Deep Dive into DeFi (Part 1)

Dec 10, 2021  •  Written by  Dennis Le and Derek

Welcome to the 6th Edition of the Web 3 newsletter produced by the Ocular team on topics surrounding L1/L2s, DeFi, Gaming, NFTs and crypto in Southeast Asia! Ocular is a crypto fund based in Singapore managed by Openspace Ventures, one of the top VCs in the region.

In previous posts, we have written about specific L1/L2 projects, as well as the history and state of NFTs. By popular demand, in the next few editions, we will do a deep dive series into DeFi (Decentralized Finance).

Welcome again to the journey of Web 3 with us. Enjoy reading. Get in touch:

Disclaimer: All write-ups in this substack channel are Ocular Ventures’ own opinions, and should not be misconstrued as investment advice or recommendations. Ocular Ventures does not hold positions in the projects discussed in this channel, unless otherwise stated. This content is intended for informational purposes only.


Have you ever heard the phrase: “Every company will be a Fintech company”? It was a term coined by Angela Strange at Andreessen Horowitz, who outlines her views in this video. To most people, this could be a controversial topic. Those at the forefront of Fintech and Crypto development may agree with this notion, which is based on the promise of ‘Embedded Finance’ (another concept that means non-financial companies could embed financial services into its offering from a third-party financial service provider). On the other hand, those who have been dealing with the legacy of traditional finance (‘TradFi’) may find it unrealistic based on their real-life pains dealing with TradFi providers.

(Source: Scott Adams)

As a group of venture capitalists who have studied the global Fintech space for many years, we gear more towards the former view and remain bullish on the accelerating development of Fintech. At the same time, we keep asking ourselves how Fintech development and adoption could progress even faster than its current pace; and how we as investors could help trigger such progress. The answer probably came in 2018 but it became a lot more obvious since the summer of 2020 when the financial use cases of crypto, collectively called Decentralized Finance (‘DeFi’), saw phenomenal growth in adoption and capital flow.

(Source: WaitButWhy)

Today, DeFi is the new Fintech that is merging with traditional Fintech and could potentially be the centre of the next Fintech wave. In this piece of writing, we will attempt to explain our thinking on where Fintech is headed and why DeFi will play a critical role in this development. This piece is also the first part of our three-part series on DeFi - the second piece will discuss the main verticals of DeFi and the trends we are observing, and the third piece will discuss the next steps for DeFi.

To understand where we are today, it’s always good to start with a walk through the history of Fintech and review where we are in the current cycle.

Past - Fintech 1.0: When the ‘Fin’ met the ‘Tech’ for the first time

Many moons ago, financial services were synonymous with banks, asset managers and to a certain extent, insurance providers. They were the primary actors of any economy handling deposits, lending and investment activities. Assets were kept physically in concrete or marble buildings, and operations were pretty much handled by pen and paper. For a long time, these financial institutions represent the trust that consumers and investors have in the financial system of a country. The bigger the building, the higher the trust.

(Source: Pinterest)

Then the mainframes and COBOL came in the 1960s and totally changed the industry. In the next half a century, financial institutions started spending money developing software and leveraging the Internet to enhance existing services or enable new services. Banks were able to interconnect, wire money across each other and even across countries electronically, allowing users to access products remotely via the Internet or ATMs.

We call this period the Fintech 1.0 era which is characterized by:

  • Growing adoption of software by traditional financial institutions from electronic payment / clearing systems to core banking / software-enabled front offices.
  • Software was mostly functionality-/ product-centric, monolithic, developed in-silos, sitting on-premises, making it hard to upgrade and replace. Even today, the cost of upgrading and replacing core banking is very expensive.
  • Financial institutions’ mindset was all about value capturing and value extraction, resulting in walled-garden ecosystems. Banks considered their technologies, systems or databases, etc. their competitive moat and refused to let non-approved actors access their back-end. Fintech players were mostly limited to tech-enabled traditional actors or software vendors to those actors.
  • Consumers conducted low-value activities virtually (e.g. checking account balance, making small transfers, repaying credit card loans) but needed to go to physical branches for high-value activities (e.g. opening a bank account, taking a loan, making a large cash transfer or withdrawal, signing a contract) because of the large trust gap of doing things virtually vs. physically and the high cost of reconciliation across various siloed systems.

