Interview with AIUSD Co-Founder: From Trillion-Dollar Multi-Year Live Trading Strategies to AGI Market Economy, the First "Money" Designed for AI
TechFlow Selected TechFlow Selected
Interview with AIUSD Co-Founder: From Trillion-Dollar Multi-Year Live Trading Strategies to AGI Market Economy, the First "Money" Designed for AI
Let's explore the specific forms of Post-AGI finance, as well as AIUSD's core competitive advantages and concrete evolutionary path under the central vision of "AI-driven financial intelligence."
Navigating Web3 tides with focused insightsBy: TechFlow
You can tell a horse from a donkey by taking it for a walk.
Over the past decade, the crypto industry has shifted from "whitepaper theorizing" to "proving performance on-chain," with markets increasingly favoring projects that deliver stronger real-world results.
On November 19, 2025, an Agentic AI monetary platform designed around live trading from day one announced it had raised nearly $10 million in Pre-Seed funding, backed by top-tier Silicon Valley investors including early institutional backers of star projects like Anthropic, Sequoia US Scout Fund, a16z Scout Fund, and leadership from Tesla FSD AI.
This is AIUSD: a mission to enable both humans and AI to "talk to money" through natural language, where systems understand intent, decompose tasks, and automatically manage unified asset routing, trading, and settlement across all major blockchains, centralized exchanges, and mainstream stablecoins—truly democratizing blockchain and advancing global finance into the era of "intelligence."
At this moment of funding announcement, TechFlow sat down with Yao Meng and Bill Sun, co-founders of AIUSD, for an in-depth conversation.
Discussing AI's role in finance—and particularly in the niche of crypto—one topic was the recent trading competition among the world’s six leading AI models. On this match, the two guests敏锐ly pointed out the fundamental flaw: “AI lacks a completeexecution environment”, and expressed eagerness for AIUSD to join such contests. Yao stated:
The AIUSD Agent doesn’t just offer opinions—it’s directly connected to real liquidity, funding rates, cross-venue execution paths, and clearing systems. If future competitions allow real execution, we’d be very willing to participate.
Confidence in AIUSD’s practical capabilities stems from multiple fronts. Bill excitedly described the current moment as a “rarewindow period” for successfully bringing AIUSD to market:
On one hand, AI is evolving from “conversational” to “executable,” while crypto infrastructure is finally maturing;
On the other, AIUSD’s underlying funding rate engine, microstructure execution engine, and cross-scenario routing system have been running live internally for over two years, achieving an annualized trading volume of around $1 trillion, with core strategies delivering average annual returns above 20%, a Sharpe ratio around 22, and no monthly drawdowns. We’ve also built robust risk management, compliance frameworks, and automated controls, ensuring this is a system truly capable of handling capital.
Facing AIUSD’s imminent launch, Yao shared a clearer roadmap:
AIUSD has successfully passed the stage from 0 to 1: 0 meaning proving technical feasibility, and 1 meaning turning it into an account layer anyone can use—just speak, and the system handles routing and settlement. Now, AIUSD stands at the inflection point from 1 to 10. Our next goal is to reliably deliver this same infrastructure to a broader consumer and developer ecosystem.
In this article, let’s follow the insights of these two Wall Street/crypto OGs to explore the concrete form of Post-AGI finance, and examine AIUSD’s core competitive advantages and evolutionary path under its vision of “AI-driven financial intelligence.”
AIUSD: Money That Both Humans and AI Can Easily Understand and Use
TechFlow: Thank you both for your time. Could you please start with brief introductions?
Yao:
Hello everyone, I’m Yao. Glad to be here.
I entered the cryptocurrency world in 2011 during university, initially mining, then gradually participating in various blockchain projects and early exchange development. Over the past decade, I’ve lived through key industry shifts—from Mt. Gox claims, FTX collapse, to the recent 10.11 liquidation event. For example, during the May 19, 2021 mass liquidation, we were wiped out overnight with a $50 million position, leaving many investors bankrupted. Last month’s 10.11 crash, our firm was likely the first large institution to dump at scale. And during last year’s election, I heavily positioned in Dogecoin—the company’s proprietary arbitrage strategy held about 20% of all Dogecoin positions in the market at one point, achieving up to 200% annualized returns, which was extraordinary for a delta-neutral strategy.
