SBET Absolute Valuation

Where it stands right now

USD · per share

Four figures, each with the comparison that belongs to it: what a share owns, what the ETH alone is worth, what the market is paying for it, and what the assets plus their yield come to.

NAV per share · economic

Market price {{ v.sbetPrice }} · a {{ v.discountToNav }} {{ v.discountNavWord }}

ETH value per share · basic-equivalent

Read with NAV per share {{ v.navPerShareEconomic }}, never on its own

Economic mNAV

Market price {{ v.sbetPrice }} ÷ NAV per share {{ v.navPerShareEconomic }}

Assets + yield

Assets {{ v.navPerShareEconomic }} + net yield {{ v.yieldPerShare }} · market {{ v.discountToLine }} {{ v.lineGapWord }}

ETH per 1,000 shares · economic

ETH per 1,000 shares · basic-equivalent

About 9% higher than the company basis on the same holdings, so never a like-for-like read against a published figure. Quoted against the 2.00 launch figure it implies growth above 100%, which is a denominator artefact rather than strategy.

Assumed diluted · company basis

Since launch 2.00 · {{ v.concVsLaunch }} · vs Jan-2026 peak 4.03 · {{ v.concVsPeak }}

Comparable to the company’s filings: the deck’s 4.03 at 27 January 2026, the weekly 8-K series of 3.66 / 3.59 / 3.87 / 3.80 / 3.94 / 4.00 through 2025, and the quarterly releases, which define the metric on this basis. The live dashboard’s {{ l.dashConc }} is on basic-equivalent shares and is comparable to that column instead. The 3.66 to 3.59 step in August 2025 is the company’s own caveat in action: it warns the metric “may experience temporary declines due to timing gaps between share issuance and ETH deployment,” and the $200m raise settled before the ETH was bought.

NAV per share above is shown on the economic basis. The other two, for comparison:

Basis

NAV per share

mNAV

Economic · treasury stock method · shown above

Basic

Fully diluted · company method

Basic ignores dilution. The company's fully diluted count includes out-of-the-money shares and credits no exercise proceeds. This app uses the economic basis: treasury stock method on the warrants that are in the money, Black-Scholes deduction on the ones that are not.

Where the price sits

Market price

NAV/sh {{ v.navPerShareEconomic }}

Assets + yield

Observed facts · cells

ETH price · USD SBET price · USD ETH · native staked ETH · lsETH + weETH Cash · $000 Liabilities · $000 Common shares outstanding

Cells, not sliders. These are externally observable. {{ l.sliderPrecisionNote }}

Judgements · sliders

Sliders carry judgement. Discount rate and horizon move assets plus yield and nothing else. {{ l.bsNote }}

Net asset value · $m

The deduction is real value that belongs to the warrant holder rather than to you. {{ l.warrantSplit }}

{{ v.lstBasis }} {{ v.lstGaap }} The carrying figure is a quarterly balance-sheet number and is fixed at that date; the spot figure moves with the inputs above.{{ v.ethSplitCaveat }}

Share count · four conventions

Four counts, two of them load-bearing. Treasury stock method for the point-in-time question above. Full dilution with its exercise proceeds for scenarios that run the share price somewhere else. Basic equivalent and the company's fully diluted count are shown for comparison and drive nothing. {{ v.basisSpread }}

Assets plus yield · per share

The discount rate and horizon are inputs, not facts. This figure also excludes the ATM option, ecosystem optionality and the governance discount. On the current inputs those net to {{ v.optionalityClause }}.

Net yield, year one · $m

That SG&A line is why this app shows both metrics. It reduces NAV per share every year. It changes ETH per 1,000 shares by exactly 0.00%, because cash appears nowhere in that formula.

Does the machine create value

Marginal action · $50m

Two tests sit below. The first asks what a buyback does at each price. The second asks what happens when the company pays its bills. Read the two columns together in both. Where they disagree, only NAV per share has registered the spend.

Repurchase price

NAV per share

ETH per 1,000 sh

Against buying ETH

The middle column never turns negative. At every price, ETH per 1,000 shares rises, because cash appears nowhere in that formula, so spending it cannot register. The left column changes sign. The benchmark in the right-hand column is not "do nothing", it is "buy ETH instead", so the breakeven sits at ETH NAV per share {{ v.ethNavPerShareBasic }} rather than at zero.

