In a rare moment of candor, Li Xiaojia, founder of Micro Connect, addressed his investors at the company's fifth-anniversary celebration: "I'm sick and tired of talking to you apologizing. I like to come back and say I'll make you proud. You will be proud, shareholders, I promise that." Micro Connect, a fintech company known for its "Li Xiaojia" connection, operates like an exchange, linking capital with small-scale economic contracts by packaging them into investable products—a relatively novel market model.
At its inception five years ago, the company considered naming itself "New Market," a homophone for "niu ma" (meaning "bull code"). Now, in its fifth year, the brand name "NUMA, New Markets" has returned to the core strategy, becoming the flagship product of Micro Connect. Few entrepreneurs are so open about the difficulties of their entrepreneurial journey, but Li Xiaojia, in a blog post reflecting on Micro Connect's five-year path, wrote honestly: "That period was quite dark, with no hope in sight."
Li Xiaojia divides Micro Connect's history into three phases: 1.0, 2.0, and 3.0. The "dark, hopeless" period he describes likely corresponds to the early 1.0 stage. In the early 1.0 days, aside from initial skepticism comparing it to a "P2P" model, Micro Connect seemed to be on a smooth path. Li Xiaojia quickly raised billions of dollars from well-known Wall Street institutions and overseas family offices. Because Micro Connect's shareholders are all professional investors, Li was puzzled by the "P2P" controversy. "We didn't raise a single cent in Mainland China. The 4.4 billion Hong Kong dollars we invested were all our own overseas funds, my shareholders' money, only offered to professional investors," Li said. "Capital within China is strictly regulated. Our overseas institutions should never consider raising funds in China. Our idea was to bring 'big money into small shops'—overseas big money into China's small shops. Of course, overseas institutions aren't doing charity; they seek returns on investment."
Li Xiaojia described his "big money into small shops" concept as "a very simple idea." "After 30 years on Wall Street, a question always troubled me: With such a good market and so much money, why can't we help our vast number of small and micro businesses? They have genuine 'money-making opportunities' and need financing. If capital can find this new blue ocean, why not?" This model was quickly validated in Micro Connect's early days. An overseas model, Revenue-Based Financing (RBF), shares a similar logic. SaaS companies like the US-based Pipe and Canada's Clearco operate on this model, positioning themselves as "cash flow asset trading markets." However, the key difference is that they don't invest in physical stores. Micro Connect's model of investing in physical stores worked because there was a genuine gap in the market for RBF-style financing for small businesses, and the "non-equity, non-debt" revenue-sharing contracts matched the cash flow characteristics of small shops. More importantly, with the widespread adoption of mobile payments and digital technologies like blockchain in China, making store cash flows visible and enabling automated revenue sharing became operationally feasible.
During the 1.0 exploration phase, Micro Connect also realized its true target wasn't "small businesses" but "small economic contracts." "A 'contract' can be a small contract from a large company or a large contract from a small company. It's the contract from all economic activities in society, not just small businesses." With a clear direction and an expanded funding base, Li Xiaojia deemed the financial model a success. "This product absolutely should be made. The whole world should consider it," he said, but he also admitted they "hit a wall." "The team severely underestimated the market's size and overestimated the dividends from digitalization." Li believed that the technology at the time was insufficient to handle massive data volumes and provide investors with intelligent, customized decision-making tools. Digitalization could only make small shop information transparent, but it couldn't help investors assess risks or form real investment decisions. This meant Micro Connect's products were popular with financing parties but couldn't be effectively sold to investors. Micro Connect's expansion seemed to stall. Li Xiaojia said, "We didn't know what tool we were waiting for to meet the technical conditions."
Consequently, Micro Connect entered its 2.0 "payback period," suspending new external investments. The sustainability of Micro Connect's model heavily relies on the operational health of its businesses. If the economy slows down and stores close, the overall return on the contract pool declines. During the 2.0 period, the entire market investment environment underwent a massive shift, increasing the uncertainty of business repayment cycles and putting immense pressure on Micro Connect's expansion. Li Xiaojia described it as, "Even if we struggled, it was hard to make a big breakthrough, so we focused on recovering the money we had already invested and reducing costs." Micro Connect thus entered a quiet waiting period.
During this wait, Micro Connect explored another path: going public through a Chapter 21 listing. "We wanted to create a fund where we could tell investors how to invest, and then Micro Connect would manage that fund, but we ultimately gave up on the idea," Li said. In discussions with regulators, the team gradually realized their true role was to build an intermediary market connecting investors and financing parties, carrying an "exchange gene." If they also participated in direct investments, it could create conflicts. By the end of 2025, Micro Connect paused its listing plans and continued to wait for a technological breakthrough. Li described the wait as "extremely painful, with the path ahead full of uncertainty." The AI boom had begun, but tools were still limited to conversational interactions, unable to execute real business tasks. The Micro Connect team tested various AI tools, but none met their expectations. This changed with the arrival of Agentic AI in early 2026. Li Xiaojia said, "It felt like a sudden savior arrived, ready to lead us out of the tunnel."
Li Xiaojia wrote in his blog that when the AI tool finally worked, his first reaction wasn't "we're so smart," but "we're so lucky." Micro Connect had entered its 3.0 era. On August 3rd, Micro Connect held a press conference to launch its core product, NUMA (New Markets). They released two white papers: "The NUMA White Paper on Real-World Contract Financing" and "The NUMA Legal White Paper on Contract Rights," detailing the company's operational logic. The conference's theme was "Stocks are connected, bonds are connected, now contracts should be connected too." After achieving cross-connectivity in traditional financial markets, Li Xiaojia seems confident in replicating this experience in the "world of contracts."
