AI eliminated the cost barrier that locked out the largest untapped market in global finance. Seen Capital built the infrastructure that opens it. This document describes what happens to the capital that flows through it.
Every fund model starts with a single unit. Here is ours. An average investment of $900 that formalises a business, funds its growth, recovers its capital, and generates the next three investments — all without a human intermediary touching the process.
Average investment per portfolio company. Small enough to deploy thousands. Large enough to transform a nano-business.
Registers the business, secures permits, opens a bank account, files for tax status. The AI pipeline handles the paperwork in the local language, in 48 hours, through WhatsApp. This is the step that converts an informal earner into a visible, bankable entity. The grant is non-recoverable — it is the cost of opening the market.
Non-recoverable. This is the price of market access.Deployed as title-retained physical assets: smartphones for mobile commerce, solar panels for off-grid productivity, cargo bicycles for last-mile delivery. The business uses the asset. Seen Capital retains title until break-even. The asset is collateral, productivity tool, and ownership incentive — simultaneously.
Title-retained. Ownership transfers at break-even.The portfolio company pays 10% of total income until the $432 working capital is recovered. Not a loan — no interest, no compounding, no penalty. A revenue share aligned with the business's ability to pay. Formalisation lifts revenue 40–80%, so the 10% share comes from new income the business would not have had without the investment.
Break-even at ~20 months. Revenue starts from week one.At break-even, the portfolio company nominates three candidates from their community. They know who runs a real business. They know who will pay. They stake their own reputation on the referral. The chain carries the trust that institutional due diligence cannot replicate — and the sourcing cost approaches zero.
3 nominations per break-even. Trust-weighted pipeline.Traditional funds deploy capital once and wait for exit. This fund deploys capital, recovers it through revenue share, and redeploys it — continuously, without raising new LP commitments. The capital recycles. The portfolio compounds.
Capital locked for fund life. Returns realised at exit only. Growth requires new LP commitments. The J-curve is structural — years of negative returns before portfolio matures.
Working capital recovers through 10% revenue share. Formalisation uplift increases revenue 40–80%, accelerating recovery. Recovered capital redeploys without new fundraising. No J-curve — revenue begins from week one.
The formalisation grant is the cost of opening a market. The working capital is the engine that pays for itself. Together, they produce a capital cycle that traditional funds cannot replicate: deploy at $900, recover $432 through revenue share, redeploy the recovered capital into new investments. Each cycle widens the portfolio without diluting returns.
At Month 84 — seven years into a 500-company seed cohort — monthly revenue share from chain-generated portfolio companies reaches $87,000. That figure grows every month as new chain generations reach revenue-share maturity. It is income the fund earns without deploying additional capital.
The market is hundreds of millions of people. No fund can reach it through traditional sourcing. The chain is how the infrastructure scales to meet the market — through the trust networks of the people it has already served.
The seed cohort of 500 companies is sourced through NGO partners — organisations that have spent decades building trust with exactly the communities the fund serves. That costs $450,000 at $900 per investment. From that point, the chain takes over.
Each company that reaches break-even at ~20 months nominates three candidates. Attrition, timing delays, and disqualification reduce the effective conversion — but even with conservative decay, six chain generations produce a 9.3× portfolio multiplier. The 500 become 4,660.
Chain nominations carry something no institutional sourcing process can replicate: the trust of a woman who has been through the programme and is staking her reputation on the referral. That trust is why chain-sourced investments outperform cold-sourced ones. It is also why the marginal cost of sourcing approaches zero — the portfolio sources itself.
The population with the highest return potential is the population with the least financial history. Traditional credit scoring cannot reach them. The compound risk model was built specifically for this problem.
Strong behavioural signals and full collateral. Banks already serve this segment. Low returns, high competition. Not the opportunity.
Good repayment history but no collateral. Served by MFIs at high interest rates. Moderate returns, crowded market.
No financial history but strong asset coverage compensates. Title-retained physical assets enable first investment. The entry point for the excluded.
