The Second Half of Tokenization: Who is Capitalizing on the 'American Institutional Dividend' through Stablecoins?

Taking three startups as examples, we explore the real paths for commercializing RWA based on the dollar system.

Written by: Joel John

Compiled by: Saoirse, Foresight News

This article was inspired by an analysis note from Marc of OMVC. Marc pointed out that stablecoins only address the financial scheduling layer but cannot solve the more fundamental and tricky core issue of dollar scarcity; to inject dollar liquidity into the global market, we cannot rely solely on stablecoins. This article will analyze based on his views and introduce a group of startups building a global financial scheduling layer on blockchain infrastructure. My core judgment is that blockchain infrastructure will become the carrier for the export of American institutions and assets to the world.

The book "The Magnificent Trade" opens with a historical account recorded by Herodotus: Carthaginian merchants once engaged in silent trade with unknown tribes on the west coast of Africa. When Carthaginian ships arrived, they unloaded their goods on the shore and then returned to the ship to light a signal fire; the locals, seeing the signal, would come ashore to inspect the goods and place gold next to them. If the Carthaginian merchants thought the gold price was too low, they would stay on the ship waiting for the other party to add more gold, negotiating back and forth until both parties agreed on the transaction price. Throughout the process, there was no verbal communication, yet they had a universal trading communication system.

This story took place 2,500 years ago, long before the World Bank, SWIFT clearing system, blockchain, and even most mature civilizations. Over two millennia, the essence of commerce has never changed: humanity has always been exploring a universal communication standard to finalize fair transaction prices. In the context of modern markets, the so-called price is the interest rate, exchange costs, or the final payment amount in dollars. All commercial activities essentially build a unified, interoperable value common language.

The World Bank and the International Monetary Fund have established a universal language for fiscal exchanges between sovereign nations; SWIFT has completed the communication standards for cross-border interbank transactions; Visa, Mastercard, and the U.S. Depository Trust & Clearing Corporation (DTCC) have extended this set of rules to payments, credit cards, and clearing and settlement fields. For over seventy years, global trade has relied on this mature institutional system to establish trust between trading parties who do not know each other—after all, deep economic ties are an important bond for maintaining international peace.

If a country's economy heavily relies on trade with another country, the probability of conflict between the two parties will significantly decrease; deep economic integration often effectively alleviates geopolitical tensions. However, the trust built by this traditional institutional system has obvious shortcomings: trust relies on the legislation of various countries, bank cooperation, and offline business processes, which cannot be verified in real-time through unified technology nor enforced through standardized technology.

A globally unified ledger, verified in real-time by all trading parties, may provide a better solution.

Visa, Mastercard, DTCC, and even SWIFT have already realized this. Various industries are embracing different blockchain technologies, with the core demand being to create a universal value language suitable for global-scale trade. We have consistently used a set of concise core arguments internally: blockchain for capital market assets is like the internet for information. The marginal cost of acquiring assets will approach zero, verification efficiency will significantly improve, and the world will eventually merge into a single capital market.

Asset tokenization is essentially creating a value language for dialogue among global economies, but the asset value carried by this language entirely depends on the institutional strength behind it. Mature economies rely on sound institutions to adjust market risks through interest rate tools and balance public interests through regulation; the United States, with its well-established institutional system, can efficiently complete technological research and development, export to foreign markets, and realize capital. Without the mature capital market and supporting institutional environment of the U.S., ecosystems like Silicon Valley would find it difficult to replicate in other regions. The vast majority of startups globally choose to register in Delaware precisely because it has abundant capital and a complete institutional framework suitable for entrepreneurship.

The dollar is the medium through which countries share the American institutional dividend; stablecoins further export the institutional stability enjoyed domestically in the U.S. overseas; asset tokenization is the standardized rule for implementing this universal value language. Various market data can also support this logic.

Currently, the total circulation of global stablecoins is approximately $315 billion. Just Tether alone directly and indirectly holds U.S. Treasury bonds worth $141 billion, ranking 17th among global holders of U.S. debt, surpassing South Korea and the UAE. In February 2026, the total monthly settlement amount of stablecoins reached $7.2 trillion, surpassing the U.S. Automated Clearing House (ACH) network for the first time. The global economy is undergoing a large-scale on-chain migration, with tokenized stocks, credit, and asset custody vaults being the core carriers of this transformation.

