# Inside the Model: A Walk-Through of the Coinbase FY2019-FY2030 Equity Research Framework

> Architecture, assumptions, and outputs of the nine-tab financial model behind the Coinbase Infrastructure Flywheel series

- Author: Siddhant Shah
- Published: 2026-03-09
- Canonical URL: https://siddhants.com/blog/coin-research-202603/model-walkthrough/
- Series: Coinbase: The Infrastructure Flywheel, part 8 of 8 (https://siddhants.com/blog/coin-research-202603/)

## Illustrative exercise

This is a **demonstrative modelling exercise**, written to show a method: how the numbers are built, which assumptions carry the answer, and where the disclosure gaps are. It is **not a price target, not a recommendation, and not a view on the security**. Any valuation range shown is an output of the stated assumptions, not a judgement about what the shares are worth.

## Thesis

The model covers Coinbase Global, Inc. (Nasdaq: COIN) from FY2019A through FY2030E — eleven years of data. FY2019-2025 actuals are drawn from standardized financials and Coinbase's SEC filings. FY2026-2030 are built from editable analyst assumptions. It is a fully integrated three-statement model culminating in a two-method DCF.

The question the model is built to answer: is Coinbase's transition from a cyclically leveraged trading business into a diversified crypto infrastructure platform durable — and at what price is that transition already priced in?

The two-method DCF produces a valuation range of $146-$279 against a spot price of approximately $205 at model date. The $205 spot sits near the midpoint — consistent with a market that has partially priced the infrastructure transition but has not fully resolved the renewal and competitive risks.

## Model Architecture

The workbook has nine tabs organized into five functional categories.

The `Cover` tab is a live-linked dashboard displaying KPIs, trading multiples, the DCF price range, and navigation links. The thesis and the valuation summary are visible before opening anything else — by design.

The `Assumptions` tab is the single control panel for the entire model. Every forecast-period driver — macro variables, revenue build-ups, operating margin targets, balance sheet mechanics — flows from here. Change one cell and it propagates through all three financial statements, the DCF, and every supporting schedule simultaneously.

The three `Financial Statement` tabs (Income Statement, Balance Sheet, Cash Flow Statement) span FY2019A through FY2030E and are fully integrated: cash is the balance sheet plug, the cash flow statement reconciles net income to cash, and every major line traces back to Assumptions.

The `DCF` tab contains the WACC build, unlevered FCFF projections, terminal value under two methods, the equity bridge to per-share value, and two 5x5 sensitivity grids.

The four `Supporting` tabs are USDC Scenarios (the 3x3 Circle renewal risk matrix), Deribit Acquisition (PPA summary and annual amortization schedule), and Supporting Schedules (five sub-models covering debt, PP&E, share count, working capital, and tax).

## Where to Start

Download: [COIN Model.xlsx](https://siddhants.com/COIN%20Model.xlsx)

The model is designed to be read in a specific sequence that builds analytical context before introducing complexity.

Start with the `Cover` tab — thesis and valuation range upfront. Move to `Assumptions` next. This is the engine room, and understanding what drives the forecasts makes every subsequent tab legible. Then work through the `Income Statement` to see how the eight revenue lines collapse into EBIT and net income, the `Balance Sheet` to confirm the cash plug and the balance check row, and the `Cash Flow Statement` to see how OCF, ICF, and FCF connect the first two statements.

Once the three-statement model is clear, the `DCF` tab follows naturally: WACC inputs, FCFF derivation from the IS, terminal value under both methods, and the equity bridge. From there, the two scenario tabs — `USDC Scenarios` and `Deribit Acquisition` — add the risk and acquisition overlays. The `Supporting Schedules` tab closes the loop on the mechanics underlying the balance sheet and share count.

## Assumptions Tab

### Macro & Market

The Federal Funds Rate is the single most important exogenous variable in the model. It drives USDC revenue directly — through reserve interest income — and interest income on Coinbase's cash balance. The modeled FOMC path runs from $4.33\%$ in FY2025A to $2.25\%$ by FY2030E, producing a stablecoin revenue profile that peaks in FY2027E and then declines as rates normalize.

Crypto market cap grows from \$$3.2$T to \$$6.5$T across the forecast period — approximately $12\%$ CAGR — serving as the macroeconomic backdrop against which trading volumes and asset balances are sized. ETH staking yield compresses from $4.0\%$ to $2.7\%$ as network participation rises from $30\%$ to $40\%$ of total ETH supply. BTC year-end price assumptions range from \$$80$K to \$$120$K; ETH average price assumptions from \$$2,500$ to \$$4,500$.

