DeFi is replicating the script of SaaS and fintech

Original Title: DeFi Is Following The SaaS And Fintech Playbooks
Original Author: Lorenzo Valente, ARK Invest
Original Translator: Ismay, BlockBeats

Editor's Note: The development of the crypto industry has never been isolated; it often resonates with historical waves of technology. This article delves into how the "splitting---reintegration" cycle of SaaS and fintech is reflected in DeFi and crypto applications. The author not only lays the groundwork with Web2 cases like Airbnb and Robinhood but also combines the evolution of Uniswap and Aave to reveal how crypto protocols have transitioned from singular primitives to modular "financial Legos," gradually integrating into super applications today. For readers trying to understand the future landscape of DeFi, this serves as a framework for observing the future through the lens of history.

As the crypto industry matures, investors are beginning to look for clues from past technological waves to predict the next major trend or turning point. Historically, digital assets have been difficult to compare directly with previous technology cycles, making it challenging for users, developers, and investors to anticipate their long-term development paths.

This situation is changing. According to our research, the "application layer" of crypto is evolving, and its patterns are quite similar to the "splitting---reorganization" cycles experienced by SaaS (Software as a Service) and fintech platforms.

In this article, I will explain how the splitting and reorganization cycles in the SaaS and fintech sectors are being replicated in DeFi and crypto applications. The evolution logic of this model is as follows:

To understand the cycle of splitting and reorganization, the concept of "composability" is crucial. In the fintech and crypto communities, "composability" is a common analytical term that refers to how financial or decentralized applications and services—especially at the application layer—can seamlessly interact, integrate, and build upon each other like Lego blocks. Based on this core concept, we will further analyze the evolution path of product structures in the following two sections.

From Vertical to Modular: The Great Splitting

In 2010, Andrew Parker of Spark Capital published a blog depicting how dozens of startups capitalized on the "splitting" opportunity presented by Craigslist. At that time, Craigslist was a "horizontal" internet marketplace offering a variety of services, from housing rentals and part-time jobs to second-hand goods trading, as shown in the figure below.

Parker concluded that many successful companies—Airbnb, Uber, GitHub, Lyft—cut out a very small vertical niche from Craigslist's broad functionality and significantly improved the experience. This trend initiated the first wave of "market splitting": the large, all-encompassing platform of Craigslist was gradually replaced by applications focused on single purposes. These newcomers not only enhanced the user experience (UX) of Craigslist but fundamentally redefined it. In other words, "splitting" deconstructed a broad platform into narrower, independently autonomous verticals, uniquely meeting user needs and thereby disrupting Craigslist.

So, what drove this wave of splitting? The underlying shift was in technological infrastructure. Advances in APIs (Application Programming Interfaces), cloud computing, mobile user experience, and embedded payments lowered the barriers to building focused applications, enabling developers to launch more precise services with world-class user experiences.

A similar splitting occurred in the banking industry. For decades, banks provided a bundled suite of financial services under the same brand and application—from savings and loans to insurance. However, over the past decade, fintech startups have gradually dismantled this bundling model in a "surgical" manner, each focusing on a specific vertical.

Traditional banks' "bundled services" include:

• Payments and transfers
• Checking and savings accounts
• Deposit growth products
• Budgeting and financial planning
• Lending and credit
• Investment and wealth management
• Insurance
• Credit and debit cards

In the past decade, this "banking package" has been systematically split, giving rise to a large number of venture-backed fintech companies, many of which have grown into unicorns, decacorns, or even close to hectocorns:

• Payments and transfers: PayPal, Venmo, Revolut, Stripe
• Bank accounts: Chime, N26, Monzo, SoFi
• Savings and growth: Marcus, Ally Bank
• Personal finance and budgeting: Mint, Truebill, Plum
• Lending and credit: Klarna, Upstart, Cash App, Affirm
• Investment and wealth management: Robinhood, eToro, Coinbase
• Insurance: Lemonade, Root, Hippo
• Cards and spending management: Brex, Ramp, Marqeta

These companies each focus on a service area where they can outperform traditional banks, leveraging new technological advantages and distribution models to modularize growth-oriented financial services.

