Author: Stani.eth

Translation: Deep Tide TechFlow

Deep Tide Introduction:A quantitative investment paper personally written by the founder of Aave: He individually calculated the capital expenditures of solar energy, data centers, robotics, nuclear energy, and space infrastructure, concluding that the real market size facing DeFi is $100-200 trillion, which is 15 times that of the total asset management scale of the world's top ten banks. Whether the argument holds true is another matter, but this analysis framework itself is worth serious reading for anyone paying attention to the RWA track.

The full text is as follows:

I previously wrote that DeFi has improved the supply side of the capital allocation problem. On-chain liquidity is highly liquid and can be programmatically shifted to opportunities with higher risk-adjusted returns. Aave has proven its ability to absorb tens of billions of dollars in liquidity, thanks to the trust it has built over many years and the superior cost structure it provides in the crypto lending model.

This liquidity creates huge opportunities for the financial foundational primitives and use cases currently emerging. The next evolution of DeFi should focus on demand-side issues, rebalancing liquidity equilibrium.

I also previously wrote that solar energy alone could present Aave with a $30-50 trillion opportunity. But this is far from the endpoint; the future opportunities Aave can tap into exceed $200 trillion.

Infrastructure that supports everything operation, financing everything

At the base level, infrastructure is what runs everything. This layer ensures that our electric vehicles have range to drive, our homes stay warm and lit, water flows properly, computers can compute, and the world stays connected.

From the perspective of capital allocation, infrastructure is seen as a safe choice. The world needs energy, water, computing power, and communication. Established infrastructure also carries the technological risk that diminishes with time and scale, gradually transitioning from a technical opportunity to a financial opportunity as it matures.

While seen as stable and secure, emerging infrastructure also provides high-return allocation opportunities. Technology is at an early stage of the cost curve, and the risk premium is correspondingly richer.

Infrastructure (the right type) is a premium product in finance because it generally involves high capital expenditures needing financing and low operating expenditures—this means that operating costs are low enough that debt can be repaid over the asset's lifecycle. From many angles, infrastructure for future financing is hard assets with cash flow.

Most importantly, when the structure is designed correctly, infrastructure finance follows Aave's lending model—lending against the asset itself rather than the user's credit, similar to how Aave operates today.

How big is the opportunity?

I believe that critical infrastructure assets essential for the world's transition to abundance include: solar farms, batteries, data centers and GPUs, electrified transportation, robotics, desalination, mineral extraction, carbon capture, nuclear energy, and space infrastructure, among others. If anything is missing from this list, it's not due to a lack of confidence in it, but rather to underscore how broad this category is. As long as an asset qualifies as an "abundant asset" and is not infrastructure heading towards decline, it is likely a reasonable choice.

Solar and batteries: Solar alone represents $15-30 trillion in capital expenditures needing financing. At this scale, solar is projected to replace fossil fuels by 2050. This is detailed in my previous articles.

Data centers and GPUs: Cumulative capital expenditures for GPUs and data centers range from $15-35 trillion, depending on the sensitivity to AI adoption rates. McKinsey estimates that $6.7 trillion in capital expenditure will be needed by 2030 alone. My consistent logic is: if you provide more computing power, computers will compute more and take on more complex tasks. Moore's Law has its limits, but we are entering the realm of atomic engineering and vertical layering of GPUs, even though these advancements won't prevent us from building more computing power. This does not take into account quantum computing, which may drive another round of distributed expansion in computing capacity.

Robotics: The automation of human tasks will become a defining feature of the world we are transitioning into. Robots, whether specialized warehousing systems or humanoid robots for everyday physical tasks, will replace human labor and grant us more freedom. By 2050, just robotics may require $8-35 trillion in capital expenditures.

Electric vehicle infrastructure: Electrification of transportation (electric vehicles, rail, airplanes, drones, charging networks, ships, ports) is on the verge of a massive transition from fossil foundations to electric-based systems. Capital expenditures are expected to reach about $10-25 trillion by 2050. Autonomous driving itself will ensure that no vehicles remain idle and will serve society around the clock.

Nuclear energy: I have a love-hate relationship with nuclear energy as an abundant asset. It is a reliable choice for generating substantial energy. However, nuclear energy is deeply tied to policy, making it difficult to innovate and even harder to finance. Projects often turn out to be more expensive and time-consuming than expected. Small modular reactors and better policy frameworks may change this situation. My estimate for capital expenditures by 2050 is relatively conservative at $3-8 trillion, constrained by policy.

Solar-driven desalination: Desalination is not new. It has existed for decades, with Middle Eastern countries relying heavily on it. It remains costly, but with economies of scale and the development of solar energy, we will be able to access almost free water anywhere in the world. Required capital expenditure: $6-12 trillion by 2050.

Carbon capture: Growth will be driven by government incentives. Estimated capital expenditures: $3-8 trillion by 2050.

Key minerals: Copper, lithium, nickel, and rare earths that drive electrification, robotics, and more. Estimated capital expenditures: $5-15 trillion by 2050.

Digital networks: Fiber optics, communication towers, and satellite ground stations. Estimated capital expenditures: $6-15 trillion by 2050.

Space infrastructure: Space as a scale factor will grow significantly as transportation and launches enjoy economies of scale. Space will become an investment opportunity in infrastructure over the coming decades. A conservative estimate for capital expenditures by 2050 is $2-6 trillion, but this figure could carry a higher multiplier. If launch costs drop by 10-50 times along historical cost curves, opportunities could expand to $10-30 trillion, and in extreme cases, reach $50 trillion. This includes: satellite constellations $3-8 trillion, launch infrastructure $1-3 trillion, orbital infrastructure (fuel tugs, service stations, essentially orbital logistics hubs) $2-7 trillion, space-based solar power $2-10 trillion, space manufacturing $1-5 trillion, and lunar infrastructure $1-5 trillion.