Throughout these years, I’ve remained focused on trading—arbitrage, high-frequency systems, institutional custody, and quantitative strategies. Now, launching AIUSD, I aim to repackage the system capabilities we’ve accumulated into a stablecoin infrastructure that both AI and humans can naturally use.
Bill Sun:
Hello everyone, I’m Bill (Qingyun) Sun. Great to be here.
I studied mathematics as an undergrad and PhD student at Stanford. When I started my PhD in 2014, deep learning was just gaining momentum. I became obsessed with understanding the mathematical structures behind these models and testing their viability in real, complex environments like finance. From the start, my research focused on two things simultaneously: the underlying principles of deep learning and their practical applications in finance.
In 2016, I worked on NLP at Google Brain, before the term Transformer was coined. We experimented with attention-based reading comprehension and Q&A, modifying model architectures and testing different tasks. What we were doing later became known externally as the Transformer—making me part of the earliest explorers of this path.
Back at Stanford, I continued researching the mathematical foundations of deep learning with two advisors: David Donoho, dual member of the U.S. National Academies and Gauss Prize winner, a leader in high-dimensional statistics and compressed sensing; and Stephen Boyd, U.S. National Academy of Engineering member, a seminal figure in optimization and founder/lead of BlackRock AI Lab. Both are deeply involved in quant finance: Donoho previously researched at Renaissance, Boyd built and leads BlackRock’s AI Lab. Working with them gave me early exposure to what problems truly matter in real markets and which model traits survive beyond academic papers.
During my PhD, I also did quant research at Citadel and Point72 Cubist, applying deep learning and reinforcement learning to stock and futures trading. After graduation, I joined Millennium WorldQuant on Wall Street as a fund manager, trading global equities and managing a sizable U.S. equity statistical arbitrage portfolio.
I got into crypto early too, around 2015. I took Dan Boneh’s cryptography course at Stanford and worked on projects at the Stanford Bitcoin Lab with Balaji and Lily Liu, exploring Lightning Network micropayments, early Ethereum smart contract designs, and privacy mechanisms in Zcash and Monero. I also made early attempts at Bitcoin fee optimization and generating stablecoins via on-chain lending.
I strongly believe in crypto’s application within AI, particularly agentic economy. Since last year, I’ve served as Co-founder and Chief Scientist at PIN AI, building an agentic economy from a consumer data autonomy and personalized AI voice+recommendation angle, translating human intent into AI actions.
Over the past decade, I’ve grown increasingly convinced that the boundaries between AI, crypto, and traditional finance are rapidly dissolving. Traditional finance brings mature efficiency and risk control, crypto reshapes infrastructure, and AI transforms everything into truly “intelligent” systems. Together, they’re rewriting the foundational logic of global finance.
AIUSD was born from this context and long-term thinking. I want to create a form of “money” that feels natural to humans and is natively understandable and callable by agents, capable of automatically routing, settling, and managing risk across multi-chain, multi-market environments. Simply put, through this new monetary form, we aim to move global finance from “automation” to true “intelligence.”
TechFlow: Both of you are very early participants in crypto. Could you share why you decided to launch AIUSD—an AI x Crypto project—and what problem it aims to solve?
Yao:
After 14 years in the space, my biggest takeaway is: blockchain has created massivewealth effects but failed to achieve true “inclusivity.”
The barrier to entry is too high. I don’t lack money—I want everyone to use digital currency. If this could be my legacy, I’d be happy. Yet after all these years, active on-chain users are mostly airdrop farmers and gamblers, not everyday users benefiting from accessible applications. My constant question: how can ordinary people—like your 50-year-old parents—use crypto? They use WeChat and Taobao effortlessly.
We should start with wallets and reinvent the entire crypto trading experience. But frankly, after 14 years, there’s still no multi-chain, one-click trading wallet on the market—a painful reality. The reason is simple: only exchanges are profitable in crypto, so all resources flow there. Wallets, despite being the largest gateway, remain poorly developed.
For example: I want to buy a token on BSC, but my funds are on Solana. I must swap SOL for gas, then BNB for gas on another chain, bridge assets across chains, calculate slippage—all requiring wallet switching, confirmation waiting, and fee handling. For average users, this is nearly “unusable.” Suppose your user isn’t a Stanford PhD or master’s student but a taxi driver—why should they endure such complexity? This prevents products from entering real life.