In SaaS and fintech, splitting not only disrupts incumbents but also creates entirely new categories, ultimately expanding the total addressable market (TAM).

From Modular Back to Bundled: The Great Reintegration

Recently, Airbnb launched a "Services and Experiences" section and redesigned its application. Now, users can not only book accommodations but also explore and purchase additional offerings such as museum visits, food tours, dining experiences, gallery strolls, fitness classes, and beauty treatments.

Initially just a peer-to-peer accommodation marketplace, Airbnb is gradually evolving into a vacation super application—it repackages travel, lifestyle, and local services into a unified, coherent platform. Moreover, over the past two years, Airbnb's products have already surpassed home rental, progressively integrating payments, travel insurance, local guides, concierge tools, and curated experiences into its core booking service.

Robinhood is also undergoing a similar transformation. Once, it disrupted the traditional brokerage industry with commission-free stock trading; now, it is expanding into a full-stack financial platform, reintegrating many vertical services that fintech startups had previously split apart.

In the past two years, Robinhood has taken the following actions:

• Launched the Robinhood Cash Card for payments and cash management;
• Added cryptocurrency trading features;
• Introduced retirement accounts;
• Brought in margin investing and credit cards;
• Acquired Pluto, an AI-driven research and wealth consulting platform.

These moves indicate that, like Airbnb, Robinhood is re-bundling previously fragmented services to build a comprehensive financial super application. By controlling more of the financial stack—savings, investment, payments, lending, and financial consulting—Robinhood is reshaping itself from a brokerage into a full-fledged consumer finance platform.

Our research shows that this "splitting---reintegration" dynamic is also affecting the crypto industry. Next, we will analyze Uniswap and Aave as case studies.

The Splitting and Reintegration Cycle of DeFi: Two Case Studies

Case Study 1: Uniswap—From Monolithic AMM to Liquidity Legos, Then to Trading Super Application

In 2018, Uniswap launched on Ethereum as a simple yet revolutionary automated market maker (AMM). Early Uniswap was a vertically integrated application: a small smart contract codebase, plus an official frontend hosted by the team. Its core AMM function—exchanging ERC-20 tokens in a constant product pool—operated on-chain as an independent protocol. Users primarily accessed this function through Uniswap's own web interface. This design was extremely successful; by mid-2023, Uniswap's cumulative on-chain trading volume had surpassed $1.5 trillion. With a highly controlled stack, Uniswap provided a smooth user experience in token exchanges, laying the groundwork for the early development of DeFi.

In the v1/v2 stages, Uniswap implemented all trading logic on-chain, without the need for external price oracles or off-chain order books. The protocol determined prices within a closed system through liquidity pool reserves (x*y=k formula). The Uniswap team developed the main user interface, which interacted directly with the contracts. In the early days, most users accessed the protocol through this frontend, which was almost equivalent to a dedicated decentralized trading entry point. Apart from Ethereum itself, Uniswap did not rely on other infrastructure. Liquidity providers and traders interacted directly with the contracts, resulting in a system that was simple yet relatively isolated.

As DeFi expanded, Uniswap evolved from a single application into composable "liquidity Legos." Its open, permissionless nature allowed other projects to integrate and extend its pools. Uniswap Labs gradually loosened its control over the stack, allowing external infrastructure and community-built functionalities to play a larger role:

Decentralized trading aggregators and wallet integrations: A significant portion of Uniswap's trading volume began to flow through external aggregators (such as 0x API, 1inch) rather than its own interface. By the end of 2022, it was estimated that 85% of trading volume was routed by aggregators like 1inch, as users sought the best prices across platforms. Wallets like MetaMask also embedded Uniswap's liquidity into their exchange functions, allowing users to trade directly through Uniswap within their wallets. This made the AMM more like a plug-and-play module within the DeFi stack, rather than relying on a native frontend.

Oracles and data indexing: Although Uniswap contracts themselves do not require oracles to execute trades, other protocols surrounding its ecosystem use Uniswap's pool prices as on-chain price references. Its official frontend also relies on external indexing services, such as querying pool data through The Graph's subgraphs, to provide a smoother interface experience. Uniswap itself does not need to run indexing nodes but utilizes community-driven data infrastructure—a modular approach that outsources the burden of data processing to specialized indexing services.