Present - Fintech 2.0: Unbundling and re-bundling of financial services

In the 2000s, the cloud infrastructure industry took a great leap and enabled software (application) adoption to hit an inflexion point. In July 2002, Amazon created subsidiary Amazon Web Services, and in March and August 2006, Amazon introduced its Simple Storage Service (S3) and Elastic Compute Cloud (EC2) respectively. These products pioneered the usage of server virtualization to deliver Infrastructure-as-a-Service (IaaS) at a cheaper and on-demand pricing basis. In April 2008, Google released Google App Engine, a Platform-as-a-Service (PaaS – the first of its kind) proving fully maintained infrastructure and a deployment platform for users to create web applications using common languages such as Python, Node.js and PHP. In February 2010, Microsoft released Microsoft Azure.

The advancement of cloud infrastructure and the abstraction of software deployment and scalability challenges enabled Fintech to transition to its second era where financial institutions (especially banks) started moving their software to the cloud, systems became more modular and Software-as-a-Service (SaaS) made it easier to rebuild, buy and sell existing financial products.

(Source: a16z)

This trend is marked by the banking infrastructure stack being broken down into various layers and tech companies offering those individual layers as a service to other companies like Uber, Shopify, Earnin, etc. to offer financial services. There are many more existing examples in developed markets like the US or Europe, as illustrated by a16z here.

To summarize, the Fintech 2.0 era is characterized by:

  • Unbundling banking infrastructure into various as-a-service layers which help non-financial companies easily re-bundle and build their own financial service offerings, shortening time to market and lowering development costs.
  • Greater number of companies offering financial products. Greater number of consumers have access to financial products, thus increasing financial inclusivity.
  • Consumers comfortably conducting a growing number of use cases online. Physical branches are becoming less important. Digital-only / neobanks are winning market shares from traditional physical banks especially in the younger/ more tech-savvy consumer segments.
  • While traditional financial institutions’ mindset is still about value capturing and value extraction, their ecosystems are getting more open (e.g. in Europe, regulators even mandated open banking infrastructure). For tech companies offering financial services, the initial mindset has been about financially enabling more unbanked/underbanked consumers and ‘stealing’ existing consumers from traditional financial institutions. It’s about value growing first and value capturing later.
  • Although the majority of fintech companies rely on more modern infrastructure, it is nearly impossible to avoid interacting with legacy tech because more than 40% of banks’ code is built on COBOL. This is still a big pain point for developers, and worse, they are also finding themselves building the same infrastructure again and again.

Outstanding issues with TradFi

While TradFi has made significant progress with the advent of the various Fintech waves, there remain a few issues with the industry that has yet to be resolved:

Inaccessible to the masses

Today, there are around 2 billion unbanked individuals globally. Some of the key reasons provided for why individuals do not have a bank account include not having sufficient money, long distances to the nearest bank and the lack of required documents. This is understandable, considering that many banks require individuals to have a minimum sum of money before processing their transactions as well as some form of identity documents and assurance of their credit history before onboarding them as customers. Given the need to perform verification checks, banks may also require individuals to register for an account in-person, creating an additional barrier to account ownership.

While it is important to carry out the necessary due diligence on consumers, the current process adopted by the traditional banking sector has led to certain individuals being excluded from the system. A recent study by the British research platform Merchant Machine revealed that Morocco, Vietnam, Egypt, Philippines and Mexico top the list of the most unbanked countries in the world with more than 60% of their population unbanked. In fact, the developed regions are not perfect role models either – around 21% of individuals in North America are unbanked. More can be done to promote greater financial inclusion.

Wavering public trust

The 2021 Edelman Trust Barometer, which surveyed 33,000+ respondents across 28 countries, showed that most individuals still do not have a high level of trust in the financial services sector. While trust levels in the sector have increased by 8% over the last decade, it remains in neutral territory with a score of 52 and it is behind many other industries, including entertainment, automotive and telecommunications.

