is how many coins the (elastic) float surrenders and at what price. The §7 numbers price a regime change: the market re-rates what every one of the 20.5M coins is — money rather than position — so the price is simply the monetary market cap divided by the full supply, no accumulation required. Stacked in order, the scenarios form a single ladder:

Each rung requires everything on the rung below it, plus one more thing going right. The supply constraint governs the middle of the ladder — the accumulation phase — and it is precisely success in that phase (a float that will not sell to trillion-dollar demand) that would force the market to the top rungs, where the question is no longer "how many coins can institutions buy" but "what is a coin worth."

There is a documented amplifier here, too. Because so little supply is liquid, net inflows have historically moved Bitcoin's market cap by a multiple of the dollars that actually flowed in — the fund's own ETF flow-elasticity work, and multiple sell-side and academic studies (the BIS's "crypto multiplier" literature, a widely-cited Bank of America estimate near 100×), put that multiplier well above 1×, with published figures ranging from low single digits into the triple digits. The number is genuinely contested and highly methodology-dependent — which is exactly why we deliberately do not bake a multiplier into the §6/§7 numbers — those price the demand dollars 1:1 into market cap, which is the conservative choice. If the multiplier is real and greater than one, the price outcomes are higher than the grids show. The supply side is thus an upside asymmetry we have chosen not to claim.

The honest counter

Two caveats keep this from being a pure-bull argument. First, the exchange-reserve metric is a fuzzier supply-squeeze signal than it used to be: a large share of the decline reflects coins moving to ETF custodians rather than into permanent cold storage, and — as CoinDesk noted in July 2026 — those custodied coins are not truly locked, since ETF shares can be sold and redeemed, returning coins to the float. Second, **supply tightness amplifies moves in both directions.** The same thin float that turns demand into explosive upside turns net redemptions into disorderly downside — which is exactly what the record ETF outflows of late 2025 and the 49% drawdown demonstrated. An inelastic float is not a one-way ticket up; it is a volatility multiplier. But it does mean the demand this note has sized, if it materializes, has nowhere near enough coins to buy at anything like today's price.

§9 The Clock Inside the Trade: The Halving Cycle and Where We Are in It

§8 established that supply is the scissor's other blade. The halving is that blade's timer — the only major market event in finance that is scheduled in code, decades in advance. It has coincided with every Bitcoin cycle so far, and its next scheduled cut lands in the exact middle of this note's forecast horizon. This section lays out the pattern, where July 2026 sits inside it, and — just as important — why the mechanism behind the pattern is changing.

The mechanism, on a fixed schedule

Every 210,000 blocks — almost exactly four years — the new-coin reward paid to miners is cut in half. This is not policy; it cannot be lobbied, delayed, or eased. Four halvings have occurred: November 2012 (50 → 25 BTC/block), July 2016 (25 → 12.5), May 2020 (12.5 → 6.25), and April 2024 (6.25 → 3.125). The fifth arrives around April 2028 (3.125 → 1.5625), cutting new issuance from ~164,000 to ~82,000 coins a year — inside this note's 2030 model window.

The pattern: four cycles, one clock

Around every halving, the same sequence has played out: a price peak roughly 12–18 months after the halving, then a deep drawdown bottoming roughly a year after the peak, then accumulation into the next halving. The regularity is striking:

• 2012 cycle: peak ~12 months after the halving (Nov 2013, ~$1,150); then ~−84% to the Jan 2015 low.
• 2016 cycle: peak ~17 months after (Dec 2017, ~$19,700); then ~−84% to Dec 2018.
• 2020 cycle: peak ~18 months after (Nov 2021, ~$69,000); then ~−77% to Nov 2022.
• 2024 cycle: peak 534 days — 17.5 months — after the halving (October 6, 2025, ~$126,000): almost to the week where the prior two cycles peaked. The current −49% drawdown began exactly on the historical schedule.

