SoFi's Galileo and Technisys technology systems, as well as platforms such as Blend Labs, form the remaining underlying technical support for Lending as a Service (LaaS). Cross River Bank, as the invisible partner bank behind dozens of financial technology companies, has issued more than 96 million loans through cooperation, totaling more than $140 billion.

Companies that can win in the long term either occupy a certain choke point and become indispensable to all participants, or they open up multiple levels vertically to form composite competitive advantages. The failed companies will be trapped in the homogeneous business layer, lacking structural voice, and can only rely on price competition until profits approach zero.

Winners: core link controllers and multi-level companies with compound advantages

SoFi: Full-Stack Composite Tools

SoFi is the only company to cover four of the five tiers:

• Directly initiate consumer loans and mortgage loans.

• Export lending infrastructure to third parties through the Galileo platform, supporting approximately 160 million active accounts.

• Relying on the self-developed risk control model to carry out loan review, the core evaluation dimensions are repayment willingness, repayment ability and stability.

• Holding a banking license and owning Galileo and Technisys core banking technology systems at the infrastructure layer.

SoFi’s revenue in 2025 will reach a historical record of US$3.6 billion, a year-on-year increase of 38%. The platform has 13.7 million members and 20.2 million financial products. Management guided for 2026 revenue of $4.7 billion and EBITDA of $1.6 billion. Not only did the business have strong revenue growth, it was also very profitable, with a profit margin of 34%. The banking license alone allows SoFi to fund loans through deposits rather than the wholesale market, directly reducing funding costs by 170 basis points.

SoFi is building the “Amazon Cloud (AWS)” of lending, a platform that both competes with and empowers other lenders. Galileo itself has been built into a billion-dollar revenue engine. Technisys, acquired for $1.1 billion in 2022, provides a core banking system layer to third-party institutions. Bank licenses form a structural moat that most fintech lenders cannot replicate, although the industry has followed suit: the U.S. Office of the Comptroller of the Currency (OCC) received 14 applications for new bank licenses in 2025 alone, signaling that the battle for the infrastructure layer is accelerating.

Upstart and Pagaya: intelligent decision-making layer

Ironically, winning in the lending industry does not necessarily require doing the lending business yourself. Upstart and Pagaya both use risk control and review engines as their core, and their risk control effects are better than the self-developed models of lending institutions, and they do not need to rely on their own balance sheets to conduct business. This is exactly the implementation of the “water seller” logic in the field of credit decision-making.

Compared with traditional FICO-based risk control models, Upstart’s model can approve 44% more borrowers under the same bad debt rate, reduce default rates by 53%, and provide borrowers with significantly lower annualized interest rates. Currently, almost all new loan originations on the platform are fully automated, significantly reducing manual intervention. This is fundamentally different from the traditional consumer credit risk control model.

Pagaya is in the same track, but faces more severe market realities. The company does not issue loans directly, but instead licenses its AI risk control engine to banks. Since its founding in 2016, Pagaya has evaluated approximately $2.6 trillion in cumulative loan applications for 31 partner banks. Its structural positioning is clear: there is no need for borrowers to know the brand, only for banks to rely on its scoring system. But the current market does not recognize this logic. In the fourth quarter of 2025, network business volume increased by only 3% year-on-year, revenue fell short of market consensus, and the performance outlook was also lower than expected. The stock price plummeted by nearly a quarter in a single day. The value of the intelligent decision-making layer is completely restricted by the credit cycle. When the bad debt rate of the cooperative network increases, even excellent AI cannot withstand the pressure of deteriorating asset quality.

But the core logic still holds: FICO only forms a single cross-sectional score based on a small number of historical variables, and as consumers’ financial situations become increasingly complex and diverse, AI risk control systems will become increasingly critical. Unlike FICO, these systems continue to learn and optimize with each score completed.

Figure: New generation clearing and settlement channel

The cost of originating a single loan through traditional channels and the Mortgage Electronic Registration System (MERS) is US$11,000, but relying on the Figure technology system including Provenance blockchain and DART system, this cost can be reduced to US$717. This new access infrastructure reduces borrowing costs by orders of magnitude.

Figure has launched over US$21 billion in home equity products (mainly home equity credit lines) through the Provenance blockchain, and the cumulative transaction size on the chain has exceeded US$50 billion. Loan originations reached $2.7 billion in the fourth quarter of 2025, up 131% year over year. The company holds more than 180 lending licenses and U.S. SEC broker-dealer registration qualifications, and has the compliance basis for large-scale operations. At the same time, it has more than 300 white label lending partners, and has added one new partner at an average rate of one per day since it submitted its S-1 listing document in September last year. Its revenue has grown from a quarterly annualized rate of $28.5 million in the first quarter of 2023 to $146.8 million today.