While consumers would like banks to be reliable, approachable and supportive, many tend to have negative experiences that affect their trust levels in these institutions. This could include financial products that are often riddled with convoluted terms and conditions and an overall lack of transparency on fees; a lack of support and real-time updates on their financial standing which makes them feel unheard or unimportant; and privacy concerns where they may have been a data leak, resulting in consumers feeling that their money or personal information is at risk.

Banks need to build the trust imperative as it affects the overall effectiveness of the financial system. A high degree of confidence in banks will lead to increased contributions to the pooling of savings and expansion of credit by banks. Trust is also essential for the stability of the financial system, as it reduces the danger and intensity of bank runs.

Trust can also affect the bottom lines of banks. A 2018 Accenture study of 900 companies found that for the 54% of companies that experienced a drop in trustworthiness, it cost them about $180 billion in total revenue over just two years. While the percentage decline in revenue will vary by industry, it appears that the banking sector will be severely affected. Revenue growth for companies in the banking sector is projected to decrease by 21.8% should there be a drop in trust, highlighting the material importance of maintaining a high degree of public trust.

Influenced by legacy

The above two issues that have been highlighted calls for the financial sector to undertake a closer look at their processes and potentially implement widespread changes in the industry. Nonetheless, the financial sector may find it difficult to push the envelope, given that it faces its own set of constraints and will likely be hampered by legacy issues.

While financial institutions themselves have long sought to expand their pool of consumers and include as many individuals as possible, the current KYC methods have proven effective in minimising default rates and there is little incentive to modify the process. In addition, financial institutions have a bottom line to meet and it may not be cost-effective to process transactions of significantly low value.

On the second issue of public trust, we understand that trust in banks has fallen sharply since the global financial crisis in 2008 and these events are likely to be embedded in a society’s memory and personal perceptions. The effects of a loss in trust are long-lasting and despite the best efforts of banks, it will likely take a while for public trust to be rebuilt.

Given the challenges faced by banks today, it is clear that there is a need for a new model to revolutionize and galvanize the financial sector. It calls for a system that is permissionless and accessible to all; open-source and transparent to build public trust; and more importantly, not burdened by past experiences such that it can be easily adapted, unbundled and recomposed to meet the needs of consumers. Can you guess where we are heading towards?

Future - Fintech 3.0 and the open-source lego bricks

Before we reveal the answer, it is important to outline some key objectives of the next wave of transformation in the financial sector. Fintech 2.0 was pretty much about unbundling and re-bundling existing applications and infrastructure to create new services or enable non-financial companies to offer financial services. Does the question then become what’s next from here? What if you can slice the above mentioned “as a service” layers even further to the most basic primitives, thus creating a lot more lego bricks and enabling an even great number of possible combinations? What if these lego bricks are composable and interoperable, allowing companies to select and assemble them in any combination to meet specific user requirements? Even greater, what if we open-source all these lego bricks so others in the community can help make them better?

The logic here is that if the current Fintech 2.0 trend continues or runs deeper, we should see primitive, composable and open-source to be the themes of the coming decades. With the traditional financial services moving away from being service-centric to being tech-enabled, we believe that the next wave of Fintech will be tech-centric or driven by developers who see the opportunities to offer financial services in a faster, better and cheaper manner. As long as the incentives are aligned, good developers globally will join forces and accelerate the next wave.


Primitive means the most basic, or in this context means components not developed nor derived from anything else, each component has been broken down as far as possible.

(Source: a16z)

As described earlier, Banking as a service (BaaS) companies typically choose a layer of the banking stack (e.g., compliance, core systems or payments), and provide that layer “as a service”. They help other companies launch new financial services in a faster and cheaper manner; and will provide tremendous value to the Fintech wannabes. But developers globally probably won’t sit still and be satisfied with this outcome yet. The world needs a programmable inventory of every type of lego bricks that could be combined in any permutation. We are only halfway there, and we have reasons to believe that such a vision is feasible since the same transformation actually has taken place in cloud computing. The introduction and application of programmatic and open-source building blocks such as microservices of applications or infrastructure as a service for computing resources are providing developers with maximum flexibility in building modern applications. With more lego bricks to assemble, developers can conduct thousands of re-bundling experiments and enable new services and new experiences.