### Transaction Revenue

Consumer transaction revenue is volume times take rate. Volume grows from \$$360$B (FY2026E) to \$$900$B (FY2030E); take rate compresses from $120$bps (FY2026E) to $85$bps (FY2030E), reflecting the combined effect of Coinbase One zero-commission waivers and increasing competitive pressure from offshore venues. Institutional volume grows from \$$982$B to \$$3,500$B — driven primarily by Deribit integration — at a blended take rate of approximately $5$bps.

Deribit is modeled as a standalone revenue line growing from \$$400$M to \$$600$M on \$$200$-\$$400$B in notional volume at $15$-$20$bps, reflecting its options-heavy mix, which commands higher per-notional fees than spot or futures. Other Transaction revenue (\$$75$M growing to \$$160$M) captures Base sequencer fees and CDP residual revenue not classifiable elsewhere.

### Subscription & Services

USDC revenue is built from first principles: on-platform USDC balance (\$$17.8$B growing to \$$28$B) multiplied by $100\%$ of reserve yield, plus off-platform USDC (\$$52$B growing to \$$92$B) multiplied by $50\%$ of reserve yield. The resulting gross revenue line is the most rate-sensitive in the model and the primary subject of the USDC Scenarios tab.

Blockchain rewards revenue is built from ETH custody volume ($2.5$M to $3.5$M ETH) multiplied by price and yield — and declines from \$$677$M in FY2025A to \$$425$M by FY2030E as yield compression outpaces asset appreciation in the base case. Coinbase One ARR grows from \$$180$M to \$$672$M as the subscriber base expands from $0.97$M to $2.8$M at an average revenue per subscriber of \$$185$-\$$240$ annually. Interest income steps up significantly in FY2026E (Deribit balance sheet addition) then declines as rates fall, ending at \$$185$M.

### Operating Expense & Margins

Gross margin expands from $80.5\%$ to $87.0\%$ as the revenue mix shifts toward higher-margin subscription and services lines. Sales & marketing declines from $14.7\%$ to $13.0\%$ of revenue. R&D declines from $23.3\%$ to $19.0\%$. G&A is elevated in FY2026E at $25.0\%$ to reflect Deribit integration costs, then normalizes to $18.0\%$ by the terminal year. SBC declines from $11.7\%$ to $8.0\%$ of revenue as the post-IPO equity program matures.

## Income Statement

The income statement spans FY2019A through FY2030E, with column C frozen for navigation. Historical figures are sourced from standardized filings; all forecast columns link directly to Assumptions.

Total revenue grows from \$$516$M (FY2019A) to \$$14.6$B (FY2030E). The composition changes materially: consumer transaction revenue — approximately $84\%$ of revenue in FY2019 — falls to approximately $52\%$ by FY2030E as institutional ($16\%$), USDC ($11\%$), subscriptions ($7\%$), and other revenue lines ($14\%$) absorb an increasing share. This is the diversification thesis made visible in the numbers.

COGS represents $13$-$20\%$ of revenue, with gross margin expanding to $87.0\%$ in the terminal year. Total operating expense falls from $60.5\%$ to $50.0\%$ of revenue. D&A includes three components: PP&E depreciation (from the PP&E schedule), Deribit acquired intangibles amortization (\$$194$M/year as detailed in the Deribit tab), and other intangibles. EBIT grows from \$$1.4$B to \$$5.4$B, with margin expanding from approximately $20\%$ to $37\%$. Below the line: interest expense of \$$230$M declining as debt is repaid, and an effective tax rate of $21\%$ from FY2026E forward. The Adjusted EBITDA memo at the bottom of the income statement shows margin expanding from $35.2\%$ to $49.1\%$ across the forecast period.

## Balance Sheet

Cash is the model's balance sheet plug. It equals prior-period cash plus the net change from the cash flow statement — no independent assumption. Under this structure, cash grows to \$$18.2$B by FY2030E and the company moves to a deeply net cash position of \$$15.4$B as debt is repaid and operating cash generation accumulates.