Multi-chain deployment: In the modular phase, Uniswap expanded beyond Ethereum to multiple blockchains and rollups, including Polygon, Arbitrum, BSC, and Optimism. Its governance mechanism authorized the core protocol to deploy on these networks, treating each blockchain as a liquidity plugin. This multi-chain strategy highlighted its composability: Uniswap can exist on any EVM-compatible chain, rather than being bundled in a single vertical environment.

However, recently Uniswap has begun to return to vertical integration, seemingly aiming to regain control over the user journey and optimize the entire stack for its own use cases. Key reintegration actions include:

Native mobile wallet: In 2023, Uniswap launched its self-custodied mobile wallet "Uniswap Wallet," followed by a browser extension, allowing users to directly store tokens and interact with products. This marks its beginning to reclaim control over the user interface layer, rather than relying entirely on wallets like MetaMask. Through its own wallet, Uniswap has re-integrated activities like token exchanges and NFT browsing into its control, potentially routing liquidity back to Uniswap.

Built-in aggregation (Uniswap X): To eliminate reliance on third-party aggregators, Uniswap launched a built-in aggregation and trade execution layer, Uniswap X. It sources liquidity from multiple AMMs and market makers through an off-chain "liquidation network" and settles on-chain. This transforms Uniswap's interface into a one-stop trading portal, integrating the best prices for users like 1inch or Paraswap. The difference is that this functionality is run by Uniswap's own protocol, eliminating the need for users to switch platforms, with the experience fully internalized. Uniswap X has been embedded into its web application and may also appear in the wallet in the future.

Dedicated chain (Unichain): In 2024, Uniswap announced the launch of its own Layer 2 blockchain "Unichain" as part of the Optimism Superchain. By extending vertical integration to the infrastructure layer, Unichain is optimized for Uniswap and DeFi trading, aiming to reduce fees by about 95% and latency to 250 milliseconds. Uniswap is no longer just an application residing on someone else's chain but fully controls the blockchain environment that hosts its contracts. By operating Unichain, Uniswap can optimize all aspects from gas costs to MEV mitigation and introduce native protocol fee sharing to UNI holders. This full-cycle transformation means that Uniswap has evolved from a decentralized application reliant on Ethereum to a vertically integrated platform with its own interface, aggregation execution layer, and dedicated blockchain.

Case Study 2: Aave—From Peer-to-Peer Lending Market to Multi-Chain Deployment, Then to Credit Super Application

Aave's origins trace back to 2017 when it was known as ETHLend, an independent lending application. In 2018, ETHLend rebranded to Aave, evolving into a decentralized peer-to-peer lending market. The team developed smart contracts for lending and provided an official web interface for users to participate. At this stage, ETHLend/Aave used an order book-like approach to match loans, handling everything from interest rate logic to lending matchmaking.

As it gradually transitioned to a liquidity pool model (similar to Compound's funding pool lending), Aave achieved vertical integration. Its v1 and v2 contracts on Ethereum introduced innovations like flash loans—allowing borrowing without collateral within the same transaction—and interest rate algorithms. Users primarily accessed services through Aave's official web dashboard. The protocol internally managed key functions like interest accrual and liquidation, relying little on third-party services. In short, Aave's early design was a monolithic money market: a decentralized application with its own UI, handling deposits, loans, and liquidations on the same platform.

Aave has also always been part of a larger DeFi symbiotic relationship. From the beginning, it incorporated MakerDAO's DAI stablecoin as a core collateral and lending asset. In fact, Aave launched almost simultaneously with Maker during the ETHLend phase and immediately supported DAI, indicating that these vertically integrated pioneers were tightly coupled and that "island protocols" did not exist early on. Even in the "vertical" phase, Aave's operations relied on products from other protocols—stablecoins.

