Stablecoins, real-world asset (RWA) tokenization and the AI Agent Economy — viewed through a financial lens.
An introductory report for investors, product builders and finance professionals. For educational purposes only. This is not investment advice.
Digital assets are moving beyond speculative narratives. The most important development is not a new wave of crypto trading, but the emergence of a new layer of financial infrastructure at the intersection of traditional finance and blockchain: stablecoins, tokenized financial assets and programmable settlement processes. For this reason, digital assets should be studied first as infrastructure, and only second as an investment trend.
Major institutions are now paying close attention. Citi Institute estimates that tokenized financial assets currently total around USD 17 billion and could reach USD 5.5 trillion by 2030 in its base case. This suggests that tokenization is increasingly seen as part of the future financial market infrastructure, rather than an experiment confined to the crypto community.
Stablecoins are closer in nature to money than to an investment asset. Within the digital asset ecosystem, stablecoins act as a means of payment rather than a speculative asset. If assets move on-chain while cash flows still run through traditional banking channels, much of the benefit in settlement speed, automation and programmability is significantly reduced.
Tokenization reduces the limitations of traditional finance but does not create magic. It can improve settlement, record-keeping, transferability and automation — but it does not by itself create buyers, cash flows, legal rights or secondary liquidity. If the underlying asset is illiquid, opaque or legally unclear, the token will inherit all of those problems.
AI Agents may become a new class of economic actors. As software begins to search, compare, shop and pay on behalf of people within pre-approved limits, a hybrid model is more likely than a purely crypto world: everyday consumption continues through apps, cards and wallets, while tokens and stablecoins prove useful where programmability, instant settlement and automated escrow are required.
Many debates around digital assets begin with a basic problem: the word “token” is used to describe many very different concepts. A token can be a speculative coin, a stablecoin, a tokenized security, a cash-flow claim, a governance right, or even a loyalty point recorded on a blockchain.
Before discussing the market, investors need a minimum shared vocabulary:
| Concept | Plain-language explanation |
|---|---|
| Digital asset | Any asset, right or unit of value represented in digital form. This includes pure crypto assets, stablecoins, tokenized assets, NFTs, utility tokens and economic rights recorded on a blockchain. |
| Blockchain Distributed ledger | A shared ledger that records transactions across a network. Its core value is not simply “data storage,” but the ability to record, transfer and verify ownership or rights in a programmable way. |
| Token | A digital unit that represents a right or a function. It can be used for payment, voting, access, ownership, sharing in cash flows or using a service. |
| Stablecoin | A type of token designed to maintain a stable value relative to a reference asset, typically the US dollar. In practice, a stablecoin behaves more like a means of payment than a speculative asset. |
| Real-world asset tokenization RWA tokenization | Representing the economic rights tied to real-world assets — such as bonds, real estate, gold, private credit, commodities or invoices — in the form of tokens. |
| AI assistant / agent AI agent | An AI system that can understand a goal, plan actions, use tools and carry out tasks within a delegated scope. Unlike a chatbot, an AI agent does not only respond; an AI agent can act. |
The first principle of analysis is to ask: what does this token actually represent? Without that answer, most of the rest of the discussion is largely noise.
Tokenization is not about moving every asset onto the blockchain overnight. The more realistic path is a long hybrid phase, in which traditional financial infrastructure and tokenized systems operate in parallel.
The first assets likely to scale are those with four characteristics: a high degree of standardization, legal clarity, available liquidity and a clear settlement problem to solve. This points toward government bonds, money market funds, listed equities, repo collateral and institutional fund units — before highly customized assets such as individual real estate projects or operating private companies.
Citi’s base case points to a tokenized financial asset market of roughly USD 5.5 trillion by 2030, with a bear case of around USD 2.7 trillion and a bull case of around USD 8.2 trillion. These are scenarios, not certainties, but they indicate the direction of institutional capital.
Tokenization only creates value when it solves a real market problem. The most credible use cases are those that can:
Without a real problem to solve, “putting an asset on the blockchain” simply adds another layer of technology without improving the economics of that asset.
Digital assets can be divided into two broad layers:
Asset layer: pure crypto assets, tokenized securities, tokenized funds, tokenized real estate and tokenized claims on cash flows.
Money layer: stablecoins, tokenized deposits and other forms of regulated digital money.
