The Founder's Case for NUMA

I know how this reads — a founder arguing his own idea will work.
Let me try to be more useful than that.

I know how this looks. A founder writing about why his own idea deserves your attention. So let me try something more useful than a pitch.

Let me lay out — as clearly as I can, and with as much distance as I can honestly muster — why I think NUMA has a real chance of working, and where the honest doubts remain. Judge the argument, not the messenger.

Today, on Micro Connect's fifth anniversary, we are publishing two white papers. They describe a new market called NUMA — short for New Markets — designed to make an entirely new asset class investable: contract rights —the cash flows generated by real commercial contracts between real economic counterparties. Not a wrapper on debt. Not a token. Not a securitisation vehicle. Something that at first glance looks like a third asset class sitting between equity and debt — but is really, I think, a way of giving equity and debt themselves an extended life in a much larger new world. The closest analogy I can find is not disruption. It is closer to Columbus finding America. Europe did not need a new kind of trade after 1492. It needed the same trade, on a scale that had suddenly become possible.

1. White Paper on Micro Connect New Markets Contract-Based Financing (CBF) Protocol

2. Legal White Paper on NUMA Contract Rights Legal Framework

That claim is either interesting or absurd. I think it is interesting. Here is my case, in five parts.

1. Five years of trying and failing is humbling — but it gave us the compass to the future.

I ran Hong Kong Exchanges and Clearing from 2010 to 2020. During that time, we built Stock Connect and Bond Connect, restructured the listing regime, and made Hong Kong a genuinely two-way market between China and the world. I retired in 2020 to work on something that could not be built inside an exchange.

The five years since have been humbling. We raised US$600 million and invested directly in small Chinese businesses through Micro Connect 1.0. Covid interrupted us. Last-mile origination was harder than we thought. Most of the capital has come back; some is still on its way. In 2025 we came close to floating Micro Connect under Chapter 21 in Hong Kong, filed the A1, and in November pulled it — being both the fund and the operator of the market others would use was a conflict I could not clean up.

Two mistakes are worth special mention. First, we overestimated our own ability to be the investor in this market — to underwrite tens of thousands of small operators one contract at a time. That is not a job any single fund can do at scale. Second, we overestimated what digitalization alone could do without AI. Data without AI is just noise at scale.

But that tuition bought us something. What we are proposing today is what we have learned from painfully hitting those walls — exactly what this market needs, and exactly what it must never be.

2. This is not another money-making product. It is about a huge part of our economy finally being served by capital markets.

Global capital markets serve about 40,000 listed companies and perhaps another few hundred thousand credit-worthy corporates. Beneath them sits an ocean of economic activity — franchise operators, subscription businesses, charging stations, delivery fleets, small SaaS, AI-compute providers — that has real revenue, real contracts, and essentially no relationship with the capital markets at all.

This is not because anyone forgot. It is because the unit economics of understanding, pricing and matching millions of small non-standard contracts have never worked. Investment banks are not built for it. Rating agencies are not built for it. Even private credit — a US$1.7 trillion industry that already reaches further than banks — cannot go much further down the size curve before diligence costs collapse the returns.

And this is not a Chinese story. It is arguably the defining structural feature of any large emerging economy with a rising middle class and rapid urbanization— India, Southeast Asia, Latin America, the Middle East, and yes, mainland China too. Wherever that pattern shows up, the same gap shows up with it.
 

I should also be honest that if we overestimated our ability as an investor, we equally underestimated the sheer size of the world this market can reach. When we started, we thought of NUMA in mostly brick-and-mortar
terms. Five years later, the world we can actually see stretches to subscription businesses, the per-use billings of AI-compute providers, the event-driven cash flows of major sporting fixtures and live entertainment, and even to partial revenue-share arrangements against public equities. Almost anything that produces a real, traceable, structurable cash flow is, in principle, a citizen of this market. What we thought was a continent turned out to be an ocean.

If a market like this can be built, it does three things at once — it completes finance, gives allocators a genuinely uncorrelated return stream, and turns a decade of offshore-onshore policy direction into working infrastructure.

This is not a crypto-style claim about disrupting finance. It is closer to the opposite: it is a claim about completing finance. Wall Street does not need to jump. It does not need a new discipline, a new mental model, or a new set of instincts. Your debt is still your debt. Your equity is still your equity. What NUMA offers is simply somewhere new to practice the discipline you already have.
 

That is why I believe this market will fly — even if we ourselves do not end up being the ones in the cockpit piloting it.

3. The reason no one has done this before is not lack of ambition. It is lack of tooling. Two things changed recently, and I did not foresee either of them.

For most of the last thirty years, the case for a "contract market" was theoretical. Contract law had been around for two thousand years; the obstacle was never legal. The obstacle was that data on millions of small commercial contracts did not exist in a form anyone could work with, and the analytical capacity to work with it — if it had existed — did not exist either.
 

Two revolutions changed that.

The first was the quiet digitisation of the last mile. Over the past twenty years, payments, settlements and reconciliations across small businesses became structured data almost by accident, as a byproduct of the mobile payments ecosystem and the rise of platform economies. By the mid-2020s, contract performance across a very large number of small enterprises was, for the first time in history, actually observable.