Slicing banking infra down to financial primitives won’t be useful if we cannot assemble them seamlessly. This comes to the concept of composability. Composability means existing resources can be used as building blocks and programmed into higher-order applications. Composability allows developers to reuse someone else’s work, resulting in more rapid and compounding innovation. To paraphrase Chris Dixon’s tweet, compounding interest is money building on itself, while composability is software building on other software.

Twitter avatar for @cdixon

Chris Dixon @cdixon

Composability is to software as compounding interest is to finance 🧵

October 23rd 2021

303 Retweets1,649 Likes

The challenge with composability is gauging its impact because, by nature, humans intuitively underestimate exponential growth. There is even a term for this phenomenon called the exponential growth bias. If compound interest is easy to quantify but difficult to feel, and software composability is even harder to quantify and to intuit. As such, historically, there has always been a trade-off between open vs close platform and self-build vs outsourcing vs off the shelf software. The ultimate consideration here is control:

  • Control over end-user experience (how a company must ensure its products work all the time to meet user experience requirements); as well as
  • Control over the economics of the supply-side resources (how a company must ensure its economics works especially when its products scale up significantly to serve a broader user base)

Cloud infrastructure is the solution to the high cost of scale and to a certain extent, the high friction of inter-application communication, as it technically abstracts away the hardware constraint and enables a new standard for software build to be more interoperable. More still needs to be done but things are progressing well. It is the trust factor that is difficult to resolve and we think that a trustless mechanism will be the ultimate piece of the puzzle needed to unlock composability. It will enable developers to build atop shared infrastructure without fear of lower-level dependencies.


Open-source is source code that is made freely available for modification and redistribution. The idea of sharing source code on the Internet is not new and dated back to the days when the Internet was quite primitive with software distributed via UUCP, Usenet, IRC, and Gopher. Today the open-source movement often refers to a decentralized software development model that encourages open community cooperation. These communities consist of individual developers as well as large companies. Most notable projects in the tech world include Linux, Git, MySQL, Node.js, Docker, Hadoop, Spark, MongoDB, etc.

In the financial service industry, the open-source model is already making a tremendous impact in several ways:

  • First, many countries (UK, Europe, Brazil, etc.) are implementing open banking regulations which mandate local banks to open up back-end infrastructure (payment, database, etc.). Open-source libraries (like The Open Banking Project) help banks cut short this process without redoing the same work that other banks have done as well as provide code / API for third parties to interact with legacy core banking.
  • Second, many projects are going beyond just opening up but also creating the new standard for specific use cases. Moov uses open-source primitives to help developers embed payment into their software. Mifos is providing open-source core banking for any microfinance business. Red Hat is standardizing the API request process and data flow. Standardization makes the global financial system more connected and accessible.
  • Last but not least, the open-source model is reducing the cost of and the barrier to banking software access, which in turn is powering thousands of fintech companies to offer financial services at cheaper costs and thus driving financial inclusion. Fintech companies unencumbered by legacy banking tech could enjoy lower costs to serve their customers and thus would be willing to lower their service fees.

Why do we need an open-source ecosystem for Fintech 3.0? Because innovation happens faster when the best and brightest all contribute to the common “knowledge goods”. Cost to build will be significantly reduced, hence incentivizing developers globally to develop whatever makes economic sense to them. The community mindset of open-source, in contrast with the control mindset described earlier, will go hand in hand with the other two components of Fintech 3.0 to complete a feedback loop of accelerated innovation.

DeFi at the centre of Fintech 3.0

With the above considerations in mind, we have high conviction that the next wave of Fintech development (aka Fintech 3.0) will be built on top of permissionless, trustless, composable and open-source primitives that are unencumbered by legacy issues. While we can wait many years for this wave to take shape, a solution already exists. It’s called DeFi (or Decentralized Finance).

DeFi is an umbrella term encompassing the vision of a financial system that functions without any intermediaries, such as banks, insurances or clearinghouses, and is operated by smart contracts. No specific institution or individual is needed for the functioning of DeFi, and users can rely solely on the mechanics of the smart contracts, which is open-source and available to all, to process their transactions.  Current DeFi development has already covered the key services that have been typically offered by TradFi, such as payments, lending, insurance, exchanges, money as well as central, commercial and investment banking services.