NAV per share

ETH per 1,000 sh

Read both columns · the 0.00%

Identical economics, different pocket, different reported result. Funded from cash, the spend cuts NAV per share and moves ETH per 1,000 shares by 0.00%. Funded by selling {{ v.ethSold }}, the same spend appears in both columns. The direction of ETH per share therefore depends on which account the bill is paid from, so an operating burn can be absent from it entirely.

What the price does as ETH moves

Chart 2 · per share

At the observed multiple {{ v.regimeObserved }} At justified {{ v.justifiedScenario }} · full dilution Assets plus yield NAV per share · parity

Price per share at each ETH level and multiple

Deliberately unlabelled

ETH price

NAV per sh

at {{ h.label }}
{{ h.sub }}

Assets + yield
full-dilution basis

Why these columns differ from the figures elsewhere in the app

No case here is called bear, base or bull. Naming them would tell you which one to believe, and the model has no basis for that claim. Each column is a multiple you might assume, each row an ETH price you might assume. The pairing is yours to make. The shaded row is where ETH trades today.

Judgements on this view

The ATM option is set by hand and stays that way. Translating a capital allocation track record into a forward option value is a judgement, not an arithmetic step, so nothing wires the scorecard into this slider. Note it is not zeroed at a discount either: an out-of-the-money option is not worthless, which is the same argument that puts a Black-Scholes deduction on the out-of-the-money warrants. The completed record reads three wins, two losses and one wash, not a clean run of accretive raises, and it is now set at 2% against that rather than the 5% carried while the record was unmeasured. Net optionality turns negative at that level, which is the point: the governance discount outweighs what the record supports paying for future issuance.

Justified mNAV · reconciliation trace

Coherence check · net optionality identity {{ v.identityOk }}

Justified {{ v.impliedJust4 }} less assets plus yield {{ v.lineInSand4 }} {{ v.identityLhs }} Net optionality {{ v.netOptionality }} × economic NAV per share {{ v.navPerShareEconomic }} {{ v.identityRhs }}

The two sides agree only when the yield term, assets plus yield and the per-share conversion all sit on one basis. If a later edit mixes bases, this row breaks first, and names which basis drifted.

This view carries two bases on purpose, and the split is the rule rather than an inconsistency. The trace above is point-in-time work, so it runs on the treasury stock method throughout: the yield term divides by economic NAV, the multiple converts at economic NAV per share, and the comparator is observed economic mNAV. The grid and chart 2 are scenario work, so they run on full dilution with proceeds, which needs no share price to resolve moneyness and therefore cannot go circular when the output is a price. Basic mNAV stays on the Overview as the company-comparable figure.

The 2025 raises, tested against the company's own characterization

August 2025 at $21.76 was clearly accretive. For NAV per share to have reached the issue price on 596,800 ETH and 139,294,471 shares, ETH would have had to be worth roughly $5,300. Its 2025 high was under $5,000.

October 2025 at $17.00 was roughly a wash. 840,100 ETH at the company's own disclosed $3,892 per ETH that week, over 192,193,191 shares, is about $17.01 per share from ETH alone.

Management described that raise as executed at a premium to NAV. The pricing release describes a 12% premium to market price: $17.00 against the $15.15 close on 15 October. The NAV claim is not repeated here as fact.

The warrant stack

At {{ v.sbetPrice }} · vol {{ v.volPct }}

Every tranche is handled twice, depending on where it sits. In the money, its shares enter the count and its exercise proceeds retire stock at market. Out of the money, no shares enter, but its Black-Scholes value is deducted from NAV. An option that is out of the money today is not worth zero.

Warrant shares

Strike

Term

State

BS value

NAV deduction

Net shares added

Separately, {{ v.prefunded }} pre-funded warrants carry a {{ v.prefundedStrike }} and are counted in the basic share count rather than here. At that strike they are shares in all but name. The {{ v.optionCount }} employee options struck at {{ v.optionStrike }} sit far out of the money: the company's count includes them as shares, this app does not.

What each convention does to the same NAV

Dilution waterfall

Convention

NAV used

NAV per share

mNAV

The company's method counts more shares and uses NAV before the warrant deduction. Both push the multiple up, so at a discount its own figure makes the stock look less cheap than it is. Basic moves the multiple the other way. The middle row is the one this app uses.

The three definitions compared

Basic Fully diluted · company method Economic {{ ch1.xLabel }}

Source note · the seeded prices

{{ l.seedNote }} The $4.69 that previously seeded the June date is the company’s weighted-average repurchase price over 24–26 June, and it is retained where it belongs, as the figure the capital allocation scorecard uses. Every mNAV on this view and the moneyness of every tranche above depend on the share price. NAV per share is price-sensitive through the treasury-stock count and the Black-Scholes deduction; ETH per 1,000 shares is not.