Crypto assets receivable and payable are modeled at \$$6$-\$$8$B gross, netting to approximately zero — consistent with Coinbase's custodial balance sheet treatment. PP&E is driven by the PP&E schedule. Goodwill is held flat at \$$4.17$B post-Deribit; no additional acquisitions are modeled. Intangibles decline from \$$3.4$B to \$$2.4$B as Deribit acquired intangibles amortize through the forecast period. Debt is modeled by named tranche, declining from \$$7.8$B to \$$2.8$B as tranches mature. Total equity grows from \$$14.8$B to \$$27.3$B.

The Balance Check row confirms model integrity every period. The two pre-IPO periods (FY2019 and FY2020) carry small residual balance check artefacts of \$$34$M and \$$316$M respectively — both traceable to pre-IPO capitalization treatment in the source data rather than model error, and flagged in the Analyst Notes tab.

## Cash Flow Statement

The cash flow statement is the integration layer connecting the income statement to the balance sheet via the cash plug.

Operating cash flow opens with net income, adds back non-cash charges (D&A, SBC, deferred tax), and incorporates working capital changes sourced from the Working Capital schedule. OCF grows from \$$2.4$B to \$$6.0$B across the forecast period, with FCF conversion above $1.0$x in every year — a reflection of the asset-light nature of Coinbase's core business.

Investing cash flow consists primarily of capex (\$$171$M-\$$292$M, modeled at $2\%$ of revenue) and the Deribit-related outflows in FY2025-2026. Financing cash flow includes no new debt issuance post-FY2025, debt repayments ranging from \$$270$M (FY2027E) to \$$1.7$B (FY2026E) as tranches mature on their contractual schedules, share buybacks of \$$1.5$B per year, and ESPP proceeds of approximately \$$100$M per year. The Levered FCF memo (CFO plus capex) grows from \$$1.9$B to \$$5.7$B.

## DCF Model

### WACC Build

The WACC is constructed from first principles with sourced inputs.

Risk-free rate: $4.02\%$, sourced from the FRED DGS10 10-year Treasury yield as of February 2026. Equity risk premium: $4.23\%$, sourced from Damodaran's implied ERP as of January 2026. Raw beta: $3.70$x, estimated from a five-year monthly OLS regression of COIN versus the S&P 500. Blume-adjusted beta — applying the standard two-thirds/one-third formula — is $2.80$x, reflecting mean reversion toward $1.0$x over time.

Cost of equity under CAPM: $15.86\%$ ($4.02\%$ + $2.80 \times 4.23\%$). Cost of debt: $3.5\%$, blended from Coinbase's actual debt tranches ($0.25\%$ convertible, $0.50\%$ convertible, $3.375\%$ senior, $3.625\%$ senior). After-tax cost of debt at $21\%$: $2.77\%$. Capital structure: $88.3\%$ equity (\$$58.8$B market cap) and $11.7\%$ debt (\$$7.8$B).

The resulting WACC is **$14.33\%$** — high by large-cap standards, and appropriate for a company whose cash flows carry $2.8$x systematic risk relative to the S&P 500.

### FCFF Bridge

Unlevered free cash flow to firm is derived as: GAAP EBIT x ($1 -$ tax rate) = NOPAT; plus D&A; minus capex; minus change in net working capital.

FY2026E FCFF is unusually low at \$$132$M due to a \$$1.6$B working capital outflow associated with Deribit integration. This is a one-time distortion — not a structural signal. FCFF normalizes to \$$2.4$-\$$4.4$B from FY2027E onward.

### Terminal Value

Two methods applied in parallel.

- **Method A — Gordon Growth Model:** Terminal growth rate of $6\%$, reflecting an aggressive but defensible assumption for a company exposed to secular crypto adoption tailwinds in a global TAM still in early innings. The $6\%$ rate is above long-run nominal GDP growth — which requires justification — but is calibrated against the trajectory of transaction volume and subscription ARR in the final explicit forecast year. Terminal value: \$$56.7$B; present value: \$$29.0$B.

- **Method B — Exit EV/Adj. EBITDA Multiple:** $18$x applied to FY2030E Adjusted EBITDA of \$$7.18$B. Terminal value: \$$129.2$B; present value: \$$66.2$B. The $18$x is calibrated against the peer set: CME Group at $22$-$24$x, ICE at $16$-$17$x, Nasdaq at $20$-$22$x, Robinhood at $25$-$33$x. At $18$x, Coinbase is discounted relative to pure-play exchange peers — appropriate given its higher cash flow volatility and crypto beta.

