An unallocated treasury is not unallocated. It is one hundred percent long a single unit of account — one that gave up roughly 21% of its purchasing power between 2019 and 2025. We size, execute, custody and report a small bitcoin sleeve against that exposure: an asset that has fallen 85% peak to trough, and is therefore only defensible at a weight the board can afford to lose in full.
Purchasing power compounds published US CPI for 2020–2025; the 2025 row is an illustrative assumption rather than published data. The 85% figure is the deepest peak-to-trough decline in the drawdown history shown below; the stress case modelled through the rest of this page is a stated 75% assumption. Nothing here is a forecast or investment advice.
Illustrative sizing at a 3% sleeve. A scenario on stated assumptions — not a recommendation, not a forecast, and not a promise of any outcome.
Minimum allocation
The smallest balance-sheet position the desk will open and report on.
Settlement
Fiat against delivery the same business day, subject to wire cut-off times.
Segregated custody
Held with a qualified custodian in your entity's name. Not lent, not rehypothecated.
Desk coverage
A named advisor during your hours, plus an out-of-hours line for execution windows.
A balance held flat since 2019 has not held its value; it has quietly financed six years of inflation. The three lines below are the same opening treasury, run through the same years, measured in what it could actually buy.
Index, base 100 at the 2019 opening date, deflated by annual US CPI. Above the dashed line the balance buys more than it did in 2019; below it, less.
Sources for 2019–2024: annual US CPI, the average 3-month T-bill yield, and bitcoin calendar-year total return. The 5% sleeve is set once at the 2019 opening date and is never rebalanced or timed, while the cash leg compounds at the bill yield. The 2025 sample is not published data: it is an illustrative row on stated assumptions, drawn dashed and marked with an asterisk, and it carries a material share of the sleeve’s cumulative result.
-20.8%
real terms, 2019–2025*
*2025 is an illustrative row on stated assumptions, not published data, and it lifts this figure materially.
A balance left in an operating account buys 79.2% of what it bought in 2019. Nothing was lost on paper — the statement never moved.
-6.8%
real terms, 2019–2025*
*2025 is an illustrative row on stated assumptions, not published data, and it lifts this figure materially.
Yield closed most of the gap and not all of it. In 2021 and 2022 the bill paid far less than the CPI print it was running against, and the ladder never made that back.
+58.5%
real terms, 2019–2025*
*2025 is an illustrative row on stated assumptions, not published data, and it lifts this figure materially.
Ninety-five percent in bills, five percent set once in 2019 and never rebalanced. One small position carries the whole result, which is exactly why it cannot be relied on to repeat.
The same path falls 19.9% in real terms into 2022, ending that year at 94.3 — below the 2019 baseline. A board has to be able to sit through that.
Figures for 2019–2024 are published annual data. The 2025 row is an illustrative assumption rather than published data, is shown dashed and asterisked throughout, and lifts the sleeve’s terminal index materially. These index paths are arithmetic on those inputs with no rebalancing, fees, taxes, or execution costs, and are not a forecast, a performance record, or investment advice. A bitcoin sleeve can lose most of its value in a single year — the 2022 leg of the same path is shown above rather than smoothed away.
Set your balance, the share of it you would commit, and how long the board expects to hold. Everything below is scenario arithmetic on the stated assumptions — not a forecast, and not a promise of any outcome.
Nothing here is sent anywhere — the model runs in your browser.
$1M – $500M
0% – 10% of treasury
1 – 10 years
Total treasury value from today to year 5, under three stated bitcoin scenarios and against holding cash alone.
The cone spans the bear and bull assumptions; it is not a confidence interval, and outcomes outside it are possible in both directions. Each path applies a fixed annual rate to the sleeve while the cash leg compounds at the T-bill assumption throughout.
Before any market move. Segments are drawn to scale.
$24.7M97%
$765K3%
Capital allocated
$765K
3% of treasury, committed at inception
Operating cash retained
$24.7M
Stays in the T-bill ladder at an assumed 4.2%
Base case, treasury at year 5
$32.3M
Bear $30.8M · Bull $34.8M
Versus doing nothing
+$964K
Cash-only benchmark reaches $31.3M
Base case in today's money
$28.3M
Purchasing power after an assumed 2.7% inflation
Blended treasury
4.61%
Expected return · volatility 2.42%
If bitcoin fell 75% tomorrow — the stress assumption used throughout this page, against a deepest decline on record of 85% — the sleeve marks down from $765K to $191K. At the treasury level that is -2.25% on a $25.5M balance sheet. The other $24.7M is untouched — it never left the cash ladder. That is the case for a 3% sleeve: the loss is bounded, quantified in advance, and survivable as an accounting event rather than a solvency one.