As we mentioned in previous articles, blockchain can upgrade financial tools into a complete open platform. This means that global developers can build products that issue assets through various financial platforms to serve end users:

• Robinhood: Provides trading channels for tokenized stocks and real-world assets (RWA);
• Centrifuge: Launches AAA-rated tokenized credit through products like JAAA;
• BlackRock BUIDL: Tokenizes money market funds;
• Ondo: Achieves on-chain issuance and trading of public securities;
• Apollo and Hamilton Lane jointly with Securitize: Moves private credit and private equity funds onto the blockchain;
• Superstate: Helps listed companies complete on-chain issuance and circulation of stocks;
• Even Blockchain Capital's venture fund shares have been tokenized for on-chain subscription.

However, ordinary users will not directly engage with these complex financial products; the dollar is the core entry point for emerging market users. Exporters, freelancers, and all groups earning dollars in emerging markets will soon discover that the dollar has three highly attractive attributes:

• Compared to local currencies, the dollar has anti-inflation properties, and long-term holding can preserve and increase value;
• Can access global capital markets such as Hyperliquid, tokenized stocks, and various alternative assets;
• Extremely liquid, allowing for one-click cross-border transfers to any region in the world.

In many emerging markets, on-chain dollar assets can even command a premium; during market pressure phases, the premium can reach 10% to 15%. Companies that can accommodate regional capital flows and standardize them into the global ledger system will possess extremely high long-term value. These companies integrate local market rules, allowing regional assets to be understood and recognized by global capital.

This is highly similar to the trajectory of internet development: Google Maps standardized global geographic information, and social information flows built a global cultural communication carrier; now, the new generation of financial products is standardizing the sources and circulation paths of assets.

Currently, several startups are entering from different tracks to implement this standardized system.

Cross-Border Payment Infrastructure

Import and export are the most frequent economic exchanges between countries, and SWIFT serves as the communication layer for these transactions. However, the clearing speed varies significantly across regions. For example, Nigerian importers purchasing goods from China are often required by Chinese suppliers to prepay 60% of the dollar amount.

However, emerging market companies often struggle to obtain sufficient dollar reserves—countries strictly control the outflow of dollars to maintain local currency foreign exchange reserves. Even if Nigerian importers connect with compliant banks and have sufficient funds in their accounts, a dollar wire transfer to China can take 7 to 10 days to clear, and most small and medium-sized merchants do not have the full set of compliance qualifications.

At the same time, if Chinese suppliers directly receive payments in stablecoins, they would lose up to 13% in export tax rebates. Although both parties have genuine trade needs, the existing payment channels cannot match smoothly: on one hand, the cross-border documentation process is cumbersome, and on the other hand, fraud risks are difficult to control; once disputes arise, the costs of cross-border judicial recovery are extremely high.

Keyrails is not positioned as a lending institution but as the payment and credit scheduling layer for this type of cross-border trade. It does not use its own funds for lending but acts as a clearing intermediary, connecting importers with various external funding sources (non-bank financial institutions, fintech companies, and recently integrated on-chain asset vaults), controlling the payment flow of borrowed funds throughout the process. Lenders can expect an annual return of about 15% to 20%, while borrowing companies face an annual financing cost of 20% to 25%. Keyrails earns a spread of 2.5% to 5% on top of payment channel fees.

It is important to distinguish: the 20% to 25% here refers to the annualized borrowing interest rate, not the exchange premium of local currency against the dollar, which is the alternative option that importers originally had to choose.

Nigerian importers exchange naira for USDT through Keyrails, matching funding sources, and directly pay suppliers in dollars via compliant SWIFT, balancing on-chain stablecoins with traditional cross-border settlement to secure export tax rebates.

Business Process: Nigerian merchants exchange naira for USDT through local OTC trading platforms, and funds are transferred to Keyrails; the platform directly pays the Chinese suppliers through its own SWIFT channel. Lenders can access the complete transaction history of merchants for the past three months via API, and credit review can be completed in just three hours. Funds do not pass through the borrower's account but are directly matched with the supplier's genuine invoice, completing settlement within China through SWIFT.