Two features of the pattern matter as much as the timing. First, amplitude is damping: halving-to-peak multiples have fallen ~95× → ~30× → ~8× → ~2×, and peak-to-trough drawdowns have shallowed −84% → −84% → −77% → (so far) −49%. Each cycle's multiple has run roughly a third to a quarter of the last. Second, the damping is exactly what the §7 store-of-value thesis predicts: an asset maturing from speculative micro-cap to macro asset should see its cycles compress toward gold-like behavior. The boom is shrinking — but so is the bust.

Where July 2026 sits on the clock

Mapping today onto the template: we are ~27 months after the April 2024 halving and ~9 months past a peak that arrived precisely on schedule, at −49%. In the prior three cycles the post-peak bear lasted 12–13.5 months, which would put the template bottom in roughly Q4 2026 — and, notably, the on-chain floor evidence in our companion Is the Bottom Forming? note (record long-term-holder absorption into the June 2026 low) is consistent with that window. The template then points to accumulation through 2027, the April 2028 halving, and — if the 12–18-month lag holds a fifth time — a next cycle peak window of roughly April to October 2029. Read against this note's model: the projected crest of the next halving cycle falls squarely inside our §6 forecast horizon, with the end-2030 measurement point sitting just past it — meaning the scenarios span one full cycle, drawdown, halving, and projected peak. The base case does not require believing in the cycle — but if the clock keeps time, the TAM build-out and the cycle crest arrive within the same window.

Why this cycle may be the last one the halving causes

Here is the honest mechanical arithmetic, and it cuts against the folklore: the 2028 halving will remove ~82,000 coins a year of new supply — about $5B a year at today's price. Our base-case institutional demand runs ~$92B a year; actual 2024–25 ETF flows alone dwarfed the 2024 halving's supply reduction. The halving's direct supply effect is now a rounding error against institutional flows — each halving halves an ever-smaller number. If the four-year rhythm persists from here, it will be less because of the supply delta itself and more because the halving has become a shared narrative clock — a Schelling point that coordinates attention, leverage, and risk appetite (and, per §8, meets a float so thin that coordinated demand moves price violently in either direction). This is the emerging sell-side consensus, too: Bernstein calls the current environment a "boring" cycle with a healthier market structure — and means it as a compliment to the store-of-value thesis — while Standard Chartered's push of its $500k target from 2028 to 2030 is, implicitly, a bet that cycle timing is flattening into a secular adoption trend. We believe the institutionalization documented in §4–§6 is the thing that kills the classical cycle — allocation-driven, rebalancing flows do not follow a four-year clock — but that the 2024–2030 window is the handoff period in which both forces are live: one more halving, one more template window, layered on top of a demand base that no longer needs it.

What would falsify the cycle read: a bottom materially later than the Q4 2026 template window, or a next peak that fails to arrive by end-2030 despite the 2028 halving, would mean the clock has broken — in which case the §6 penetration math, which never assumed a cycle, still stands on its own.

§10 The Bear Case, Steelmanned

This note was put through four independent adversarial reviews, each tasked with destroying one of its pillars. They landed real hits — including one outright sign error (§8), now corrected. Rather than bury the survivors, we state the four strongest at full force, then give the honest scorecard. A research note that cannot survive its own steelman should not be trusted; here is ours.

Attack 1 — "The base case launches from a base that is currently shrinking." The note's own preferred instrument is measured stocks, not surveys. That instrument has turned: professional 13F Bitcoin-ETF holdings fell from 313k to 261k BTC in Q1 2026 (−17% in coins, −35% in dollars), hedge funds −39% and brokerages −53%, after net professional selling in Q4 2025. A +35%/yr compounding path starting from a base the freshest data shows contracting is a forecast at war with its own evidence; the AIMA "71% plan to increase" cohort had two quarters to act and the aggregate coin count went down. Our response — largely CONCEDE. We have rewritten §2 to lead with the reversal, not the trailing FY2025 growth figure. The genuine defense is compositional, not directional: banks doubled their holdings and Abu Dhabi's sovereign funds kept adding while the fast money left — a healthier holder mix, but a smaller aggregate. The base case now rests explicitly on a resumption of the 2024–25 accumulation regime, which the current tape has not delivered. The honest read is that the near-term risk is skewed to the bear, and we say so in §0.