Sleeve after the fall
$191K
Mark-to-market loss
-$574K
Treasury impact
-2.25%
Treasury still standing
$24.9M
Drawdowns of this depth have taken as long as three years to recover, a future decline can be deeper than any on record, and there is no assurance that any of them recovers at all. Size the sleeve so that the answer to “what if we are wrong” is a number your board can sign off in advance.
Scenario arithmetic on the assumptions shown — not a forecast, not a recommendation, and not investment advice. The bear, base and bull paths apply a fixed annual growth rate to the sleeve; real bitcoin returns arrive in violent, unevenly spaced moves, and a path that ends well can spend years underwater on the way. Past performance is not indicative of future results. Take the output to your own advisers before committing balance-sheet capital.
A volatility ceiling is a control your committee already knows how to write. Set the ceiling, and the arithmetic returns the largest sleeve that fits inside it. No optimiser, no view on price.
Blended treasury of a T-bill ladder plus a bitcoin sleeve, 0% to 20%. One scale, both axes annualised.
Cash is modelled at the risk-free rate, so this curve has no interior optimum to find: expected return and volatility both rise with the sleeve the whole way along. The binding constraint is the ceiling your policy sets, not a maximum somewhere on the line. Assumptions, not forecasts.
A 5.0% volatility ceiling permits up to a 7.6% bitcoin sleeve. Expected treasury return rises from 4.20% to 5.25%; blended volatility 5.00%.
That is a pickup of 105 basis points of expected return, on these assumptions. The other side of it: the stated stress case used throughout this page, a -75% fall in the sleeve, would take roughly 5.7% off total treasury value. The deepest decline on record is deeper still, at -85% — see the drawdown history below. Sizing from a ceiling is what keeps that second number inside what the board has already agreed it can absorb.
Every figure in this section is arithmetic on these inputs. They are assumptions, not forecasts, and your own investment committee should set them before any of it becomes policy.
| Leg | Expected return | Volatility |
|---|---|---|
| Cash, T-bill ladder | 4.20% | 1.20% |
| Bitcoin sleeve | 18.00% | 60.00% |
Scenario arithmetic on the stated assumptions, not a forecast, not a recommendation, and not investment advice. The expected returns, volatilities and correlation above are inputs chosen for illustration; realised outcomes will differ, and a bitcoin sleeve can lose the greater part of its value. Past performance is not indicative of future results.
An allocation you cannot survive is an allocation you should not make. Every peak-to-trough decline of the modern market is below, at full depth, with the time it took to get back to the prior high.
Deepest
−85%
Longest recovery
36 mo
Bitcoin spot, measured from each cycle high to its trough. Bar length is the loss; the figure at the right of each row is the time from trough back to the prior peak.
Historical price history, not a projection. Recovery is measured in calendar months back to the prior nominal high and ignores the opportunity cost of the capital while it was underwater.
This is the sum a treasurer is already doing in their head. A sleeve held at 3% of treasury turns each of those declines into the figure in the right-hand column, with the rest of the balance sheet held unchanged.
| Event | Sleeve drawdown | Cost at 3% of treasury |
|---|---|---|
| Mt. Gox collapse | −85% | −2.55% |
| ICO unwind | −84% | −2.52% |
| Rate shock & credit contagion | −77% | −2.31% |
| COVID liquidity shock | −50% | −1.50% |
Arithmetic on historical drawdowns at a stated 3% sleeve weight, not a forecast and not investment advice. A future decline can be deeper than any of the above, and a sleeve can lose its entire value.
Recovery took years, not weeks. The two deepest declines needed 36 months (3.0 yrs) and 35 months (2.9 yrs) to reach the prior high again. A treasury that may need the capital inside that window should size accordingly, or not allocate at all.
Nothing about the mandate changes that. Custody, execution quality and reporting control operational risk; they do not control price. If a −85% mark on the sleeve would force a sale, breach a covenant, or put a payroll or a debt service date at risk, the right weight is smaller than the one under consideration, and it may be zero.
The honest test is not whether the position recovers. It is whether the business can hold it while it does not, without the decision being made for it.
Every mandate buys the same asset through the same desk under the same custody. What differs is cadence, how the block is worked, and how much governance sits around it.
A ring-fenced position sized to a board-approved risk budget.
$250,000
Opening allocation
A scheduled buy programme that removes the timing decision.