Figure’s core business has little to do with crypto-assets, but its stock price trends are highly similar to Bitcoin. The company's settlement system reflects the logic of cost structure restructuring: final confirmation of settlement only takes a few seconds, while the traditional method takes more than a day; loan origination costs are only a fraction of the traditional model. Securitization-related cost savings exceed 100 basis points over the life of the loan – representing a potential cost reduction of more than $30 billion in the $3 trillion annual securitization market.

Aave: The core controller in the DeFi field

Aave accounts for more than half of the DeFi lending market. Liquidity will generate more liquidity, and borrowers will continue to gather on the platform with the deepest capital pool (network effect). Its cumulative loan disbursements have exceeded $1 trillion, and the agreement officially crossed the $1 trillion mark in cumulative loans last month.

In addition to its dominant position in the DeFi field, Aave’s most interesting structural aspect is its institutional lending business line Horizon. Horizon has taken in $580 million in deposits and aims to top $1 billion in 2026. It serves as a bridge between DeFi liquidity and traditional credit needs. If Aave can introduce on-chain funds into institutional-level lending products, it will become the capital supply layer for traditional lending institutions, opening up a potential total space (TAM) that is far larger than the retail DeFi market.

DeFi lending also offers a structural risk advantage that is often underestimated. Overcollateralization in DeFi typically ranges from 150%–180%, compared to only 50%–70% in traditional peer-to-peer lending. Bad debts in DeFi mainly come from oracles or technical failures rather than credit qualification defaults.

Affirm: Distribution Channel Lockdown

Affirm leads the buy now, pay later (BNPL) space by being deeply embedded in merchant payment settlement infrastructure. Critics focus on its consumer credit risks, but ignore the core structural logic: Affirm is not a consumer loan institution in the traditional sense, but a credit distribution channel for sales terminals. System integration with merchants is its moat. Given that BNPL is expected to cover 13% of all digital transactions, platforms that embed checkout processes at scale will charge structural “channel fees” from the commercial transactions themselves.

Losing Pattern: Four Structural Failure Patterns

We intentionally do not name companies that fit these patterns. If you are an investor or operator in the credit field, you naturally know who they are. More important than the specific names is understanding why these structural positions are doomed to fail, because the same patterns will create new victims in the next cycle.

Balance sheet-only lenders

The only competitive advantage of this type of enterprise is access to funds. They issue loans using traditional risk control methods, provide funds with their own balance sheets, and have no dedicated technology layer. They are just "brainless conduits" for money.

In a world where private credit management has reached $3.5 trillion and is on its way to $5 trillion, capital is not scarce; what is scarce is smart decision-making and infrastructure. Such companies can only compete on price, which squeezes profits to zero in every interest rate cycle and forces them to take excessive risks. These lenders end up extending credit to risky businesses and suffer losses when the cycle turns.

Most of these participants are traditional consumer lending institutions, small banks, and financial technology lending companies that have never established a technology moat beyond their initial loan products. When capital becomes homogeneous, lending without technical advantages and relying solely on its own balance sheet is tantamount to slowly handing over shareholder equity to borrowers.

CeFi lending victim

The centralized crypto lending (CeFi) platform that collapsed in 2022 was not a victim of the bear market. They fell on the credit industry's oldest failure patterns: maturity mismatches, misappropriation of customer funds, lending against illiquid assets and a lack of transparent risk management.

The decentralized lending (DeFi) protocol that automatically enforces mortgage discipline through smart contracts and makes the mortgage rate publicly visible on the chain has not exploded. What’s really going wrong are those CeFi platforms that rely on human judgment and have opaque balance sheets. Any lending platform — whether in crypto or traditional finance — that only lets you trust its balance sheet without showing you the collateral is repeating an old structural path that has failed before.

Ghost Protocol

There is a type of DeFi lending protocol that is technically alive but structurally dead. After they went online, they attracted initial lock-up funds through token incentives, but stagnated after the incentives faded. The code works and the lock-in value (TVL) is not zero, but the usage curve is flat or declining, and there is no clear natural demand growth path.

The reason is that DeFi lending exhibits extreme power law distribution characteristics: liquidity will concentrate on platforms with network effects - as evidenced by Aave's absolutely dominant market share. Protocols that fail to break through the critical scale will fall into a structural "no man's land": they are too small to attract natural liquidity and supporting integration; they are not too small to be shut down gracefully. As profit-seeking funds flow to leading platforms, their locked-in value continues to slowly drain away, and this process is irreversible. These are zombie protocols that survive on the sunk cost of governance tokens.

Lending institutions that missed the platform transformation

Some companies built strong loan origination businesses in the previous cycle but never developed platform capabilities. They have no API distribution channels, no embedded financial partnerships, and no technology licensing model. It has strong loan origination capabilities, but cannot export its capabilities externally.

As the credit industry becomes modular, being able to become a component in someone else’s system is as important as directly originating loans. For companies that can only lend directly to end borrowers, their growth will be limited by their own channel coverage; for companies that can provide lending capabilities to other institutions, there is no upper limit on their potential market space (TAM). Pure loan originators usually have good single-customer economic models, but have a flat growth curve because the reachable market is limited to their own brands and channels. In a modular architecture, becoming an excellent lending institution is a necessary condition, and becoming an excellent lender that can be accessed by other lending institutions is the real winning position.

