How it works — Preflop

01 · The covenant instrument

A covenant is a programmable income-share agreement deployed as a smart contract. An issuer receives capital upfront in exchange for a fixed percentage of future income over a defined period, subject to a payment cap.

The repayment at any period $t$ is:

R(t) = \min\bigl(s \cdot I(t),\; \frac{C \cdot P - \sum_{\tau=1}^{t-1} R(\tau)}{T - t + 1}\bigr)

where $s$ is the share rate, $I(t)$ is gross income in period $t$ , $C$ is the cap multiple, $P$ is the principal, and $T$ is the term in periods. The $\min$ ensures the issuer never pays more than their share rate ormore than what’s needed to exhaust the cap.

The five terms

Principal (P)

Capital raised upfront

$10K – $250K

Share rate (s)

% of gross income paid to all token holders combined

2% – 10%

Term (T)

Duration of obligation

24 – 120 months

Cap (C)

Maximum total repayment as multiple of P

1.5× – 5×

Floor (F)

Min annual income for payment to trigger

$30K – $60K

02 · Token economics — the constant supply model

Every covenant mints exactly $N = 1{,}000$ tokens at deployment. No additional tokens are ever created — this fixed supply is the foundation for market clearing and price discovery on the secondary market.

Token price at issuance

P_{\text{token}} = \frac{P}{N}

If an issuer raises $P = \$100{,}000$ , each token costs $\$100{,}000 / 1{,}000 = \$100$ at issuance. This is the entry price for backers.

Tokens received by backer

n_i = \frac{A_i}{P} \times N

A backer who allocates $A_i = \$15{,}000$ into a $\$100{,}000$ raise receives $(15{,}000 / 100{,}000) \times 1{,}000 = 150$ tokens, representing $15\%$ ownership of the covenant.

Dividend per token per period

d(t) = \frac{R(t)}{N}

If the issuer earns $I(t) = \$200{,}000$ and the share rate is $s = 5\%$ , total payment is $\$10{,}000$ . Each token receives $\$10{,}000 / 1{,}000 = \$10$ . A backer holding 150 tokens gets $150 \times \$10 = \$1{,}500$ .

Maximum return per token

\text{Max}(token) = \frac{C \cdot P}{N} = C \cdot P_{\text{token}}

With a $3\times$ cap, max return per $\$100$ token is $\$300$ . The cap bounds total payouts at $C \times P = \$300{,}000$ across all token holders. After that, the covenant terminates.

Critical distinction — share rate vs. ownership

The income share rate $s$ is a property of the covenant, not of any individual backer. A covenant with $s = 5\%$ means the issuer pays 5% of income to all token holders combined. Each backer’s claim is:

\text{Backer}_i\text{'s claim} = \frac{n_i}{N} \times s \times I(t)

A backer holding 150 out of 1,000 tokens receives $15\% \times 5\% = 0.75\%$ of the issuer’s income. Five backers holding 150 tokens each would collectively claim $75\% \times 5\% = 3.75\%$ — not $5 \times 5\% = 25\%$ .

03 · Present value framework — pricing the token

The fair value of a covenant token at any time $t$ is the expected present value of all remaining cash flows:

V(t) = \sum_{\tau=t+1}^{T} \frac{\mathbb{E}_t\bigl[d(\tau)\bigr]}{(1+r)^{\tau - t}}

where $d(\tau)$ = dividend per token in period $\tau$ , $r$ = risk-adjusted discount rate, $\mathbb{E}_t$ = expectation conditional on info at time $t$

At issuance ( $t = 0$ ), this simplifies to:

V_0 = \frac{1}{N}\sum_{t=1}^{T} \frac{\mathbb{E}\bigl[\min(s \cdot I(t),\; \text{cap remaining})\bigr]}{(1+r)^t}

The discount rate

The discount rate $r$ must account for four distinct risk premia:

r = r_f + \pi_{\text{income}} + \pi_{\text{liquidity}} + \pi_{\text{moral hazard}}

$r_f \approx 5\%$ — risk-free rate (time value of money)

$\pi_{\text{income}}$ — income uncertainty premium. High for pre-revenue founders ( $\sim 15\text{–}25\%$ ), lower for established creators with track records ( $\sim 5\text{–}10\%$ )

$\pi_{\text{liquidity}}$ — illiquidity discount. In Phase 1 tokens are non-tradeable ( $\sim 10\%$ ). Converges to $\sim 2\%$ as secondary market develops

$\pi_{\text{moral hazard}}$ — risk of income underreporting or career pivots ( $\sim 3\text{–}5\%$ )

Income trajectory modeling

Income is modeled as a stochastic process with regime changes — capturing the non-linear jumps typical of startup trajectories (the $\$0 \to \$3M$ raise event):

I(t) = I(0) \cdot \exp\bigl(\mu t + \sigma W(t)\bigr) \cdot R(t)

$I(0)$ = income at issuance (often $\$0$ for pre-revenue)

$\mu$ = drift rate (career growth trend)

$\sigma$ = volatility of income growth

$W(t)$ = Wiener process (random shocks)

$R(t) \in \{0, 1\}$ = regime indicator (0 = building, 1 = breakthrough). This is a Poisson jump process that captures the non-linearity

Early backers are essentially buying an option on the regime change $R(t): 0 \to 1$ . The expected value of this option drives the token price at issuance.

04 · Secondary market — price discovery and capital gains

Once the pricing engine assigns fair values, tokens trade. The constant supply of $N = 1{,}000$ tokens per covenant ensures well-defined market clearing. At any time $t$ , the token price updates as new information arrives:

V(t) = V(t-1) + \Delta V\bigl(\text{new info}\bigr)

Capital gain for early backer = $V(t) - V(0)$ per token

This is analogous to taking an individual public. The covenant is the S-1 filing. The token mint is the IPO. Income reports are quarterly earnings. The conviction score is the analyst rating. The secondary market is the exchange.