$1,000,000
Committed programme
Balance-sheet scale, with governance and liquidity to match.
$10,000,000
Mandate size
Minimums are opening allocations, not standing balances, and can be scaled or unwound on written instruction. Every mandate is documented in a written investment policy statement agreed with your board before any capital moves. Custody, insurance and liquidity terms are set out in the custody agreement and the mandate rather than here; ask the desk for the custodian’s policy summary, including its limits and exclusions, before you rely on it.
Five stages, each with a named control and an audit trail behind it. Every ticket is dual-approved, every fill is measured against a reference, and the asset is segregated the moment it settles.
Authorised signatories submit an order through the corporate portal or the desk line. Dual approval is enforced on every ticket.
The desk works the block across OTC counterparties and regulated venues rather than sweeping a single order book.
Fiat settles against delivery on the same business day. Nothing leaves your control before the asset is confirmed.
Assets land in segregated cold storage held by a qualified custodian, addressable only under your entity's key policy.
Holdings are published to the reserve attestation and reconciled to your ledger with a cost-basis file your auditor can take.
No single person can move value. An instruction needs two authorised signatories before the desk works it, and both are named in the audit record.
Each fill is reported against a volume-weighted average price reference for its execution window. The reference is a measurement of the fill, not a guaranteed price.
Coins sit in segregated cold storage with a qualified custodian. They are not lent, staked, or pledged as collateral at any point.
These controls govern operational risk — who may instruct, where the asset sits, what the auditor receives. They do not reduce market risk. Bitcoin remains volatile, a disciplined process does not stop the position falling in value, and drawdowns deeper than 70% have occurred within a single cycle.
Five steps, each with a named owner on our side. Corporate KYB is the long pole; everything downstream is documentation your board already knows how to read.
Day 1
A 45-minute working session with the desk covering objectives, risk budget, and the accounting treatment your auditor will expect.
Days 2–5
Corporate KYB: certificate of incorporation, ownership structure, authorised signatory list, and source-of-funds documentation.
Week 2
We draft the investment policy statement and execution mandate for board review — allocation ceiling, rebalancing rules, and reporting cadence.
Week 2–3
Custody accounts are opened, approval thresholds configured, and a test settlement is run end to end before any size is committed.
Week 3
The opening tranche is worked, settled, and reconciled. Your reporting pack and portal access go live the same day.
Six controls that turn a treasury allocation into a documented, reviewable position: an adoptable policy pack for the board, and an independent record for the audit file.
Under ASU 2023-08 crypto holdings are measured at fair value each period with changes through net income — no more one-way impairment. We supply the period-end marks and the movement schedule.
A drafted investment policy statement, risk memo, and resolution template your board can adopt or amend, so approval is a review rather than a research project.
Assets sit with a qualified custodian in segregated cold storage — never commingled with the firm's balance sheet and never lent out.
Holdings are reflected in the published proof-of-reserves attestation, with an addressable record your auditor can verify independently.
Signatory lists, per-ticket limits, and dual approval on every instruction. Changes require written notice from an existing authorised officer.
Monthly statements, quarterly attestations, and an annual cost-basis pack aligned to your financial year end.
Accounting treatment summarised for orientation only; it is not accounting, tax, or legal advice. Confirm the treatment for your reporting framework and financial year with your own auditor before adopting a policy.
Accounting treatment, sizing, liquidity, custody, cost and tax — answered at the level of detail an audit committee expects.
As an intangible asset measured at fair value. ASU 2023-08, effective for fiscal years beginning after 15 December 2024, requires crypto assets to be remeasured each reporting period with gains and losses recognised in net income. That replaced the old impairment-only model, under which a recovery in price could never be written back up. We provide period-end marks, the fair-value movement schedule, and the disclosure inputs your auditor will request.
General information for corporate treasury teams, not investment, legal, accounting or tax advice. Accounting and tax treatment depends on your jurisdiction, reporting framework and structure; confirm the position with your own auditor and tax adviser before acting.
Tell us the shape of the balance sheet and the desk comes back with sizing, custody structure and reporting mapped to your entity, before anything is committed.
A treasury allocation is a governance decision before it is a market decision. The briefing answers the four questions a board asks: how much, held how, marked how, and what the balance sheet looks like through the 75% stress case this page uses — and through the 85% decline that is the deepest on record.
The downside arithmetic arrives in the same document as the upside, on stated assumptions rather than as a forecast. If a bitcoin sleeve is wrong for your cash profile, we would rather say so on the first call than the third.
The desk