As DeFi expanded, Aave began to "split" and embrace a modular architecture, outsourcing parts of its infrastructure while encouraging others to develop on its platform. Several key shifts towards composability and external dependencies include:

External oracle networks: Aave no longer fully relied on its own price sources but adopted Chainlink's decentralized oracles to provide reliable asset pricing for assessing collateral value. Price oracles are crucial for lending protocols as they determine when loans are under-collateralized. Aave governance chose Chainlink as the primary price source for most assets, outsourcing pricing infrastructure to a specialized third-party network. While this modular approach enhances security (e.g., Chainlink aggregates multiple data sources), it also means Aave's stability relies to some extent on external services.

Wallet and application integrations: Aave's liquidity pools gradually became foundational modules for many dApps. Portfolio management tools and dashboards (like Zapper, Zerion), DeFi automation tools (like DeFi Saver), yield optimizers, etc., all accessed its contracts through Aave's open SDK. Users can even access or lend through third-party interfaces, with the official interface being just one of many entry points. DEX aggregators also indirectly utilize Aave's flash loan functionality to execute complex multi-step trades initiated by services like 1inch. Aave's open-source design fosters composability: other protocols can package its functionalities into their own logic, such as calling Aave flash loans in Uniswap arbitrage bots. This positioning as a liquidity module rather than an isolated application extends its reach throughout the DeFi ecosystem.

Multi-chain deployment and isolation mode: Similar to Uniswap, Aave has also expanded to multiple networks, including Polygon, Avalanche, Arbitrum, Optimism, etc., effectively achieving cross-chain modularity. Aave v3 introduced "isolated markets," setting independent risk parameters for specific assets, sometimes even operating outside the main liquidity pool. Additionally, Aave launched a "permissioned" version—Aave Arc—for KYC-compliant institutional clients, essentially an independent modular instance of Aave.

These examples indicate that Aave demonstrated flexibility during the splitting phase, no longer confined to an integrated environment but leveraging broader infrastructure: Chainlink providing data, The Graph providing indexing, wallets and dashboards providing user entry points, and tokens from other protocols (like DAI or Lido's stETH) serving as collateral. Modularity enhances composability and avoids reinventing the wheel. However, the trade-off is a decrease in Aave's control over parts of its stack and increased risks from external dependencies.

In recent years, Aave has begun to return to vertical integration, developing key components that it previously relied on externally. In 2023, Aave launched its own stablecoin GHO. Historically, Aave primarily matched users borrowing other stablecoins, especially DAI, which had significant scale on Aave. The introduction of GHO means that the Aave platform itself now has a native stablecoin and serves as a distribution channel for other protocols' stablecoins. Like DAI, GHO is an over-collateralized, decentralized stablecoin pegged to the dollar. Users can mint GHO by depositing on Aave v3, allowing Aave to reclaim the "stablecoin issuance" aspect of the lending stack.

The result is:

Aave is no longer just a place for borrowing stablecoins but a stable asset issuer, directly controlling the parameters and revenue of the stablecoin. GHO competes with DAI, allowing Aave to cycle interest back into its ecosystem, with GHO's yields directly benefiting AAVE token stakers rather than indirectly increasing MakerDAO's revenue.

Introducing GHO also requires dedicated infrastructure. Aave established "facilitators" (including the main liquidity pool) to mint and burn GHO and set governance rules. With this new functional layer, Aave has essentially created a "self-version" of a MakerDAO product tailored for its community.

Furthermore, Aave is leveraging mechanisms like Chainlink's Smart Value Routing (SVR) to help users recover MEV (maximum extractable value, similar to order flow payments in stock markets). The deep coupling with the oracle layer allows arbitrage profits to flow back to the protocol, blurring the boundaries between Aave and the underlying blockchain mechanisms. This indicates Aave's intention to further customize its underlying infrastructure, such as oracle behavior and MEV capture, for self-benefit.

Although Aave has not yet launched its own wallet or chain like Uniswap, other projects initiated by its founders suggest that its goal is to build a self-sufficient ecosystem. For example, its Lens Protocol social network may integrate with Aave in the future, exploring finance based on social reputation. Architecturally, Aave is gradually providing all core financial primitives: lending, stablecoins (GHO), and possibly decentralized identity (Lens), rather than relying on external protocols. The core of its product strategy is to "deepen the platform," enhancing user stickiness and protocol revenue through services like stablecoins and lending.