Why this model measures what it measures

Three choices in this model are non-obvious. Each is defended here, including where it is weak.

A · Why economic mNAV is the right multiple

All three definitions answer one question: how much am I paying per dollar of NAV? The answer depends on how the claims on that NAV are counted. Basic and fully diluted count them differently, and move the multiple in opposite directions.

Basic mNAV ignores dilution. If a holder owns 10.0m warrants struck at $8.15 and the stock reaches $20, those shares will exist and will claim NAV. Basic leaves them out, which widens the apparent discount.

Fully diluted mNAV, the company method, does two things. It counts every warrant and option share regardless of moneyness, including the 3,146 options struck at $122.88. And it credits no exercise proceeds: if the $8.15 warrants are exercised the company receives roughly $81.6m of cash that lands in NAV. The 10-K and 10-Q state that Assumed Diluted Shares Outstanding is not calculated using the treasury stock method.

A larger denominator and a NAV before the warrant deduction both push the multiple up, so at a discount the company's own fully diluted figure makes the stock look less cheap than it is.

Economic mNAV treats moneyness as a state rather than a fact. A warrant is a probabilistic claim, so it is handled twice. If it is in the money now, the treasury stock method applies: shares in, exercise proceeds used to retire shares at market, only the net counted. If it might be in the money later, its Black-Scholes value is deducted from NAV, because that option value belongs to the holder rather than to you. Basic captures neither effect. The company method adds the shares without the proceeds, and leaves the deduction out.

The company concedes the gap this closes. Its dashboard disclaimer states that these metrics do “not take into account that our assets are subject to all existing and future liabilities.” Economic mNAV is the version that does.

Live comparison at the current share price

Where it bites. The divergence is widest at a discount and narrows as the stock rises, the opposite of the intuition. At $5 the company count sits roughly 7.4% above the economic count; by $30 that falls to about 1.8%. As the price climbs, exercise proceeds shrink relative to the price, so the treasury stock offset gets smaller and the two methods converge on full dilution. So the choice of metric matters most right now, when an investor is sizing a discount.

Its own weakness. The treasury stock method assumes exercise proceeds are used to repurchase stock at the prevailing price. That is an accounting convention, not a prediction of what management will do. And the Black-Scholes deduction depends on the volatility input, which is an estimate; 100% is used here. Best of the three, not perfect.

B · Assets plus yield

Assets plus yield is not a price target. It is a decomposition, and it answers one question: what is SBET worth if management never touches the capital markets again? Two components, neither of which requires anything new to happen.

Part 1 · assets that already exist. Total ETH holdings plus cash, less liabilities, divided by shares. Worth this much whether or not anyone does anything. This is NAV per share.

Part 2 · cash flows the existing base throws off. Staking, liquid staking and restaking accretion, plus protocol incentives and fund returns, less the SG&A required to run a public company. Discounted over a finite horizon. Hold the ETH, stake it, collect the incentives, pay the bills.

Anything above that figure requires a claim about future capital allocation: that management will sell equity above NAV and manufacture ETH per share for existing holders. That machine is real. It took ETH concentration from 2.00 at launch to a peak of 4.03 in January 2026, on assumed diluted shares. It has not held there: concentration is {{ v.concVsPeak }} off that peak, so the machine has been running in reverse rather than flat. It needs mNAV back above 1.0× and a marginal buyer willing to pay the premium. Neither is contractual, neither is within management's control, and both were absent during 2026.

So the figure tells you what you are underwriting. Below it, you are buying assets plus yield at a discount. Above it, you are also paying for management's future ability to sell stock above NAV to somebody else.

Caveats. The discount rate and horizon are inputs, not facts. At a lower rate in perpetuity the yield component is far larger and the figure moves up substantially. It also excludes the ATM option, ecosystem optionality and the governance discount. On the current inputs those net to {{ v.optionalityClause }}, which puts the justified price of {{ v.impliedJust }} {{ v.justVsLineWord }} assets plus yield of {{ v.lineInSand }}.

C · What ETH per share does and does not measure

ETH concentration is ETH units divided by shares per thousand. Cash appears nowhere in that formula. So a cash-funded buyback leaves the numerator untouched and shrinks the denominator: the metric rises, arithmetically, at any price paid.