The core logic that enables this model to work: merchants are not comparing with low-interest bank loans (which emerging markets cannot access), but with the high exchange premiums in parallel foreign exchange markets. During market pressure periods, traditional channels can see exchange premiums as high as 20% to 30%, with clearing taking 7 to 10 days; whereas Keyrails' annualized financing products at 20% to 25% can complete settlement in 6 to 8 hours, making the short-term borrowing comprehensive cost lower. For a three-month term, the annualized interest rate of 20% to 25% translates to an interest of only about 1% (excluding fees and collateral discounts), far lower than a one-time payment of 20% to 30% exchange premium.

For Chinese suppliers, the advantages are equally clear: dollar funds are compliant and can be credited to corporate accounts, allowing for normal export tax rebate claims, which cannot be achieved by directly receiving stablecoins; the core pain point is the severe lack of liquidity in local currency to dollar transactions in emerging markets.

Keyrails' core competitiveness does not lie in having the lowest funding costs but in locking credit funds and controlling risks through exclusive payment channels. It digitally collects global trade documents, builds cross-regional payment links, and restricts borrowed funds to be used only for purchases corresponding to genuine invoices. Its moat is built on three main elements: complete credit data, compliant clearing channels, and dedicated fund management. Each transaction enhances the platform's database, continuously improving the standardization capability of regional trade funds connecting to global dollar capital.

Programmable Credit

Keyrails focuses on the rapid circulation of funds in informal markets, while Semiliquid builds collateral sharing infrastructure to serve familiar institutional trading counterparts.

The underlying logic of both companies is interconnected: full-cycle control of funds/assets. Within the Keyrails system, funds can only flow out through designated compliant channels; within the Semiliquid system, collateral does not need to leave the custody account at any point.

Asset tokenization merely changes the digital representation of assets and does not automatically confer collateral financing attributes. If banks, funds, or traders hold tokenized U.S. Treasury bonds, money market funds, stocks, or credit products in a custodial institution, only those that support collateralized lending can fully unlock the value of the assets. Without guaranteed financing functions, tokenized assets are merely digital shells of their underlying physical assets; only by adding credit functions can they connect to a complete financial circulation system.

Semiliquid's Programmable Credit Protocol (PCP) allows both borrowing and lending parties to complete token asset financing without transferring collateral out of the custody account. Borrowers retain their assets in custody, continuously earning returns on the underlying assets; lenders obtain legally enforceable collateralized debt rights. Once the borrower repays, the collateral lock is automatically released; in the event of default, lenders can dispose of the collateral according to pre-agreed rules. The core mechanism can be summarized as "cash transfer, collateral lock": funds can flow freely, collateral remains in the custody account, and ownership only changes during default disposal.

Institutional borrowers' tokenized U.S. Treasury bonds remain in custody, only locked for collateral to finance lenders, while asset returns still belong to the borrower, and lenders obtain legally binding collateralized debt rights.

We can intuitively understand the value through a set of calculations: assuming an institution holds $100 million in tokenized U.S. Treasury bonds with an annual yield of 5%; at a 98% collateralization rate, it can borrow $98 million in liquidity, with an annual borrowing interest rate of 6%. The annual interest cost of the loan is $5.88 million, but the U.S. Treasury bonds can generate $5 million in interest each year, leaving the borrower with a net cost of only $880,000, corresponding to an annual net cost of about 0.9% on the $98 million loan (excluding protocol fees, asset depreciation, and custody fees).

Borrowers do not simply bear a 6% borrowing cost; what they actually pay is the difference between the borrowing rate and the collateral returns. In traditional banking systems, the interest generated by collateralized assets is often intercepted by banks and custodial institutions; however, under the programmable collateral system, even if the assets are in a collateral lock state, the returns still fully belong to the borrower. Semiliquid opens this mechanism to all market participants holding tokenized assets with borrowing needs.

It fundamentally differs from decentralized lending pools like Aave: institutions do not need to transfer assets into public smart contract liquidity pools, avoiding the high-risk premiums associated with unpermissioned lending, and do not have to accept public liquidation rules; assets can continue to be held in regulated custodial institutions while achieving collateralized financing. For lenders, the due diligence process is significantly simplified, and asset disposal rights are clear and controllable; custodial institutions can verify the status of collateral in real-time, and the underlying protocol eliminates the risk of double pledging the same asset.