In summary, Aave's evolution path is: from a closed lending dApp to an open "liquidity Lego," relying on external protocols like Chainlink and Maker, to now reintegrating key functions and building a larger vertical financial suite. The launch of GHO particularly highlights Aave's desire to regain control over the stablecoin layer that was previously outsourced to MakerDAO.

Our research indicates that the trajectories of projects like Uniswap, Aave, MakerDAO, and Jito reveal a broader cyclical pattern in the crypto industry. In the early days, vertical integration—building monolithic products around specific scenarios—was crucial for pioneering new functionalities like automated trading, decentralized lending, stablecoins, or MEV capture. This self-contained design allowed projects to iterate quickly in the early market while maintaining quality control. As the industry matures, modularity and composability have become mainstream: protocols begin to split their stacks, leveraging external advantages to launch new features or create more value for external stakeholders, gradually becoming "financial Legos."

However, the success of modularity and composability also brings new challenges. Relying on external modules increases risks and limits the protocol's ability to capture the value it creates. Today, leading protocols with strong PMF (product-market fit) and revenue sources are shifting their strategic focus back to vertical integration. They are not abandoning decentralization and composability but are strategically reintegrating key components: self-built chains, wallets, stablecoins, frontends, and other infrastructures. The aim is to provide a smoother user experience, capture additional revenue sources, and avoid being constrained by competitors.

Uniswap is building a wallet and dedicated chain; Aave is issuing GHO; MakerDAO is forking Solana to launch NewChain; Jito is merging staking/re-staking with MEV. In our view, any sufficiently large DeFi application will ultimately seek its own vertical integration solution.

Conclusion

History does not repeat itself exactly, but it often rhymes. The crypto industry is playing a familiar tune. Just as the revolutions of SaaS and e-commerce platforms over the past decade, DeFi and application layer protocols are performing the trajectory of "splitting---reintegration" around new technological primitives, changing user expectations, and a stronger desire for value capture.

In the 2010s, a batch of startups focused on single verticals within Craigslist's vast market effectively split it into independent companies. This "splitting" birthed giants—Airbnb, Uber, Robinhood, Coinbase—who later embarked on their own "reintegration" journeys, merging new vertical businesses and services into unified and sticky super platforms.

The crypto industry is rapidly following the same path.

The initial explorations were highly focused vertical experiments: Uniswap as an AMM, Aave as a money market, Maker as a stablecoin vault. With open and permissionless modular designs, they gradually evolved into "financial Legos," releasing liquidity, outsourcing key functions, and allowing composability to flourish. Now, as usage scales and markets differentiate, the pendulum is swinging back toward integration.

Today, Uniswap is evolving into a trading super application, equipped with its own wallet, dedicated chain, cross-chain standards, and routing logic; Aave has launched its own stablecoin and is repackaging lending, governance, and credit primitives; Maker is building a new chain to optimize the governance of its monetary system; Jito is unifying staking, MEV, and validation logic into a full-stack protocol; Hyperliquid is integrating exchanges, L1 infrastructure, and EVM to create a seamless on-chain financial operating system.

In the crypto world, primitives are inherently modular, but the best user experiences—and the most defensible business models—often come from reintegration. This is not a betrayal of composability but its completion: building the optimal Lego blocks and using them to construct the sturdiest castle.

DeFi is compressing this complete cycle into just a few years. Why? Because DeFi is built on a completely different track:

Permissionless infrastructure: Greatly reduces experimental friction, allowing any developer to fork, copy, or extend existing protocols in hours rather than months.

Capital formation happens instantaneously: Through tokens, teams can fund new projects, ideas, or incentive mechanisms at unprecedented speeds.

Liquidity is highly mobile: Total locked value migrates rapidly with incentives, making it easier for new experiments to gain attention, and successful projects can achieve exponential growth.

Accessible markets are broader: Protocols target global, permissionless users and capital pools from day one, often scaling faster than Web2 counterparts constrained by geography, regulation, or distribution channels.

DeFi's super applications are expanding rapidly in real-time. In our view, the winners will not be the projects with the highest degree of modularity but those that can accurately determine which stacks must be owned, which can be shared, and when to flexibly switch between the two. ```