Micro Connect's official business description is "a financing method based on real-world contracts—Contract-Based Financing (CBF)." This differs from the already-validated RBF (Revenue-Based Financing) model, though both are "non-equity, non-debt" equity models. According to Micro Connect, NUMA aims to establish a third equity asset class alongside traditional equity and debt. "It allows every contract in the real world to be discovered, understood, and invested in, opening a new path for capital to empower the real economy," Li said. "We are now in the execution phase. It will succeed, I just don't know when or how fast. Whether it will be completed in our lifetime is uncertain, but I have no doubt about its success." When asked why he is so confident, Li responded, "The market is too big to ignore anymore."
The economic landscape is filled with numerous contracts, and the variety of assets that can be traded is vast. "Examples include charging piles, photovoltaic panels, performance projects, discounted equity of unlisted companies, commissions from various platforms, e-commerce inventory, and creditor claims from small shop leases in shopping malls. As long as funds, labor, or materials are invested to create future revenue rights, they can be financed and circulated." "However, the factors influencing each contract are completely different, making standardization difficult. This requires powerful AI tools to process the data." In Micro Connect's conceptual world, the investor is the driver, facing ten instrument panels. Micro Connect calls these ten panels the "Ten-Dimensional Code," which helps investors quantify and evaluate asset portfolios. In this logic, humans define the coordinate axes for the ten panels and set scoring rules from 1 to 9. Investors adjust the weights and set constraints. The AI is responsible for scanning and parsing the original contract data, outputting the coordinate data for the Ten-Dimensional Code, and performing massive portfolio simulations.
For example, the first panel is the business's own creditworthiness. The X-axis is business scale, and the Y-axis is the counterparty's credit level. A coordinate of (5,5) represents a medium-sized business with medium credit. An investor can set a rule not to accept targets below (5,5), requiring lower-scoring targets to compensate with higher returns. The second panel is the cash flow characteristics of the business. The X-axis is the cash flow frequency (1 being very infrequent, 9 being very frequent), and the Y-axis is the business lifecycle (1 being a short duration, 9 being a long duration). A coordinate of (1,1) represents a low-frequency, one-time business, like a single concert, characterized by high risk and potentially high returns. A coordinate of (9,9) represents a high-frequency, long-cycle project, like a chain of stores. The third panel is the degree of digital performance. The X-axis is the "degree of digitalization of the real-world business," and the Y-axis is the "degree of digitalization of contract performance." A coordinate of (9,9) might involve settlement via AI Token computing power, with direct usage-based settlement and minimal disputes. A coordinate of (1,1) represents a purely offline, simple contract with no digital performance records, carrying a high risk of disputes. Li Xiaojia explains that the Ten-Dimensional Code coordinates are calculated using a weighted average of the weights set by humans, outputting two standards: the contract's expected return and its risk. This is mapped onto a two-dimensional plane, with the horizontal axis representing expected return (higher to the right) and the vertical axis representing risk (higher to the top). This creates four quadrants: the Southeast quadrant (high return, low risk) is the investor's preferred target and will be prioritized; the Northwest quadrant (low return, high risk); the Northeast quadrant (high return, high risk); and the Southwest quadrant (low return, low risk). Investors can manually adjust the weights of dimensions, such as increasing the weight for digitalization, or set hard thresholds, like not allowing a business's credit quality to fall below a certain coordinate. Of course, a single contract might fall into a high-risk zone, but investors can bundle multiple contracts into a portfolio for risk hedging. Investors can also express their risk preferences, for example, "shift the overall portfolio slightly towards the high-return, high-risk Northeast quadrant," or "allocate a certain number of portfolios to the Southeast, Northeast, and Southwest quadrants."
With AI as a quantitative tool for financial products, is the contract pool Micro Connect has accumulated over five years sufficient? Li Xiaojia believes the contract pool doesn't need to be massive from the start, but it must be sufficiently diverse. "In theory, 100 contracts can run the entire system, but they must meet the diverse risk preferences of investors. As long as contract assets are continuously produced and investors enter the market simultaneously, the two can catalyze each other, and the market will grow steadily, step by step."
An investor present at the Micro Connect press conference admitted they had been anxious for years without seeing a return on their investment. However, they said they were willing to trust Li Xiaojia because of his directness, efficiency, and openness. "The era has brought us AI, and we understand the Micro Connect model under this AI tool. We can wait a little longer." Li Xiaojia said, "Without the accumulation and lessons learned over the past five years, especially the pitfalls we stumbled into, it would be impossible to build the foundation of Micro Connect's products as it is today. Without AI, the many things we want to do wouldn't be possible at all."
Zhu Xi, a philosopher who lived in seclusion on Wuyi Mountain for fifty years, once wrote: "Last night, the spring water rose in the river; the huge ship floated as lightly as a feather. All the effort of pushing it in vain before; now it sails freely in the current." He wrote "The Collected Commentaries on the Four Books" and other works on Neo-Confucianism during this period. In ancient times, when time moved slowly, if Zhu Xi had had AI, would he have needed fifty years to write? In an era where AI is becoming increasingly practical, perhaps from its fifth year onwards, Micro Connect and Li Xiaojia can also experience the smooth sailing that comes with the rising tide, going with the flow.
Comments