No financial record. No credit score. No banking relationship. But the compound model scores them on two independent axes — behavioural trust and asset coverage — and the substitution curves determine precisely how much asset backing compensates for a trust deficit. This is the most excluded population on earth. It is also the largest market in global finance. The model is trained continuously on portfolio data, improving with every investment.
Three tranches. Each tranche attracts a different class of capital with a different risk-return expectation. The first-loss tranche de-risks the impact tranche. The impact tranche de-risks the commercial tranche. Everyone is aligned. Everyone is protected by the tranche beneath them.
$500K–$1M. Accepts 0–2% return in exchange for first-loss protection. Catalytic capital — exists to de-risk the tranches above it and prove the model works. The Bill & Melinda Gates Foundation, USAID, DFID, and analogous institutions fund this tranche because it multiplies the impact of every dollar above it.
$3M–$5M. Targets 4–7% net IRR. Protected by the first-loss tranche beneath. These are family offices, endowments, and impact mandates that need measurable social returns alongside financial performance. The revenue-share model delivers both — every dollar returned is proof of a business growing.
$5M–$10M. Targets 10–15% net IRR. Available from Fund II onwards, once the track record from Fund I demonstrates risk-adjusted returns at scale. Protected by two tranches of subordinated capital. This is where the model reaches institutional scale — pension funds, sovereign wealth, commercial allocators.
2% management fee. 20% carry above hurdle. Seven-year fund life. Standard terms that institutional LPs expect, applied to an infrastructure that delivers non-standard access to the world's largest excluded market.
Traditional venture and PE funds show negative returns for years as capital deploys and portfolios mature. The revenue-share model begins generating cash flow from the first month of operations. Capital recovers and recycles. The fund never sits in a valley waiting for exits that may not come.
Fund I proves the model. Fund II scales it. Fund III institutionalises it. Platform licensing begins in Year 3 and creates a revenue stream independent of AUM.
500 companies across 6 geographies. First chain generation validated. Compound risk model calibrated on real data. Revenue-share economics demonstrated. NGO network operational. Platform licensing piloted with 1–2 DFI clients.
2,000–5,000 companies across 15+ geographies. Commercial tranche opens. Chain mechanism operating across multiple generations. Platform licensing at $500K/client/year with 3–5 DFI clients. Risk model trained on 10,000+ data points.
Chain mechanism generating deal flow that exceeds capital availability. Multiple chain generations validated across geographies. Platform licensing generating $2.5M+ recurring revenue. Total revenue across all streams reaches $14.9M by Year 5.
Platform licensing is the strategic inflection. It transforms the business from a fund manager into an infrastructure provider. A DFI paying $500K per year to run their own portfolios through the AI pipeline is buying technology — not fund exposure. That revenue is valued at technology multiples regardless of how the fund itself is categorised.
The pipeline removes the cost floor. Seven AI agents take a woman from first WhatsApp contact to funded portfolio company in 48 hours, in any language, at a marginal cost that makes $900 investments viable. No human intermediary. No local office. No cost barrier.
The fund structure removes the capital ceiling. Three tranches — first-loss, impact, commercial — each de-risking the one above. Blended capital attracts institutional allocators who cannot touch the market alone. The structure scales from $10M to $420M without changing the model.
The chain removes the sourcing constraint. Every investment that reaches break-even generates three peer-nominated candidates. The portfolio sources itself. Six chain generations multiply 500 seed investments into 4,660. The marginal cost of origination approaches zero.
The data removes the risk perception. The compound risk model — two independent axes, continuously trained on empirical portfolio data — scores candidates with zero financial history. It converts the most excluded population on earth from an unquantifiable risk into a measured, managed, investable opportunity.
What remains is the largest untapped market in global finance — and the only infrastructure that can reach it.
The pipeline removes the cost floor. The fund structure removes the capital ceiling. The chain removes the sourcing constraint. The data removes the risk perception. What remains is the largest untapped market in global finance — and the only infrastructure that can reach it.