Attack 2 — "The store-of-value thesis just failed its best test." §7's entire second lens is conditional on Bitcoin decorrelating from risk assets. 2025–26 delivered the ideal test: a record gold rally (~$5,600/oz, +~80%, passing US Treasuries in official reserves) driven by exactly the fiscal-stress/debasement fear the thesis invokes — and in that window Bitcoin fell ~49% and stayed correlated to equities. When sovereigns fled Treasuries they bought gold, unanimously, with the Bitcoin option fully available; the US "Strategic Reserve" holds only seized coins and has bought none. Our response — CONCEDE and demote. We have stripped the "decorrelation is already resolving" language from §7. The thesis is now labeled what it is: an unproven, single-observation option whose most recent and most favorable test cut against it. A worse possibility we cannot yet exclude: institutionalization may prevent decorrelation permanently — the more Bitcoin sits inside 60/40 books, the more it trades with them in risk-off. If that is the structural truth, §7 never activates and the §6 allocation lens is the whole story.

Attack 3 — "The dynamic terminal is mostly debasement, on a pool that is mostly non-competable." Two-thirds-plus of the headline $6.5M/2050 is simply the measuring stick shrinking at the 5.5% debasement rate — in today's purchasing power the base path is ~$3.4M, not $6.5M. And the $163T pool rests on a $100–170T "competable premium" we admit is unsourced; much of it is legally non-competable (sovereign bonds held for LCR/HQLA and collateral that a 1,250%-risk-weight asset cannot satisfy; levered, immovable real estate). Haircut the pool to a defensible ~$50–70T and the base 2050 falls below VanEck, not 2.2× above it — so the note's "beats the benchmark" boast is downstream of its most generous assumption. Our response — CONCEDE the framing, defend the shape. §7 now shows real-dollar figures alongside nominal, presents the pool haircut, and restates the VanEck comparison as "straddles, not clears." What survives is the 2030 waypoint (~$200k, nearly independent of the disputed parameters) and the mechanism argument (savings adoption is unilateral; settlement is not). Everything past ~2035 is scenario sketch, and the S-curve's terminal share K = 25% is a labeled assumption with no monetary-substitution precedent, not a measurement.

Attack 4 — "The GLD precedent is symmetric, and the wave may have been the whole trade." §5 sold GLD as a smooth $77B→$74B success. The real path: gold ETFs shed ~$24B+ in 2013 alone (only 17 net-inflow days that year), fell ~a third from the 2011 peak, and took years to recover. On the note's own compressed clock, Bitcoin's $147.5B→$77B round-trip may be the analogue of GLD's 2013 bleed, not a pause — i.e., the 2024 ETF launch may have released a finite stock of pent-up demand (GBTC conversions, DAT premium financing, both now spent) rather than opening a durable flow. Our response — CONCEDE and correct. §5 now shows GLD's full path, and "the wave was the trade" is elevated from throwaway to a live bear scenario with real weight.

The scorecard. Three pillars survive modified and humbler; one claim (§8's "tight float raises the top") was simply wrong and is fixed. What remains genuinely robust is the §6 2030 allocation lens (~$180–250k) — it never depended on decorrelation, cycle timing, or the disputed pool, and even the bear attacks leave it standing (a more elastic float makes the base more feasible, not less). Everything beyond 2030 is a conditional scenario whose conditions have not yet held, and the near-term balance of evidence favors the bear. That is the honest shape of the thing: a solid floor under a wide, unproven, upward-skewed tail.

§11 Further Objections, Briefly

"The DAT bid is exhausted." (Standard Chartered's Kendrick, paraphrased.) Partly conceded: Strategy's first-ever sale (July 6, 2026) and the collapse of premium-to-NAV financing mean the corporate channel cannot repeat 2024–25. Our base case already assumes it nearly stalls (1.2% → 2% in four and a half years). The bear risk is reversal — 1.27M BTC of corporate holdings becoming supply in a forced-deleveraging tape. Two-thirds of that risk is one company.

"Banks are locked out, so the infrastructure story is capped." True on balance-sheet holdings (1,250% risk weight); wrong on intermediation — lending, custody, collateral, and structured product are all compounding under rules that already exist. Basel's own expedited review is the free option.