Stablecoins matter because most on-chain transactions need a unit of payment with relatively stable value. If the asset is digital but the payment leg still depends entirely on the traditional banking system, the transaction is only half digitized. The real infrastructure breakthrough occurs when both the asset and the money move on the same programmable layer.
This is what makes models such as delivery-versus-payment, automated escrow, 24/7 settlement and machine-initiated payments more feasible.
| Type | Nature | Central question |
|---|---|---|
| Pure crypto assets | Assets created inside the blockchain ecosystem, such as Bitcoin or Ethereum. | Does it represent any right outside the network, or is its value entirely determined by the market? |
| Stablecoin | A token pegged to a reference asset, typically the USD. | What backs it, who issues it, how does redemption to cash work and how is it regulated? |
| Tokenized assets | A traditional asset or economic right represented by a token. | What is the legal wrapper, who custodies the asset, what rights does the token holder have and how does liquidity work? |
This distinction matters because their risks are entirely different. A stablecoin mainly faces questions about reserve assets, redemption capacity and issuer quality. A tokenized asset mainly faces questions about legal rights, custody, valuation and liquidity.
Imagine a morning in 2028. On your way to the office, you tell your personal AI assistant:
“Order me a latte with 50% ice and 30% sugar, under VND 80,000, delivered to the lobby before 8:30.”
You do not open an app. You do not compare stores. You do not enter an address. You do not search for discount codes. The AI assistant reads your schedule, location, delivery window, past preferences and current offers across Starbucks, Grab and ShopeeFood, selects the optimal option and pays within the spending limit you have approved — through a supplementary credit card with a limit you set specifically for the AI (much like a supplementary card for your child).
The transaction is mundane: a cup of coffee. But the economic structure has changed. The human sets the goal and the constraints; the AI carries out the research, comparison and execution.
Citi describes this as part of the “Do It For Me” economy. In the early stages, the agent may only suggest actions and wait for human approval. As trust grows, users may allow the agent to execute low-value, repetitive and low-risk transactions within predefined limits.
The key point is that not every agent transaction needs a blockchain. For domestic consumer payments, cards, e-wallets, QR payments and virtual cards may remain the dominant channels.
Tokens and stablecoins become more relevant in specific contexts: 24/7 settlement, cross-border payments, programmable escrow, repetitive micropayments, machine-to-machine commerce and software-initiated payments that require automated verification.
| Scenario | Role of token / stablecoin |
|---|---|
| Coffee and food delivery | Traditional app-based payment is likely to remain dominant. Stablecoins only become useful if merchants or platforms support them. |
| A birthday gift for a spouse | The AI can choose based on the recipient’s profile, budget, delivery timing and return policy. Smart contracts or escrow may be more appropriate for high-value or cross-border transactions. |
| Flights and hotels | The agent can compare prices, flight times, baggage rules, loyalty points and cancellation risk. Traditional payment remains the front-end interface, while agent-to-agent settlement may develop in the back end. |
| Digital services & APIs | The agent can autonomously buy data, compute, plugins or API access. Micropayments in tokens/stablecoins are useful because they are small, repetitive, automated and 24/7. |
The AI agent economy does not mean “everything will run on crypto.” A more realistic outcome is a hybrid model: everyday consumer payments still rest largely on apps, cards and wallets, while tokens and stablecoins move into the layers that require programmability, instant settlement and automated escrow.
RWA tokenization is attractive because many real-world assets are hard to access, fractionalize, transfer and manage. Real estate, private credit, commodities, invoices, investment funds and contractual cash flows can all potentially be packaged into economic rights and distributed through digital infrastructure.
A typical tokenization structure requires four components:
The biggest misconception: that tokenizing an illiquid asset will automatically turn it into a liquid one. It will not.
Tokenization can reduce operational friction, but it cannot by itself create buyers, cash flows, reliable valuation or legal certainty. The token is only a layer of data recording. The real economics still depend on the contracts, custody arrangements, legal enforceability, quality of disclosure and dispute-resolution mechanisms.
In other words, the quality of a tokenized asset is limited by the quality of the underlying asset and the legal structure behind it.
A serious analysis of digital assets should start with risk, not with yield. The five questions below help filter out many weak products:
What does the holder actually own? Ownership of the asset, a cash-flow claim, a debt claim against the issuer, a redemption right, or simply economic exposure? If the legal rights are unclear, the token’s market price may matter less than the contract behind it.