I skipped hydrogen production because I am unsure how it will evolve in the broader electrification transition.

In total, financing infrastructure could bring about $100-200 trillion in opportunities for DeFi. For reference, the world's top ten banks manage about $13 trillion in assets. Successfully financing a substantial portion of this transition will make Aave the largest financial network to date.

Selecting the Right Form for Aave

Infrastructure financing in DeFi can take two main forms.

Path One: Yield-bearing Stablecoins (YBS)

YBS is becoming a strong example of distributing off-chain income to on-chain users. Ethena achieves this primarily through basis trading; USD.ai does so through GPU financing. The annualized yield for staking sUSDai is 10-15%.

From Aave's perspective, the growth of YBS directly translates into protocol growth. Aave is a circular machine: if the yield from YBS infrastructure products exceeds Aave's cost of funds (about 4-5%), there is a circular opportunity: borrow liquidity from Aave using YBS as collateral and redeploy it. I view YBS as an on-chain yield allocation wrapper, sharing characteristics similar to traditional off-chain funds.

Path Two: Direct Collateralization

Using tokenized infrastructure directly as collateral means that yields or economic benefits remain off-chain or with the borrower, but through collateralization and borrowing demand flows into Aave, generating yield for depositors by supplying stablecoins. This path also does not aim to achieve stable net asset value, making it very suitable for assets where net asset values fluctuate and cannot be tested through stablecoins.

Which path will prevail? It's hard to say. Each has its advantages, and Aave has supported both models well. Examples of YBS include Ethena's sUSDe and Maple's SyrupUSDT. Examples of direct collateralization include Tether's gold (xAUT), Bitcoin and Ethereum mortgages, as well as the JAAA RWA fund—in this fund, the underlying economic benefits belong to the asset owners in exchange for interest paid indirectly to on-chain depositors through Aave. Notably, Aave's own aToken (like aUSDC) is, in a sense, the earliest form of on-chain YBS in this category.

The two paths depend on user types. The former's user profile may consist of on-chain allocators maximizing YBS yields; the latter's borrowers may be operators or funds looking to expand liquidity allocation and build more infrastructure to scale without directly demanding on-chain yield distribution.

Is the yield sufficient?

While DeFi does have excess capital in the current interest rate environment, infrastructure financing should provide ample upside potential to absorb this capital. Average equity internal rates of return across fields include: solar 10%, batteries 12%, data centers 13%, electric vehicle charging infrastructure 13%, water infrastructure 9%, and space infrastructure around 18%. The higher the technological risk and the earlier it is in the cost curve, the higher the assumed returns.

Yields can be further enhanced through strategies. Vaults on Aave V4 can be allocated to solar farms with yields of 8-12%, using that asset as collateral to borrow GHO (creating high-yield profit space for Aave), then redeploying GHO into battery farms with yields of 12-18%, or even opportunities in GPU data centers with annualized returns of 10-20%.

DeFi users are typically sensitive to redemption risk and lock-up periods (this may change in the future as the sector matures). Infrastructure products usually produce cash flow, helping to mitigate redemption risk. Using Aave as a liquidity conduit can make these products more accessible to users—allowing users to supply liquidity to specialized hubs focused on these specific economic characteristics and trust assumptions, ensuring access to infrastructure opportunities while isolating and controlling risk. A key distinction is that directly tokenizing the assets themselves allows for auction-based liquidation, improving liquidity characteristics of these assets compared to slowly layered debt-wrapped funds.

Aave as the Financial Infrastructure Layer

The best path for Aave into RWA and infrastructure opportunities is to serve as the base layer of financing liquidity, starting from the less risky, mature end (solar energy) and then gradually moving towards riskier assets with the fine risk control provided by the Aave V4 hub's radiating architecture.

Today, most RWA tokenization focuses on assets with existing deep liquidity markets: treasury bills, money market funds, corporate credit. These assets trade smoothly, and users already have enough channels to borrow these assets. Likewise, while private credit may superficially seem like a compelling DeFi use case, it also has downsides. Private credit is typically for CLOs, corporate, and private equity financing. If the infrastructure layer is what I call the base layer, then this is the top layer. In a world transitioning faster than ever before, especially at the top layer, assets need to tilt towards the future we are building and not towards the past we are leaving behind. A seemingly outstanding asset-backed financial product may shine on paper but lose its positioning in the world of tomorrow.

Tokenization of traditional financial assets will continue to grow and will certainly be a part of the Aave story, just as crypto-native assets and their growth continue to exist. But the bigger opportunity lies in becoming the future infrastructure financing layer. This is what excites me about RWA and Aave.

What does it mean for fintech companies?

Large fintech companies are increasingly becoming the distribution and experience layer: interfacing quality financial products to end users. I have previously written that leveraging DeFi allows fintech companies to unlock a more streamlined cost structure for new financial products. DeFi operates almost autonomously, is more transparent, and guarantees execution through smart contracts. It requires less operating expenditure, enabling tighter profit margins and opening up new financial opportunities.

In a world where financial access becomes commoditized and no longer offers differentiated value propositions, the ability to secure unique yield opportunities brings new value to fintech companies (and even banks) and their users. Fintech companies actively participating in the stablecoin issuance space may also uncover new use case opportunities, including real lending demand for loans secured by infrastructure collateral.

Fintech companies and banks can become the perfect distribution channels for the yields generated from infrastructure collateral on Aave V4 via Aave Kit and Aave App, with these collateral elements closely tied to the future we are building. Integrating Aave into fintech companies and banks to inject capital could accelerate the transition to an abundant world by 10-15 years. This presents Aave and its various integration partners with a unique opportunity to capture and share $200 trillion in market value.