"Index inclusion cuts the other way." It can — the JPMorgan estimate that an MSCI exclusion of DATs would force ~$2.8B out was a real scenario. But when MSCI actually consulted on it in early 2026, it chose not to exclude them (January decision, reaffirmed February), so passive plumbing stayed a one-way ratchet this cycle. It works only so long as classification committees cooperate — which is exactly why we keep it on the signpost list rather than banking it.

"BlackRock says adoption kills the return." We agree — and the model respects it. The grid prices a stock of holdings, not a momentum story; if the bull case arrives, forward returns from 2030 should compress toward gold-like, exactly as BlackRock argues. The TAM thesis is a one-time repricing thesis, not a perpetual-growth thesis.

"Surveys overstate everything." Also agreed — every major survey (Coinbase/EY, AIMA) pre-screens for crypto-engaged respondents, which is why our model is anchored to measured stocks (13Fs, ETF assets, treasury trackers) and uses surveys only for direction, never level.

And the deepest objection — reflexivity. $414B of base-case net demand cannot buy 6M coins at today's price; inelastic supply means the price adjusts until sellers appear. That is why the implied-price grid exists: the model's dollars and the price column are jointly determined, and the grid simply makes the simultaneity explicit rather than pretending to solve it.

§12 Signposts (each dated, each checkable)

1. DOL final rule on 401(k) alternatives: the proposed rule cleared OMB review and was released March 24, 2026, with public comment through June 1, 2026 — a clean, finalized fiduciary safe harbor for managed crypto sleeves is now the single largest base-case dependency, and the finalization is the event to watch.
2. Wirehouse solicitation dominoes: Morgan Stanley advisor-recommendation go-live; confirmation and terms of Merrill's reported 4% cap; UBS/Wells following. Watch platform AUM disclosures, not press releases.
3. US Strategic Bitcoin Reserve codification — NDAA season, late 2026. Codification without purchase authority = signal; budget-neutral acquisition authority = regime change (moves us toward bull-case central-bank penetration).
4. Basel targeted review outcome (opened November 2025): any risk-weight relief for hedged/custodied BTC reopens the bank balance-sheet channel we currently carry at zero.
5. MSCI/S&P classification decisions on DATs: MSCI declined to exclude DATs in early 2026 (the ratchet held), but the question can reopen at any review — a renewed exclusion move (~$2.8B JPMorgan-estimated forced flow) would be a bear-case marker for the corporate channel.
6. 13F institutional share of ETF assets (now ~28%): sustained climb through 35% would confirm the end-owner mix shifting from households to institutions; a stall below 30% weakens the whole §6 base case.
7. **A first corporate-pension or public-pension policy allocation** (not a pilot): the consultant-blessing moment that started the alternatives S-curve in the 1990s.
8. Exchange reserves and Glassnode illiquid supply (the supply gauge from §8): reserves are ~2.2–2.4M and falling, illiquid supply ~78% and rising. A continued decline through the next demand wave is the supply-shock thesis playing out; a sharp reversal (coins flooding back to exchanges) is the early-warning that demand has turned and the thin float is about to cut the other way.

§13 What the Analysis Implies — a Lens, Not a Recommendation

This note sizes a market; it does not advise anyone to buy it. Three understandings fall out of the work. First, the institutional Bitcoin trade is not "institutions are coming" — they are here, at 0.10%, and the entire question is whether the next stop is 0.5%, 1%, or nowhere. The measured stock (13Fs, ETF assets, treasury trackers) — not surveys, not conference sentiment — is where the answer will show up first. Second, the binding constraints are now administrative, not technological: a DOL rule, a wirehouse compliance memo, a Basel risk weight, an index committee. These are checkable, dated events — which is why §12 exists. Third, the drawdown is the test that makes the data meaningful, and it is currently splitting the field: patient balance sheets (banks doubling, Abu Dhabi's sovereign funds adding) kept buying through the 49% decline, while the fast money (hedge funds −39%, brokerages −53% in Q1 2026) capitulated — so the aggregate professional coin count is falling even as its composition improves. Whether that resolves into the base case or the bear turns on which behavior dominates the next leg; on the freshest data the bear has the edge, and the note is written to let a reader watch that resolve rather than to tell them how it ends. The aim throughout is comprehension.