Objects worthy of attention

The above-mentioned winning companies have become market consensus or close to consensus, but this is not the case for the following companies. They have the structural qualities to become masters of core links, but they have not yet been proven at scale. These are targets worth continuing to track.

Morpho

Morpho’s total locked value (TVL) has reached US$6.6 billion, a year-on-year increase of 164%, and its market value exceeds US$800 million. Its structural logic is completely different from Aave: Aave is a commercial bank in decentralized finance (adopting a unified lending capital pool model), while Morpho is building a modular lending layer that allows institutional participants to customize their own lending market based on their own risk parameters, collateral types and interest rate models. If the lending system truly becomes modular, Morpho will become a lending-as-a-service protocol at the on-chain level.

Maple Finance

Maple has issued a total of US$11.3 billion in loans in 2025, serving 65 active borrowers, and its assets under management (AUM) have increased significantly from US$516 million to US$4.6 billion, an increase of 767%. The company aims to achieve $100 million in annual recurring revenue (ARR) by 2026. Maple is one of the few protocols that is truly committed to bringing real-world corporate lending to blockchain infrastructure, enabling business by connecting institutional credit needs with on-chain funding and settlement systems. The explosive growth of its assets under management shows that institutions’ interest in the on-chain credit market is shifting from theoretical conception to actual implementation.

Cross River Bank

Since 2008, Cross River has partnered to originate more than 96 million loans totaling more than $140 billion. It’s the partner bank behind Affirm, Upstart, and dozens of other fintech lenders. According to sources, the bank is preparing for an IPO. Cross River is an “invisible bank” that serves as the infrastructure layer to support the operation of a considerable part of the financial technology lending business. As the cooperative banking model matures, its market position brings a voice that no single financial technology lender can replicate. The key to the bank's success is to make fintech companies unable to carry out lending business without its support.

License battle

The U.S. Office of the Comptroller of the Currency (OCC) received 14 applications for new bank licenses in 2025 alone—almost equal to the total of the past four years. The total number of license applications submitted by financial technology institutions has reached a record high of 20. Affirm, Stripe and Nubank are all actively applying for licenses. These companies regard licenses as the core competitiveness of the final restructuring of their credit business.

Companies that started as technology service providers are now capturing the economic value of the entire industry chain by obtaining regulatory qualifications. The status of a bank license in the lending field is comparable to that of a regional node in cloud computing. The reasons are:

• The construction cost is extremely high;

• Industry participants cannot avoid it;

• Once acquired, it forms a permanent structural advantage.

The business logic is very clear: every 1 basis point optimization of capital costs can increase the pre-tax return on equity by several percentage points. For large-scale enterprises, the advantages brought by licenses are extremely significant. But for small and medium-sized institutions, licenses may become a trap: they have to bear all the compliance costs, regulatory inspection pressure and capital requirements, but do not have enough business scale to cover these expenses. Only companies that already have a huge business volume can use licenses as growth accelerators.

Credit structure in 2030

If there is one core analytical framework to remember from this article, it is the following three questions. They apply to all lending companies, whether listed, unlisted or on-chain institutions.

First: Which level does the company occupy? Loan origination and homogeneous capital supply belong to the Red Ocean track, and profit margins will continue to compress along with the industry cycle. AI risk control, blockchain settlement, and bank licenses are core chokepoints, and their value will continue to accumulate with compound interest. If a company is stuck in the Red Sea track and cannot cut into the core links, no matter how good the team is, its long-term profitability will be continuously eroded.

Second: Is it a platform or a single product? A single product serves end borrowers, and its scale grows linearly with its own channels; the platform empowers other lending institutions, and its growth relies on the size of the entire ecosystem, not just its own business. SoFi has both attributes, and Pagaya is a pure platform company. Companies that only lend directly to their own customers will have a ceiling on growth, while platform companies do not have this limitation.

Third: Do you have a regulatory moat? Whether it’s a banking license, the 180 state-by-state lending licenses, or programmatic compliance through smart contracts, it all falls into this category. In the lending industry, regulation is not an additional cost but core infrastructure. Companies that recognize this early will build an advantage that will take competitors years and huge amounts of capital to catch up.

By 2030, the credit industry will look less like traditional banking and more like the cloud computing industry. A few full-stack platforms will cover multiple levels and form compound interest advantages in each link: the most typical representative in the traditional financial field is SoFi, and the on-chain field is Aave. Centering around these core platforms, a large number of specialized hierarchical service providers will access through APIs and on-chain channels, each focusing on segmented functions and charging service fees.

In the global debt market of US$348 trillion, the penetration rate of financial technology is less than 0.2%. This market is not left to be divided among hundreds of lending institutions, but will be dominated by more than a dozen platforms and become the underlying support of the entire industry.