How to accurately value cryptocurrencies?

This article aims to build a valuation framework suited to token characteristics.

Written by: Four Pillars

Translated by: AididaoJP, Foresight News

Key Points

• Tokens ≠ Equity. Valuations should use enterprise value / holder income, not enterprise value / protocol income.
• The accrual ratio (the proportion of protocol income ultimately received by holders) is a key diagnostic indicator. In the projects we compare, this ratio ranges from 25% to 100%.
• “Dilution” also differs. Team incentives are real operating costs (should be included in valuation multiples), while investor unlock sales are market events (should not be included).
• Treasury value depends on “extractability.” The issue isn’t “how much is in the treasury,” but “can holders actually access it?”

A common misconception in crypto valuation is pulling out a protocol with $500 million annualized revenue, dividing market cap by that number, and concluding it’s “cheap.” This approach is flawed because both the denominator and numerator are wrong. Investors think they’re buying at a 5x multiple, but considering what they can actually receive, the real multiple might be 20x.

Price-to-earnings ratio is a good starting point, but it ignores the balance sheet and capital structure—hence the use of enterprise value multiples (EV/EBITDA) in traditional finance. However, applying EV/EBITDA concepts to tokens encounters three fundamental issues:

• Treasury assets: Holders have no legal claim.
• Protocol income: Most may never reach holders.
• Largest costs: Not on the income statement but reflected as new token issuance.

This article aims to develop a valuation framework adapted to token features. The core metric is enterprise value / holder income—that is, the price paid per dollar of income that ultimately reaches you (as a token holder), considering balance sheet and operational costs. I will illustrate with five protocols (HYPE, PUMP, MAPLE, JUP, SKY). This is not investment advice, just a methodological demonstration.

1. How to Calculate a Token’s “Enterprise Value”?

A common initial mistake in token valuation is using market cap directly, but market cap does not equal enterprise value.

In traditional finance, the logic is clear:

Enterprise Value = Market Cap + Debt – Cash

Because if you buy the entire company, you assume its debt and take its cash. Subtracting cash is reasonable because that money legally belongs to you.

In crypto, things are more complex. From token burns (USDC inflow, tokens permanently destroyed, no one can claim that USDC) to foundation wallets (holding hundreds of millions, but without governance rights or distribution mechanisms), the situations vary. The key question isn’t “what’s in the treasury,” but “can holders access it?” (Of course, if someone acquires the entire protocol, the discount disappears, similar to traditional finance. The “claim rights discount” mainly applies to minority equity holders like us.)

I use the term “enterprise value” because the logic is similar: you’re calculating how much you need to pay to acquire the core business, excluding parts of the balance sheet that aren’t yours. The formula:

Token Enterprise Value = Market Cap + Token Debt – Extractable Treasury Assets

Most protocols currently have no “token debt,” so the focus is usually on treasury assets.

First, break down what’s in the treasury. A protocol’s treasury typically holds three types of assets:

• Stablecoins: real cash, theoretically fully extractable.
• Native tokens: the protocol’s own tokens. Subtracting this part is like “subtracting yourself,” usually requiring at least a 50% discount.
• Protocol-owned liquidity (POL) and other assets.

Total Treasury Assets = Stablecoins + Native Tokens × (1 – appropriate discount rate) + POL

But total assets ≠ extractable assets—that’s the core issue this framework addresses.

Some protocols even lack assets that can be discounted. For example, pure burn mechanisms (USDC inflow used for buybacks and burns) do not generate any balance sheet assets that can be claimed. In such cases, extractable treasury assets = 0, so enterprise value = market cap. This is the clearest scenario, requiring no subjective judgment.

For protocols holding actual assets, I introduce a “claim rights discount” framework, based on how much the holder can actually control, ranging from 0% to 100%:

• 0% discount: automatic buyback and burn, no governance vote needed; or funds are fully at the token holder’s discretion.
• 25% discount: active DAO with a history of actual distributions.
• 50% discount: governance rights exist but are only on paper, never truly exercised.
• 75% discount: treasury controlled by the team, weak governance.
• 100% discount: funds controlled by the foundation, no claim rights for holders.

These percentages are the most subjective and easily contested part of the framework—I admit that. But a debate between 25% and 50% is far more meaningful than both ignoring the treasury and only discussing P/E ratios.