Information events that reprice tokens

Income report above expectation

↑moderate

Fundraising close (founders)

↑↑large

Viral growth event (creators)

↑↑large

Research publication / citation

↑moderate

Income report below expectation

↓moderate

Extended income floor breach

↓↓large

Career pivot away from high-earning path

↓↓large

Worked example — capital gains from a regime change

Issuance: Founder raises $P = \$100{,}000$ at $s = 5\%$ , cap $C = 3\times$ , 7-year term. Current income: $\$0$ . Token price: $P_{\text{token}} = \$100$ .

Month 8: Founder closes $\$3M$ Series A. Expected future income jumps. The pricing engine updates:

V(8) = \sum_{t=9}^{84} \frac{\mathbb{E}_8[d(t)]}{(1+r)^{t-8}} \approx \$340 \text{ per token}

Backer return: Backer who bought 150 tokens for $\$15{,}000$ at issuance now holds $150 \times \$340 = \$51{,}000$ in token value — a $3.4\times$ capital gain before any dividends.

Total return: Capital gains + accumulated dividends. The dividend stream continues flowing even as the token appreciates.

05 · Three pricing mechanisms — market maturity drives the model

Phase 1

Backer-determined

Backers set the effective price by choosing which covenants to fund. The platform doesn't price — early inefficiency creates alpha for informed backers.

— Backer demand (allocation speed)— Issuer conviction score— Comparable ISA transactions

Phase 2

Prediction market consensus

Prediction markets on key milestones (Series A? 1M subscribers?) generate probability estimates that feed into the valuation model — similar to Polymarket but for individual trajectories.

— Milestone probability markets— Peer-relative performance— Event-driven repricing

Phase 3

ML-driven DCF

With sufficient income data, ML models predict individual income paths — accounting for regime changes, sector dynamics, and macro factors. Each token gets a dynamic NAV.

— Historical income trajectories— Sector distributions— Comparable outcomes— Default rates

06 · Conviction scoring — the curation layer

Not every applicant lists. The conviction score determines who gets curated — the market prices after that. This is not a credit score (probability of default). It prices probability of outperformance:

\text{CS} = \sum_{i=1}^{6} w_i \cdot \text{Signal}_i

Weights are initialized by the founding team, then retrained via regression against actual income outcomes as data accumulates.

Signal weights (initial)

Optionality sacrifice

Offers turned down, opportunity cost

20%

Consistency

Days active on project, GitHub streak

20%

Traction signals

Revenue, users, followers, citations

20%

Network quality

Who already believes in them

15%

Problem importance

Domain scoring (AI safety > ad tech)

15%

Education quality

School tier, major relevance

10%

07 · The four phases — from marketplace to market

01Covenant

On-chain ISA contracts

Now — MVP

The legal and technical primitive. Each covenant is a smart contract encoding ISA terms, minting a fixed supply of 1,000 ERC-20 tokens, and enforcing payment logic including cap and floor.

— Term structuring UI— Smart contract (Base L2)— Fixed token supply (1,000)— Legal wrapper (Reg D)— Cap & floor enforcement— Automated payout

02Origination

Issuer × backer marketplace

Months 1–6

Two-sided marketplace. Issuers apply and get curated via conviction scoring. Backers browse, filter by industry/domain/stage, and fund covenants. Income data begins accumulating — this is the training set for Phase 3.

— YC-style application— Conviction scoring— Advanced backer filters— Plaid income verification— Pro-rata dividends— Issuer & backer dashboards

03Pricing

AI valuation engine

Months 6–12

With enough income trajectories from Phase 2, we train models that price individual human capital using DCF analysis, prediction market signals, and conviction scoring. Tokens become marked-to-market assets.

— AI-driven DCF per issuer— Prediction market signals— Dynamic NAV repricing— Risk & default modeling— Comparable transaction data— Institutional data licensing

04Market

Secondary trading & derivatives

Month 12+

Tokens trade peer-to-peer. Indices aggregate tokens by sector and vintage. Options price the probability of regime changes. Tranches structure risk for institutional buyers. This is the full human capital market.

— P2P secondary market— European call options— Sector & vintage indices— Tranched risk pools— Jump-diffusion options pricing— Institutional data feeds

08 · Derivatives — options, indices, and structured products

Standard Black-Scholes assumes log-normal prices. ISA tokens have jump discontinuities (regime changes). We use a jump-diffusion model:

C = C_{\text{BS}}(V, K, r, \sigma, T) + \lambda \cdot \mathbb{E}[\max(V_{\text{jump}} - K, 0)]

where $\lambda$ = jump intensity (Poisson rate of regime changes), $V_{\text{jump}}$ = post-jump token value

Call options

European calls on individual tokens. Priced using a jump-diffusion model that accounts for regime change probability.

Token at $100. Call at $150 strike, 12-month expiry. Founder closes Series A → token reprices to $340. Option pays $190.

Sector indices

Weighted baskets of tokens by domain. Institutional buyers take diversified positions on cohorts instead of individual bets.

Preflop Founder Index: top 20 founder tokens by conviction score. Rebalanced quarterly. One product, diversified human capital exposure.

Tranched pools

Senior and junior tranches on token pools. Senior gets first claim on dividends (lower risk). Junior absorbs defaults but captures outsized upside.

Pool of 50 tokens. Senior: 6% yield, 95% recovery. Junior: 15%+ yield, absorbs first 10% of defaults.

Taking individuals public.