Where is the underlying asset held, who controls it, and is it separated from the issuer’s balance sheet? For stablecoins: reserve assets, audits, redemption mechanisms and issuer quality. For tokenized assets: custody, legal segregation, verification and protection in the event of bankruptcy.
Can the token actually be traded or redeemed for cash? A listed token does not automatically have liquidity. It is necessary to assess actual trading volume, market makers, transfer restrictions, lock-ups, whitelist rules and redemption mechanisms.
How is the asset valued, and who supplies the data? If the underlying asset is off-chain, there must be a mechanism to bring information on-chain — creating oracle risk. It is important to understand how NAV is calculated, how often prices are updated, who verifies the data and what happens when the data is wrong.
Which national legal system governs this product? Can investors participate legally? Where are disputes resolved? Could the token be treated as a security? Which KYC, AML and suitability rules apply?
A credible RWA product needs more than a wallet address: a clear legal structure, verifiable backing assets, real liquidity, transparent valuation and a compliance process appropriate for the target investor base.
Before treating any tokenized asset as a serious investment product, ask the following questions:
| Area | Question |
|---|---|
| Asset | What is the underlying asset? Where do the cash flows come from? Is it pledged, disputed or subject to legal encumbrances? |
| Legal rights | Does the holder own the asset, receive the cash flows, hold redemption rights, or only have a debt claim against the issuer? |
| Custody | Who holds the asset? Is there an independent custodian, audit or attestation? Is the asset ring-fenced? |
| Liquidity | Is there a secondary market? Can it be redeemed for cash? Is there a market maker? Are there transfer restrictions? |
| Valuation | How is NAV or price calculated? Who values the asset? How is off-chain data verified? |
| Compliance | Which jurisdiction’s law applies? Who is eligible to invest? Which KYC/AML rules apply? Where are disputes resolved? |
If you cannot answer five basic questions — what do I own, where does the real asset sit, how do I exit, how is the price determined, and where are disputes resolved — then the product should not yet be treated as a serious, institutional-grade investment opportunity.
This market is not yet a major hub for tokenized capital markets, but the need for education about digital assets is growing. This is especially true for people who are already familiar with equities, real estate, banking and traditional financial products.
That creates a clear gap: structured research, education and risk-analysis tools written in clear, accessible language.
AIVA approaches digital assets from a research and risk-management perspective, not from the perspective of token issuance, exchange operation, asset custody or investment recommendations. The immediate goal is to help readers understand stablecoins, RWA tokenization and the AI Agent Economy in financial language, while building checklists that help identify risks before participating.
Over the longer term, if user demand is clear enough, AIVA could grow into an analytical support layer connecting traditional financial markets and digital assets. The appropriate role of AI is to support reading reports, identifying risks, comparing strategies and monitoring portfolios within a clear control framework — not to fully replace human judgment.
For financial assets, the default principle is that a human must always remain in the control loop. AI can support research and decision-making, but it should not autonomously execute financial transactions in the absence of a clear legal framework, identity verification, authorization limits and risk controls. As software begins to transact on behalf of people, the infrastructure for money, assets, identity and risk control must also become more programmable.
This report is an introductory educational document. It synthesizes public sources and reinterprets them through a finance-first lens for its readers.
| Source | Use |
|---|---|
| Citi Institute — Tokenization 2030: Wall Street On-Chain (June 2026) | Market size, 2030 scenarios, asset composition and the role of stablecoins. |
| Citi Institute — Agentic AI: Finance & the “Do It For Me” Economy (January 2025) | The assistant economy, AI wallets, agentic commerce and the AI × crypto intersection. |
| Publications from BIS, FSB, IOSCO and the IMF | Policy context, systemic risk, disclosure, settlement money and investor protection. |
| AIVA synthesis | Interpretation for the market, the risk checklist and the implications for products. |
Market figures are cited per their respective sources. Charts are redrawn for educational purposes and should not be treated as precise forecasts. The views in this report are AIVA’s interpretation and should not be considered investment advice.
AIVA Digital Assets Research provides research and educational content only. This is not investment advice, financial advice, legal advice, brokerage, custody, token issuance, trade execution, or distribution of financial products. Digital assets carry significant risk. Readers should do their own research and seek professional advice where needed.
AIVA Digital Assets Research provides research and educational content only. This is not investment advice, financial advice, legal advice, brokerage, custody, token issuance, trade execution, or distribution of financial products. Digital assets carry significant risk. Readers should do their own research and seek professional advice where needed.