The second was agentic AI. Structured data without analytical capacity is just noise at scale. What agentic AI provides — and what we could not have solved with any number of analysts, lawyers or accountants — is the ability to read, classify, price and monitor millions of contracts with the kind of continuous attention that a human portfolio manager could only apply to one.

I want to be honest here. I did not set out in 2021 to build an AI-native market. I did not know then that AI would be the missing piece. What we did have was five years of walking into walls — and every wall taught us something. The ideas we ended up with are, looking back, almost embarrassingly simple. They are the sort of ideas you only reach by being stupid enough, for long enough, to be pushed into the corner where they become obvious.

And every one of those hard-won ideas turns out to be impossible to execute without AI — not difficult, impossible. AI at the level we needed only really arrived since the beginning of 2026. We got the problem. Something else sent the solution, and it arrived just in time. The timing is largely luck. But the timing is now.
 

4. The design has real load-bearing structure, and that really counts.

As you will see in the white papers, there are a handful of pieces I would ask you to examine. Each one is the answer to a specific wall we walked into.

The Protocol Layer. NUMA is not an exchange, not a fund, and not a platform in the app-store sense. It is a protocol — closer in spirit to what SWIFT is to cross-border payments, or what ISDA is to derivatives, than to anything you would normally call a marketplace. Protocols do not require new legislation to exist; they require adoption. That means NUMA can be built inside the existing legal and regulatory perimeter rather than waiting for one to be written for it.

The Contract Structure. NUMA is not a market where a borrower arrives with a project seeking funds. Only a party who has already put value into a real commercial contract — already funded it, delivered under it, or otherwise carried the first-loss position — can list a claim on that contract for others to invest in. Originators cannot manufacture supply out of nothing. Skin comes before game. This is the single design choice that most protects the market from adverse selection, and it is deliberately non-negotiable.

The Discovery Engine. Millions of small contracts are not searchable by humans. NUMA plots every contract on a map of return versus risk, with each contract occupying a probabilistic region — more like a weather forecast than a rating. Investors do not pick contracts. They set a target region on the map, and the market assembles a portfolio for them. The coordinate system behind that map — what we call the ten-dimensional code — is a shared analytical grammar that gets richer the more the market uses it. The consequence is what matters: for the first time, small non-standard contracts become genuinely searchable, and diversification becomes something you actually get, not something you claim.

The Atomic Unit. Every note in this market is denominated at HKD 1,000. That is a structural choice, not a marketing one. At that unit size, nobody comes here to pick a single winner — they come to build portfolios, and the market is portfolio-driven by construction. Because the ledger and the code do work that lawyers and rating agencies used to do, a genuinely new kind of asset-backed structure becomes possible — one that small participants can reach, without the usual apparatus of special-purpose vehicles, arrangers and rating shops. Debt investors keep measuring what debt investors measure. Equity investors keep measuring what equity investors measure. Only now they can do it at a scale and granularity that traditional markets could not offer. Same discipline. New continent.

The Anchoring in Sharpe. The white papers spend an entire chapter on the Sharpe ratio, not on tokens or blockchains. That is deliberate. On the debt-like side of the market, NUMA measures risk as expected impairment against remaining exposure — a construct any structured credit analyst will recognise. On the equity-like side, it uses cross-sectional dispersion of realised outcomes — a construct academics have used for illiquid asset classes since the mid-2000s. Same Sharpe. Different denominators. Different terrain. Nothing here asks the reader to accept a new law of finance. The gravity is the same gravity.

5. Where does regulation fit here?

NUMA does not execute trades, does not custody assets, does not sell products, does not offer advice. Licensed brokers execute; licensed custodians hold; licensed asset managers advise. NUMA operates the pipes and the ledger. Contract law does the underlying work; securities law does the investor-facing work. We seek no regulatory exemption.

On the capital side, participation is limited to institutional and professional investors domiciled outside mainland China — a red line, drawn on purpose. On the asset side, contracts from the mainland real economy are welcome — routing offshore capital into the real economy, not the other way. There will be regulatory questions along the way. Almost certainly. I would be worried if a founder in my position told you otherwise.

I am not promising this will work quickly. New markets do not launch; they accrete. Stock Connect took years to become plumbing Hong Kong could not live without. Bond Connect the same. NUMA will follow the same slow-burn curve, and probably slower, because it is a genuinely new asset class rather than a new pipe between existing ones.

What I am really asking for is the fifteen minutes of diligence you would give any serious new proposition. And read the white paper.

To signal that we are aligned with the outcome and not just the launch, Micro Connect is committing roughly HKD 1 billion of its own balance sheet as aco-build fund — mostly as co-investment matching alongside first-mover institutional participants.

We are confident this market will work. It will work. But it will take a lot of work — and we invite you to be part of it.

What NUMA becomes, then, is not a Hong Kong experiment. It becomes the third leg of global finance, and the first genuinely new asset class in ageneration. You can choose to be among the first to walk in. You can choose to watch from the sidelines for a season. Either is reasonable. But I would ask you not to walk away thinking the question is whether this market will happen. The question is only when, and with whom.

Stocks connect. Bonds connect. Contracts should connect too.

Charles Li

Founder & Chairman, Micro Connect

3 August 2026 · Hong Kong