The 2022 crises involving Three Arrows Capital and Archegos Capital clearly illustrate the necessity of this mechanism. Three Arrows Capital left creditors with approximately $3.5 billion in bad debts; Archegos' family office had overlapping swap exposures with multiple investment banks, resulting in a loss of $5.5 billion for Credit Suisse alone. The core issue in both crises was not the decline in asset prices, but the inability of lenders to share complete collateral data in real-time: they could not confirm the location and pledge status of the assets, and the repayment capacity of the same asset was repeatedly disclosed to multiple trading counterparts.

Semiliquid fills the trust gap in traditional finance with real-time collateral verification. Borrowing data does not need to be publicly disclosed to the entire market, but all relevant parties can verify in real-time the actual existence of collateral, its locked status, and the right to dispose of it.

Traditional secured lending processes require coordination among lawyers, back-office operations, and custodial institutions, making the process lengthy for a single financing project; Semiliquid compresses the entire process using programmable infrastructure. More critically, it revitalizes idle token assets, supporting margin trading and repo financing businesses. The next phase of competition in the tokenization industry will not be about putting more assets on-chain, but about enabling on-chain assets to possess practical financial attributes: institutions can pledge and borrow without relinquishing custody rights, earning additional returns.

AI Computing Asset Financialization

GPUs are the core production materials of the artificial intelligence industry, supporting all large model computations. At the same time, the procurement costs of GPUs are extremely high: Meta, Amazon, Alphabet, and Microsoft are projected to spend a total of $410 billion on computing capital expenditures by 2025, with plans to invest $725 billion in 2026, a year-on-year increase of 77%.

A simple analogy: in an agriculture-dominated economy, tractors are the core production tools; in today's AI era, GPUs are the production devices that can convert capital investment into continuous cash flow.

Computing power companies generally procure GPUs through credit, relying on renting out computing power to generate revenue to repay loans. However, small and medium-sized data centers and AI infrastructure service providers cannot obtain low-cost credit like cloud vendors such as Google, Amazon, and Microsoft; GPU pricing, efficiency, and residual value fluctuate with the iteration of large models and ongoing changes in enterprise demand, making it difficult for traditional financial institutions to accurately assess loan risks.

As of the time of writing this article, the total assets locked in the USD.AI protocol amount to $398 million, with an active lending scale of $202 million. The scale of single loans has grown from $1-5 million at launch to $98 million per loan, corresponding to a large credit line for a cluster of 2,304 GPUs, with the collateral asset model having moved beyond pilot projects to achieve large-scale implementation.

USD.AI allows stablecoin lenders to provide funds, while computing power operators use GPU devices as collateral for loans, relying on GPU leasing income to repay loans, creating a source of real computing power returns for stablecoin supply.

USD.AI opens a new avenue for lenders, sharing the dividends brought by the expansion of cloud vendors and the growth of AI computing power demand; computing power companies can financialize their own physical devices without relying on traditional credit channels, quickly obtaining liquidity. Both parties can rely on this system to safeguard their rights through collateral disposal in the event of operational risks with the equipment.

Core Industry Conclusion: Value Lies in Scenarios and Industry Data

These types of companies are fundamentally different from traditional DeFi products. They rely on tokenization, blockchain channels, and stablecoins to serve the inadequacies and scattered processes of traditional finance in the physical market. Their moat is not just the underlying code, but the industry scenario data accumulated from each transaction. Semiliquid's programmable credit, USD.AI's computing power lending, and Keyrails' cross-border trade financing share common advantages: with each completed transaction, the platform accumulates a layer of trust and industry data that is difficult to replicate, while continuously optimizing its risk control and business processes.

The long-term value of enterprises comes not only from transaction flows but also from the accumulated user, borrower, and trading counterpart scenario data across all dimensions, which is logically very similar to traditional banks: the longer users stay with the bank, the more the bank can market credit cards, mortgages, fixed-income products, and other services to them, tapping into the lifetime value of individual customers. At the same time, this type of physical financial demand can avoid the strong cyclical fluctuations of the crypto industry: unlike crypto exchanges and speculative trading products, cross-border trade, margin, and credit are essential needs across cycles, covering the entire industrial chain of the real economy.

This is also the path for the crypto industry to move beyond speculation and become the underlying operating system for the formation of real capital.

Commercial trade always requires a universal value language: tokenization provides a standardized value carrier, and protocols build the grammar to execute these rules; companies that can integrate market scenarios from various global regions and guide physical assets into this universal trading system will capture the majority of incremental value in the industry over the next decade.