§14 Disclosures

Information only. This material is published by TON618 Capital for information purposes only. Nothing here is an offer to sell or a solicitation of an offer to buy any security, fund interest, or digital asset, and nothing here is personalized investment advice or a recommendation regarding any instrument.

Publisher's exclusion. All research is published solely as general, impersonal information of regular circulation. It is not tailored to the objectives or circumstances of any individual and is not issued in connection with compensation from any client. The Fund has no clients and distributes all research free of charge. On that basis it publishes in reliance on the publisher's exclusion from the definition of "investment adviser" under the Investment Advisers Act of 1940 (§202(a)(11)(D); cf. Lowe v. SEC, 472 U.S. 181 (1985)).

Registration & conflicts. TON618 Capital is not registered as an investment adviser or broker-dealer in any capacity. The Fund is a Bitcoin fund and maintains a strategic Bitcoin allocation; this note's subject — the growth of institutional Bitcoin demand — bears directly on the value of that allocation, and the reader should weigh that as a material interest. As of the date of this note the Fund holds no position in any single security named herein (including MSTR, MARA, Metaplanet, or any ETF sponsor's equity); Bitcoin ETF products discussed are vehicles the Fund could use for its Bitcoin exposure. The Fund receives no compensation from any party in connection with this research.

Use of AI. Artificial intelligence is used in the creation of this research. All methodology and data integrity are reviewed and approved before publication by TON618 Capital's Chief Investment Officer, Keyth Beck; errors may nonetheless occur, and readers should verify independently.

CFA. This report was prepared to align with CFA Institute analytical standards (methodology only). CFA® and Chartered Financial Analyst® are registered trademarks owned by CFA Institute. That reference describes the analytical framework applied; it does not imply the report was prepared, reviewed, or authored by a CFA charterholder, and the report is not issued, reviewed, endorsed, certified, or approved by — nor affiliated with — CFA Institute.

Risk & feedback. Past performance is not indicative of future results. Digital assets are volatile and may result in total loss of capital. Corrections and feedback are welcome — please direct them to CIO Keyth Beck at keyth@ton618capital.com.

Source & Verification Note

This note synthesizes a five-angle research sweep (measured holdings stocks; allocator surveys with source-provenance verification; sovereign/bank channels; sell-side TAM frameworks; denominator pools and adoption precedents) plus a dedicated second pass on channels the first sweep missed (retirement rails, wirehouse platforms, insurance, Basel bank treatment, index inclusion, collateral infrastructure), and was then run through an adversarial claim-verification pass before release. That pass confirmed the large majority of the note's discrete numeric and attribution claims against primary sources and corrected several fast-moving 2026 items that had moved on before publication: MSCI consulted on excluding digital-asset-treasury companies but decided against it in January 2026 and reaffirmed that in February (the note previously framed the JPMorgan ~$2.8B exclusion scenario as an open risk); the State of Wisconsin Investment Board fully sold its $321M IBIT pilot by May 2025 (previously cited as a live pension position); Harvard cut its IBIT stake ~43% and exited its ETH-ETF position in its Q1 2026 filing (previously understated); the pension denominator was updated to WTW's GPAS 2026 vintage ($68.3T P22, year-end 2025) from the older $58.5T/$63T figures, which lowers current penetration further; the US public-pension alternatives figure was corrected to ~35% (2022) from a stale 27%; the DOL 401(k) rule is past OMB and public comment (released March 24, 2026), not merely submitted; the Coinbase/EY survey allocation figure was set to the January-2026 vintage (18%→29% planning >5%); and the Standard Chartered DAT view is now paraphrased, not quoted. An earlier verification vote also confirmed IBIT's $46.82B net assets and Jan 5, 2024 inception/0.25% fee against BlackRock's own fund page (3-0).