Terminal value represents $77\%$ of total enterprise value under the Gordon Growth method and $89\%$ under the exit multiple method. Both figures are typical for high-growth, back-loaded cash flow profiles and underscore how sensitive the valuation is to terminal assumptions.

### Equity Bridge & Price Targets

Enterprise value equals discounted FCFFs plus the present value of terminal value. Add net cash of \$$3.45$B, divide by $280.3$M diluted shares.

Gordon Growth method: **\$$146$ per share** (approximately $29\%$ downside to spot). Exit multiple method: **\$$279$ per share** (approximately $36\%$ upside to spot). The \$$205$ spot sits near the midpoint — consistent with a market that has partially priced the infrastructure transition but hasn't fully resolved the renewal and competitive risks.

### Sensitivity Tables

Two 5x5 grids let you stress the valuation against changing assumptions. The first crosses WACC ($10\%$-$18\%$) against terminal growth rate ($4\%$-$8\%$) for the Gordon Growth method. The second crosses WACC ($10\%$-$18\%$) against exit multiple ($14$x-$22$x) for the exit multiple method. Base case assumptions are centered in each grid. The grids show how quickly the upside case collapses at higher discount rates — and what the downside case looks like at lower terminal assumptions.

## USDC Scenarios Tab

This tab isolates the single largest binary risk in the model: the Circle-Coinbase revenue-sharing agreement renewal, due in 2026. Current terms — confirmed in Circle's S-1 — entitle Coinbase to $100\%$ of reserve interest on on-platform USDC balances and $50\%$ on off-platform balances.

The tab presents a 3x3 matrix crossing three Federal Funds Rate scenarios ($4.0\%$, $3.5\%$, $3.0\%$) against three sharing-rate scenarios.

The **Base case** preserves current terms ($100\%$/$50\%$). The **Bear case** reflects a renegotiation to $80\%$/$40\%$. The **Stress case** reflects a scenario where off-platform sharing drops to zero — pricing in a structure similar to Circle's disclosed Binance agreement (\$$60.25$M upfront plus monthly distribution fees), where Coinbase receives no flow-through from off-platform balances.

The output range across all nine cells: \$$570$M (stress terms, $3.0\%$ FFR) to \$$1,480$M (base terms, $4.0\%$ FFR). The model's base case USDC revenue assumption of approximately \$$1,436$M sits in the upper-left corner of the matrix — a reminder that any combination of rate cuts and renegotiated terms compresses this line significantly. The FY2025A implied yield of $3.08\%$ ($71\%$ of the average Fed Funds Rate) provides the sanity check anchor for the yield assumptions applied across scenarios.

## Deribit Acquisition Tab

This tab documents the \$$2.9$B acquisition (August 14, 2025) and its balance sheet and income statement consequences.

Total consideration: \$$700$M in cash plus approximately $11$ million COIN shares at approximately \$$200$/share, implying roughly \$$2.2$B in equity consideration. The deal was struck at approximately $9.7$x revenue on Deribit's estimated \$$300$M FY2024 revenue — a premium consistent with the platform's options market leadership and open interest depth.

Purchase price allocation is an analyst estimate; official PPA has not been published by Coinbase. The estimated breakdown: goodwill of \$$2,929$M; customer relationships of \$$710$M amortized over 8 years (\$$88.75$M/year); technology platform of \$$480$M over 6 years (\$$80$M/year); trade name of \$$120$M over 10 years (\$$12$M/year); non-compete agreements of \$$40$M over 3 years (\$$13.3$M/year); net working capital step-up of \$$91.5$M; deferred tax liability of −\$$471$M. Total annual amortization of acquired intangibles (ARI) is \$$194.1$M per year for FY2026-2028, declining to \$$180.8$M in FY2029-2030 as the non-compete agreements expire, and \$$100.8$M from FY2031 onward as the technology platform intangible rolls off.

This ARI flows directly into the D&A line on the income statement and into the intangibles balance on the balance sheet. It's a GAAP cost that doesn't appear in Adjusted EBITDA — meaning the \$$127$M step-up from partial-year FY2025 to full-year FY2026 affects GAAP EBIT comparability but not the Adjusted metrics. Any FY2026E EBIT comparison against FY2025 actuals that doesn't account for this step-up will draw the wrong conclusions about operating leverage.