(Source: Consensys)

So how does DeFi work? According to Multicoin Capital, DeFi can be viewed as a stack of 6 distinct levels. The levels are distinct yet composable and can be built on each other. We will provide a brief summary of each layer below, but do check out their article for a detailed breakdown.

(Source: Multicoin Capital)

  • Level 1: Units of Value. There is a need for assets to serve as collateral for derivatives, loans and leverage. In DeFi, these atomic units of value can be ETH, money market tokens or stablecoins, such as USDT and USDC.
  • Level 2: Transaction Layer. With the atomic units of value in hand, users must be able to make transactions on-chain for DeFi to work. For this, it requires a layer-1 or layer-2 protocol with sufficient capacity to process the transactions. Examples of protocols include Solana (which we had covered previously), Ethereum, Avalanche, etc.
  • Level 3: Oracles. Secure and verifiable inputs of market data are critical to the functioning of DeFi protocols. Price oracles serve to provide financial primitives with accurate data and enable them to trigger certain events, such as liquidations. The most popular oracles today include Chainlink, Band, Nest and Coinbase.

Levels 1, 2, and 3 comprise the core infrastructure of DeFi. Developers then build on these levels to create sophisticated and interoperable financial primitives, which will form levels 4-6.

  • Level 4: DeFi Primitives. This is a level that many will be familiar with, as it refers to the DeFi applications that users interact with. It includes lending protocols, trading pools, order book exchanges, derivatives networks and asset management platforms. The primitives can also be thought of as a network, as each primitive can be used independently or in conjunction with other primitives to meet the needs of the users. Notable examples include Compound, Aave, Curve, Serum, dYdX and Set.
  • Level 5: Aggregators. Aggregators live on top of primitives, and there are both supply-side and demand-side aggregators. Aggregators do not custody collateral assets – instead, they often provide smart contract instructions that enable users to interact with other DeFi applications. Aggregators include Yearn Finance, 1inch and Paraswap.
  • Level 6: Wallets / Front Ends. The final level is wallets and other front ends (including relayers), which users rely on to store their units of value and engage DeFi Primitives. Notable examples include MetaMask, Phantom, Coinbase and Ledger.

Advantages of DeFi

We will discuss the use cases of DeFi (i.e. Level 4-6) in greater detail in Part 2 of this series, so stay tuned! In the meantime, with the workings of DeFi stack in mind, we would like to understand how DeFi measures up to TradFi. The table below summarises the features of both systems.


As seen above, DeFi has the potential to fulfil the promise of Fintech 3.0 and address the issues that TradFi currently faces. Given the advantages that DeFi brings, it is a model that financial systems around the world may move towards in the future, though their trajectories may be slightly different. In developed regions such as the US and Europe, they are currently in the midst of a transition from Fintech 2.0 to Fintech 3.0, where DeFi features prominently in the movement. Adoption of DeFi primitives may thus come slightly more seamlessly for these regions and we expect it to gain mainstream popularity soon.

On the other hand, for developing regions including Southeast Asia, we note that Fintech 2.0 has only just started to take shape, as the ecosystem is less mature. Nonetheless, given the high crypto adoption rates in the region, we may observe the region leapfrog into Fintech 3.0 directly, instead of waiting for the Fintech 2.0 wave to run its course. It is a scenario that could mirror that of e-payments previously – while developed countries had transitioned from cash to debit/credit cards to contactless payment (e.g. use of QR codes), some developing countries jumped straight to contactless payment from cash, bypassing debit/credit cards.

While we recognise the merits of DeFi, DeFi is not without its issues as well. In the next part of this DeFi series, we will cover the main verticals of DeFi as well as dive deeper into the key trends and concerns surrounding the sector. For Part 3 of the series, we will discuss the views of the DeFi community on what lies ahead for the sector, and their proposed solutions to address the issues that we had raised above. There will be interviews, exciting new ideas and more in-depth analysis of DeFi, so be sure to check out the articles when they are released!