Adversarial steelman pass (v1.2). The note was then subjected to four independent adversarial reviews, each tasked with destroying one pillar (the institutional base case, the store-of-value premise, the supply-shock argument, and the dynamic substitution model). They produced the material changes reflected in §10 and throughout: (1) §8's v1.1 claim that a tight float "raises the top" was a sign error — the float is price-elastic (LTHs released ~1.8–2M coins into each recent bull market; the market absorbed Germany's ~50k-BTC forced sale in mid-2024 with only a transient dip) — corrected, and the supply-adjusted bull widened back to ~$490–900k; (2) the Q1 2026 13F reversal (professional holdings −17% in coins, hedge funds −39%) was verified against CoinShares and now leads §2 in place of the trailing FY2025 growth figure; (3) §7's "decorrelation is already resolving" was removed — the 2025–26 gold-record/BTC-drawdown period is the thesis's failed natural experiment; (4) §7 now shows real-dollar figures beside nominal (~72% of the $6.5M/2050 is debasement of the measuring stick) and concedes the competable-pool haircut that makes the VanEck comparison a straddle, not a clean beat; (5) §5 now shows GLD's full post-2011 path (a ~one-third AUM collapse and years of recovery), elevating the "the wave was the trade" bear. The verified new facts (Q1 2026 13F figures; GLD 2013 outflows ~$24B+/552t) were confirmed against primary reporting before insertion. The steelman did not overturn the note's defensible core — the §6 2030 allocation lens (~$180–250k) — but it materially humbled everything beyond it and shifted the near-term balance toward the bear.

High-confidence, multiply-sourced: the ETF stock and flow figures (SoSoValue daily data cross-checked against BlackRock's fund page); DAT totals (~1.26–1.27M BTC, cross-checked across bitcointreasuries.net and The Block's independent tracker, which differ on company count — 197 vs 85 — but agree on BTC totals within 0.5%); Strategy's 843,775 BTC and its July 6, 2026 first sale; the US government's ~328k BTC and un-codified SBR status (CoinDesk, July 6, 2026); Mubadala/Al Warda/Wisconsin/Harvard/Brown 13F positions; the AIMA/PwC 7%/71%/55% provenance and skew; WTW pool sizes (P22 pensions $68.3T at year-end 2025 per GPAS 2026; insurance $40.2T, SWF $13.7T, endowments/foundations $2.3T per the 2025 study); Global SWF's $60T→$80T; GLD's adoption milestones; the Basel 1,250%/Group-2-limit framework and its November 2025 review; the August 7, 2025 401(k) EO and the DOL rule's March 24, 2026 OMB clearance/public-comment close; MSTR's Nasdaq-100 entry and S&P 500 pass-over.

Supply side (§8), sourced: ~20.05M BTC mined of a 21M cap (July 2026); post-April-2024 issuance 3.125 BTC/block ≈ 164,000/yr (≈82,000 after the April-2028 halving), so ~0.5M new coins across the 2026–2030 horizon; Glassnode illiquid supply ~14.4M (roughly 72–74% of supply, near record highs) with liquid ~2.6M and highly-liquid ~3M; exchange reserves ~2.2–2.4M, a 7–9-year low (down from ~3.2M in 2023); lost/dormant ~3.1M (range 2.7–3.9M). Our four-way float decomposition maps lost + long-term-illiquid to Glassnode's "illiquid" bucket and uses the exchange-reserve figure (~2.4M) as the tighter "available float." The four-way float decomposition is illustrative and labeled as such; the flow-imbalance figures and the supply-feasible price corridors are model outputs from model.py. Methodological corrections: (v1.1) §8's absorption analysis originally divided each scenario's full dollar delta by a clearing price, which conflates appreciation of coins already held with fresh purchases; it was replaced with the stock-consistent method (institutions' 2030 coin holdings bounded between today's ~2.24M and ~5.14M by the absorbable float). (v1.2) An adversarial-review pass then flagged that v1.1 over-claimed the supply constraint "raises the top" (toward $650–900k) — a sign error. The float is price-elastic (long-term holders released ~1.8–2M coins into each of the last two bull markets, and the market absorbed Germany's ~50k-BTC forced sale in mid-2024 with a transient dip), so a tighter float is a two-way volatility amplifier and price-transmission mechanism, not a directional bull driver; the supply-adjusted bull band was widened back to the full ~$490–900k. The net-inflow "multiplier" remains referenced as contested and deliberately not used in the price grids (which price demand 1:1, the conservative choice).