Real-world examples:

• Maple: treasury holds $9.36 million (99.7% stablecoins), small amount. Enterprise value adjusts from $272 million to $265 million, minimal impact.
• SKY: treasury holds $140.3 million, but 99.9% is its own token. After applying a 50% discount, extractable value is ~$70.2 million, reducing enterprise value from $1.69 billion to $1.62 billion.
• PUMP: reports holding about $700 million in stablecoins, but no governance or distribution channels, so holders cannot access it. Extractable assets = 0, enterprise value = market cap.
• HYPE and JUP: similarly, pure burn or closed treasury, no judgment needed, enterprise value = market cap.

2. Income and Token Costs: How Much Can Actually Reach Me?

The gap between what the protocol earns and what holders actually receive is where most valuation frameworks fail—and it’s the key factor influencing valuation multiples.

Imagine income as a three-layer waterfall:

• Fees: total paid by users.
• Protocol income: the portion left after paying LPs, validators, etc.
• Holder income: the part that finally reaches token holders via buybacks, burns, or direct distribution.

Two key conversion rates:

• Retention rate = Protocol income ÷ Fees (how much protocol keeps from total fees)
• Accrual ratio = Holder income ÷ Protocol income (how much of the retained income actually reaches holders)

These ratios combined can lead to vastly different outcomes:

• HYPE: retention 89.6%, accrual 100%. Out of nearly $900 million in fees, $805.7 million ultimately flows to holders.
• Maple: retention 13% ($141.5 million fees → $1.83 million protocol income), accrual 25.1% ($1.83 million protocol income → $4.6 million holder income). Total pass-through only 3%, while HYPE is 90%.

Within the same framework, one project has a 3% ratio, another 90%. Using “EV / Fees” or even “EV / Protocol Income” as comparison metrics would be wildly misleading.

Why use “holder income” in the denominator instead of “protocol income”?

In traditional finance, EV/ revenue works because equity holders have residual claim rights—legally theirs. But token holders lack this right; they only get what the tokenomics design provides. If income is held in the team-controlled treasury with no distribution mechanism, just holding governance tokens doesn’t give you that income.

Using “protocol income” as the denominator can inflate the apparent valuation of protocols with low accrual ratios, creating an “accrual discount.”

For example, Maple:

• EV / Protocol Income = 14.5x
• EV / Holder Income = 57.7x

A fourfold difference! The same data, but based on different denominators, leads to vastly different market valuation judgments.

3. Costs: Dilution Comes in Three Flavors

The term “dilution” is overused in crypto; misclassification leads to valuation errors.

First category: Team incentives (equity incentives)—a real operating cost

Warren Buffett said decades ago: if incentives aren’t counted as costs, what are they? Gifts? In traditional finance, they show up on the income statement, reducing profits. In crypto, they manifest as new tokens entering the market, but the economic essence is the same—it’s a real business cost.

• HYPE: team incentives annualized at $464.9 million, consuming 57.7% of holder income.
• PUMP: team incentives annualized at $128.5 million.

These should be included in valuation multiples.

Second category: Operational token costs (ecosystem incentives, user acquisition, etc.)—also operational costs

They function like customer acquisition costs, representing real expenses, and should be included in multiples. PUMP, besides team incentives, has $77 million in operational token costs, totaling $205.5 million.

Judgment criterion: does it create new token supply?

If the protocol only distributes existing income to stakers without minting new tokens, the cost is already reflected in the cash flows (protocol income minus holder income). If the protocol mints or unlocks previously non-circulating tokens, that’s genuine dilution and a business cost.

Third category: Investor lock-up and unlock—market events, not operational costs

You wouldn’t subtract VC sales from Apple’s profits to get an “adjusted profit.” Similarly, this shouldn’t be included in operational multiples.

PUMP’s annualized potential sell pressure is $83.5 million, about 7.3% of market cap. While market impact is significant, it’s a market event, not a business cost. I place it in a separate diagnostic metric called “Total Token Holder Tax” (token costs + potential investor sell pressure as a percentage of holder income), but it’s not part of the core valuation multiple.

4. Four Core Multiples and One Diagnostic Indicator

Based on the above logic, we define the following metrics (here, the definitions are unified; later sections will directly reference):

• EV / Holder Income (core metric): how much you pay per dollar of income that ultimately reaches your pocket.
• Market Cap / Holder Income: same as above, but without treasury adjustments. The difference reflects balance sheet effects.
• EV / (Holder Income – Token Costs) (cost-adjusted multiple): deducts real business costs (team incentives, operational costs), excluding investor sell pressure.
• EV / Protocol Income (for reference): the gap with EV / Holder Income indicates the “accrual discount.”
• Total Token Holder Tax (diagnostic): = (Token Costs + Investor Sell Pressure) ÷ Holder Income. It reflects both business costs and supply pressure. For example, PUMP’s ratio is 60.3%, meaning for every dollar of income reaching holders, an additional $0.603 is injected via new supply. This number doesn’t directly indicate valuation level but signals cash flow and supply dynamics.