The fair objection: cash is fungible with ETH, so spending it has an opportunity cost. That is right. The right benchmark is not "do nothing", it is "buy ETH instead". Measured against that benchmark the breakeven is ETH NAV per share, not zero. Buying back below ETH NAV per share beats buying ETH; above it, buying ETH wins.

That breakeven is on basic-equivalent shares, the base a repurchase actually retires. The same holdings on economic shares are {{ v.ethValuePerShareEconomic }}.

The metric still rises at every price. A repurchase at three times NAV reduces NAV per share, and ETH per share goes up. Only NAV per share registers the price paid. That is why this model never displays one without the other.

Operating expense is the larger case, not buybacks. SG&A of roughly {{ v.sgaPlain }} a year, funded from cash, reduces NAV per share every year and moves ETH concentration by 0.00%. The same is true of the ecosystem grants. The operating burn does not appear in the metric at all.

Fund that same {{ v.sgaPlain }} by selling ETH and the metric falls. Identical economics, different pocket, different reported result. The direction of the metric depends on which account the bill is paid from.

Basis is not a detail here. The company publishes this metric on two different denominators. Its filings and press releases — the weekly 8-K series, the January 2026 deck figure of 4.03, the quarterly releases — define it on assumed diluted shares. Its live investor dashboard defines it on basic-equivalent shares: per the dashboard's own tooltip, ETH and as-if redeemed ETH equivalents, measured at midnight UTC on the most recent Sunday, divided by each 1,000 basic-equivalent shares outstanding, where basic-equivalent means common shares issued and outstanding plus shares issuable on exercise of pre-funded warrants. That is the definition used in the basic-equivalent column here: {{ l.basicEqSum }}. This model's basis is the company's published definition, not a construction of its own. Neither column is uniquely comparable to the company; each maps to a different company source. The assumed diluted count here is reconstructed from 10-Q components and need not match the company's exactly. Any history chart uses the assumed diluted column and says so on the chart. Mixing bases produces growth that the strategy did not deliver.

The company publishes basic-equivalent NAV per share on its own live investor dashboard, where it reads {{ l.dashNavPS }}, so the corrective is available in its own disclosure. That dashboard, not a filing, is the source for this one figure.

D · What the capital allocation scorecard cannot cover

The scorecard covers 3 of 6 deals, not the record. Its composition: FY2025 buybacks at $16.37 against NAV per share of $11.59, -$9.27m; the June 2026 registered direct, share leg, -$19.85m; June 2026 buybacks at a 29% discount, +$4.13m; partial total -$24.98m. That partial total is kept off the view rather than shown with a caveat, because a headline figure carrying a footnote gets read as a headline.

The three 2025 raises are blank by decision, not oversight. NAV per share at those dates is not derivable from disclosure: cash is reported inconsistently and on a yet-to-be-deployed basis, interim liabilities are never disclosed, and the pre-funded warrant split needed to reach basic-equivalent is not given. There is no input here inviting a user to fill them.

E · How the yield strip is capitalized

The strip is no longer uniform. Protocol incentives decay {{ l.decayPct }} a year, as documented, so they are discounted on their own declining schedule; every other component of net yield is capitalized as a level annuity over the horizon. The earlier convention held incentives flat for the full ten years, which overstated them.

Present value of the strip is {{ v.yieldPVM }}, giving assets plus yield of {{ v.lineInSand }}. The discount rate and horizon inputs move both terms; the incentive decay rate is fixed at the documented {{ l.decayPct }} and is not exposed as a slider.

F · AI provenance

Built with AI assistance. The workbook is generated by a Python script; Claude was used for construction, source retrieval from SEC filings, and arithmetic checking; Claude Design built this interface, and Claude Code packaged, audited, and shipped it. Every build is validated by an automated regression harness (18 checks) that verifies the workbook against its own source cells and recomputes key results from first principles.

Analytical judgements are the author's. Valuation conventions, share-count bases, the treatment of protocol incentives, the ATM option value, and the decision to leave unsourceable figures blank rather than estimate them are editorial choices, not model outputs.

Figures are tied to SEC filings and identified market data. Where a figure could not be sourced it is left blank and the gap is documented rather than filled.

The AI made errors during construction, including cross-referencing the wrong cells, quoting figures from superseded builds, and asserting a defect in code that was correct. Each was caught by review or by the harness. That is the reason for the harness, and the reason judgement stays with the author.