Halving cycle (§9), sourced: halving dates (Nov 28 2012, Jul 9 2016, May 11 2020, Apr 19-20 2024; next est. Mar–Apr 2028 at block 1,050,000) are protocol facts; cycle peaks used: ~Nov 30 2013 (~$1,150), Dec 17 2017 (~$19,700), Nov 10 2021 (~$69,000), Oct 6 2025 (~$126,200 — verified earlier in this note), giving peak lags of ~367/526/548/534 days; cycle troughs: ~Jan 14 2015, Dec 15 2018, Nov 21 2022 (~777/889/924 days post-halving; ~−84/−84/−77% from peak). Halving-to-peak multiples (~95×/~30×/~8×/~2×) are computed from approximate halving-day prices (~$12/~$650/~$8,700/~$64k) and are stated as approximations. The "template bottom Q4 2026" and "next peak mid-2029–2030" are pattern extrapolations from n=4 cycles, labeled as such, not forecasts.

Second-lens (§7) monetary pools, sourced but order-of-magnitude: gold above-ground ~$29T at ~$4,150/oz (World Gold Council ~216,265 tonnes), of which ~$13–15T is investment/official gold; the 27%-of-official-reserves / passing-US-Treasuries milestone (WGC/ECB, end-2025, first time since 1996); global bonds ~$143T (2026 industry estimates in the SIFMA/ICMA tradition); global real estate ~$393T (Savills, start-2025); global broad money ~$99T (M2 aggregates, early 2026); Tether ~97,000 BTC and its 15%-of-profits policy; Strategy's ~$8.2B convertible stack and STRK/STRF/STRD preferreds. These aggregates are inherently fuzzy and definitionally sensitive; §7's capture rates are explicitly modeling choices, and the whole section is flagged as more speculative than the rest of the note and conditional on a correlation/volatility transition that has not yet occurred. The terminal-arithmetic decomposition — that ~20–30% of bonds, ~10–20% of real estate, and ~30–50% of broad money constitutes competable store-of-value premium (~$100–170T) — is our own illustrative estimate with no external source, stated so the reader can substitute their own fractions. The dynamic substitution model's parameters are likewise argued choices, disclosed in full: pool $163T growing 5.5%/yr (deliberately below trend global M2 of ~6–7%), p₀ = 0.78% (measured: today's market cap over the pool), r = 0.20–0.30 (anchored to the measured ~30%/yr penetration-growth of 2020–2026, haircut for scale; that history is a single adoption run, n = 1), K = 5/25/50% terminal share. All paths are reproducible from model.py. VanEck's $2.9M/2050 base, trade-settlement mechanism, 2.5%-of-reserves assumption, and ~$53M hyper-tail are their published figures, engaged with rather than endorsed.

Deliberately hedged, not asserted as hard fact: the Bank of America/Merrill "up to 4%" advisor policy (single-outlet reporting; treated as "reported"); China's and the UK's government BTC totals (estimates of seized coins, not audited disclosures); the ~28% institutional share of ETF assets (a Bernstein estimate, methodology-dependent); the 60% liveliness/free-float figure (ARK's derivation from Glassnode, labeled experimental by ARK itself); the consumer BTC-loan "$1T in a decade" projection (a single SVB-cited scenario, quoted as such); and all steady-state penetration percentages in §6, which are modeling choices — disciplined by the cited anchors, but choices. The model's arithmetic is reproducible from model.py in this note's repository folder.

Version 1.2 · analyst: TON618 Thematic Research. (v1.1 added the §7 store-of-value lens, §8 supply constraint, §9 halving cycle, and the stablecoin/credit/family-office/miner channels. v1.2 ran the note through four adversarial steelman reviews — correcting the §8 supply sign error, leading §2 with the Q1 2026 13F reversal, demoting the §7 decorrelation claim, adding real-dollar terms and the pool haircut, restoring GLD's full path, and adding the §10 steelmanned bear case.)