5. Data Overview and Case Highlights

• HYPE: accrual ratio 100%, 9.4x holder income. High team incentives push the cost-adjusted multiple to 22.2x. Revenue structure is clear; complexity isn’t on the income side.
• PUMP: appears cheapest (2.4x), accrual 98.8%. But treasury is inaccessible, with a large unlock scheduled for August 2026. Cost-adjusted multiple rises to 4.2x, with a total token holder tax of 60.3% (highest in sample).
• MAPLE: highest accrual discount (4x). Protocol income at 14.5x vs. holder income at 57.7x, a huge gap. No token costs, so cost-adjusted multiple remains unchanged.
• JUP: the cleanest balance sheet. “Net-zero emissions” governance, no token costs, no investor sell pressure, no extractable treasury. All multiples approach 7.7x.
• SKY: accrual ratio 45.8%, a prime example of how “denominator choice affects valuation.” Protocol income multiple is 7.3x (seemingly cheap), but holder income multiple is 16.0x (less cheap). Treasury is mostly (99.9%) its own token, worth discounted.

6. Conclusion

This framework has limitations:

• Claim rights discount on treasury is subjective: I use 25%, you might choose 50%, and neither can be definitively right.
• Determining whether new issuance occurs can be complex: some protocols have minting functions active but distribution channels dead, tokens stuck in unallocated pools, making assessment fuzzy.
• Data sources have noise: DeFiLlama’s 30-day annualized data can vary depending on snapshot timing, making the same protocol look cheap or expensive by a factor of two.

But it’s at least a practical starting point. EV / holder income, adjusted for balance sheet and real business costs, helps clarify how much of each dollar paid actually ends up in your pocket.

The gap between protocol earnings and what holders receive is the biggest fundamental mismatch in today’s market. Many protocols generate hundreds of millions in fees, but holders only get a fraction, and most valuation frameworks fail to distinguish this.

Fortunately, the industry is beginning to focus on value capture: fee switches are being turned on, buybacks are replacing inflationary staking, governance is voting to pause incentives. We are building tools to better measure what’s truly happening.

7. Data Sources and Methodology

Income data: DeFiLlama annualized figures (last 30 days × 12). Pros: more sensitive than half-year data; cons: single-month volatility can introduce noise.

Holder income: directly from DeFiLlama’s “holder income” field, including only buybacks, burns, and direct distributions.

Treasury data:

• MAPLE: $9.36 million (DeFiLlama, 99.7% stablecoins)
• SKY: $140.3 million (DeFiLlama, 99.9% own tokens)
• JUP: $0 (closed)
• PUMP: estimated $500 million median stablecoins (range $286 million–$800 million)

Token costs:

• MAPLE: $0. No staking distribution after MIP-019 (October 2025). Although a 5% inflation smart contract may still mint, no distribution channels exist. (Sources: docs.maple.finance, The Defiant 2025/10/31)
• SKY: $0. Savings module (STR) now distributes SPK and Chronicle Points, not SKY tokens. (Verified March 2026 at app.sky.money/rewards). The “6 billion SKY per year” figure mentioned by Rune in August 2024 is outdated, but governance can restart anytime. (Sources: sky.money FAQ, vote.sky.money)
• JUP: $0. The “net-zero emissions” proposal passed on Feb 22, 2026 (75% approval). DAO treasury is closed until 2027.

Investor sell pressure:

• PUMP: steady-state annualized potential sell pressure of $83.5 million. Actual cliff unlock begins August 2026; over the next 12 months, estimated actual sell pressure is about $48.7 million (assuming 7/12 months).

Lending protocol metrics:

• MAPLE: uses actual assets under management (AUM) ($3.79 billion, Q1 2026 report) instead of DeFiLlama’s TVL ($1.945 billion). Net interest margin (NIM) = protocol income / AUM. Detailed metrics in Excel appendix.

Cash operating expenses: not estimated, due to lack of disclosure, which could lead to inaccuracies.

Equity incentive valuation: based on current token price. Sensitive to price changes.