## Supporting Schedules

Five sub-models underpin the balance sheet and cash flow mechanics.

The `Debt schedule` models each named tranche individually, calculates interest at approximately $3\%$ of beginning-of-period balance, and splits outstanding balance between current and long-term portions based on maturity. Total debt declines from \$$7.8$B to \$$2.8$B as tranches mature through the forecast period.

The `PP&E schedule` applies a standard roll-forward: beginning balance plus capex (modeled at $2\%$ of revenue, or \$$171$M-\$$292$M) minus depreciation ($55\%$ of prior-period net PP&E). Net PP&E grows modestly from \$$354$M to \$$486$M — consistent with Coinbase's asset-light model, where infrastructure investment is primarily software and cloud rather than physical assets.

The `Shares schedule` starts with a basic share count of $256$M and applies \$$1.5$B per year in buybacks at the prevailing share price. Diluted share count includes outstanding RSUs and options under the treasury stock method, declining from $287$M to approximately $259$M as buybacks more than offset new equity awards over the forecast period.

The `Working Capital schedule` derives accounts receivable from DSO assumptions (declining from $16$ to $12$ days), accounts payable from DPO assumptions (declining from $31$ to $14$ days), and accrued liabilities from $14$-$15\%$ of operating expenses. The large working capital jump from \$$41$M to \$$1,679$M in FY2026E reflects the Deribit balance sheet reclassification — a one-time integration step that drives the low FY2026E FCFF in the DCF. Not a structural signal.

The `Tax schedule` calculates the income tax provision as pretax income multiplied by the $21\%$ effective tax rate from FY2026E forward, consistent with Coinbase's disclosed forward ETR guidance. The deferred tax asset starts at \$$571$M and draws down at approximately \$$50$M per year as temporary timing differences reverse.

## Sources & Data Provenance

Historical financials are sourced from standardized financials (COIN US Equity) and confirmed against Coinbase's 10-K and 10-Q filings on SEC EDGAR. The Deribit PPA is an analyst estimate; official allocation has not been published. WACC inputs are sourced as follows: risk-free rate from FRED DGS10 (February 2026); equity risk premium from Damodaran's implied ERP (January 2026); beta from a five-year monthly OLS regression; cost of debt from Coinbase's publicly filed bond indentures. Macro assumptions are based on the FOMC dot-plot as of February 2026 for the Fed Funds Rate path and analyst estimates for crypto market cap. Peer multiples are sourced from consensus data as of March 2026.

## Key Risks & Sensitivity Levers

Seven variables drive the majority of valuation uncertainty in the model.

A $100$bps upward shift in the Fed Funds Rate adds approximately \$$350$-\$$450$M to annual USDC revenue relative to the base case path. A Circle renewal that renegotiates terms to $80\%$/$40\%$ removes approximately \$$220$-\$$300$M per year from the P&L — permanently, depending on the rate path at renewal. A $10$bps compression in the consumer take rate — through Coinbase One adoption, competitive pressure, or mix shift — reduces annual revenue by approximately \$$500$M at projected FY2028 volumes.

Terminal value represents $77$-$89\%$ of enterprise value. A ±$2\%$ change in the terminal growth rate moves implied per-share price by approximately \$$50$-\$$100$. A beta compression from $2.8$x to $2.0$x would reduce WACC to approximately $12\%$, adding \$$40$-\$$60$ per share to equity value.

Deribit integration execution and SBC dilution intensity are qualitative risks that don't translate cleanly to a point sensitivity but are flagged throughout the Analyst Notes tab.

## Disclaimer

This model is produced for educational and illustrative purposes only. It does not constitute investment advice, a recommendation, or an offer to buy or sell any security. All forward-looking estimates are my own opinion and are subject to significant uncertainty. The model has not been independently audited. Users should conduct their own due diligence and consult a qualified financial advisor before making any investment decision. Cryptocurrency markets are highly volatile and speculative in nature. I may hold positions in the securities discussed.

## Disclosure

The views expressed here are my own personal opinions. This is **not investment advice** and should not be relied upon as such. Nothing here is connected to, endorsed by, or written on behalf of Rosenblatt Securities.

---

Source: https://siddhants.com/blog/coin-research-202603/model-walkthrough/
About the author: https://siddhants.com/about.md
Site index for agents: https://siddhants.com/llms.txt