AI offers a breakthrough. AI excels at understanding intent, making decisions, and automating complex processes. AIUSD leverages AI to abstract away these tedious operations, enabling automatic cross-chain routing, settlement, and yield distribution.
In short, AIUSD’s core is a unified wallet layer enabling both AI and human users to perform financial operations without gas fees orcross-chainbarriers.
We believe AI will be the next primary interface: future money won’t be “manually clicked” by people, but understood by AI, which then automatically finds optimal paths, executes settlements, generates yield, and manages risk. So AIUSD aims to solve the problem of making money “AI-native.” You won’t need to understand ten chains or twenty token standards—just tell an AI “send $1,000 to a friend” or “buy some BTC and put it in a stable yield pool,” and the AI handles routing, trading, and settlement.
We often say AIUSD wants to be “One AI to rule them all”—a unified infrastructure allowing all people and agents to naturally use funds.
Not everyone needs to become a blockchain expert. A single sentence should let the system handle everything. This is the direction we’ve believed in from day one: abstracting complex financial systems into instructions callable via natural language.
This project continues what we’ve always done: making capital operations programmable, verifiable, and reusable. Previously it was HFT bots; now it’s AI agents. The underlying logic remains unchanged—only smarter and more abstract.
Bill Sun:
I entered crypto in 2015, witnessing the full transition from ideological infancy to infrastructure maturity. Over the years, I’ve formed two clear convictions.
First, stablecoins and RWA will be the critical gateways for real-world liquidity to enter on-chain. Regardless of market volatility, demand for these assets is most stable and closely tied to real economic activity.
Second, to make on-chain money truly usable, we must completely abstract away operational complexity. For ordinary users and institutions alike, managing addresses, selecting chains, changing RPCs—none of these should be barriers.
I believe stablecoins should ultimately become First-Class Citizens in the blockchain world. Today, stablecoins are still “second-class citizens” beneath assets like ETH, SOL, or BNB, which serve as first-class Gas tokens. The future should reverse this: when users hold USDC or USDT, they shouldn’t need to know which chain it’s on—just as transferring dollars between Chase, Interactive Brokers, or Robinhood doesn’t require awareness of underlying systems.
Today’s multi-chain structure resembles multiple religious nations—separate and closed—with risky, complex bridges between them. A key goal of AIUSD is to abstract away this fragmentation, making using money on-chain as natural as in the real world.
From an AI perspective, I’ve long felt the industry lacks a truly Machine-Native currency. It should support micropayments, be precisely executable via API or function calls, and ideally allow models to generate deterministic DSLs, expressing fund flows at the code level. That’s why we merged two goals:
-
On the crypto layer, unify multi-chain, multi-pool, multi-asset experiences into the “money” layer;
-
On the AI layer, make money something machines can directly understand and invoke.
Ultimately, I hope AIUSD becomes not just an M0-style stablecoin, but an intelligent asset system with M2 characteristics: earning interest like a money market when idle, yet capable of leverage, hedging, or spot conversion when needed—packing the core capabilities of Interactive Brokers and money market funds into the stablecoin’s backend. This way, it’s friendly to both humans and agents, naturally becoming foundational assets for the next-gen financial system.
The Rare Window Period for AIUSD to Reach Consumers
TechFlow: AIUSD has three core keywords: Crypto, AI, and Stablecoin. As insiders in Wall Street, how do traditional institutions view the potential of these three areas?
Yao:
To be honest, Wall Street’s attitude toward these three areas is highly stratified.
Crypto’s underlying infrastructure—custody, settlement, stablecoin clearing networks—is no longer seen as “rebellious tech” but as next-generation financial plumbing. More and more funds are exploring “how to make their capital efficiency approach on-chain speeds.”
AI is another story. Quant and risk management on Wall Street are already highly AI-driven, especially in strategy generation, data cleaning, and sentiment analysis. But the industry realizes AI can analyze—but cannot directly execute capital moves. This is a huge gap.
Stablecoins bridge the two. They’re the fuel for AI to execute trades and the settlement unit in the crypto world. Over the past five years, USDT and USDC have proven the value of “on-chain dollars.” In the next five, the market needs “smarter money”—capable of understanding strategy, yield, and risk, and programmable.
From a Wall Street OG perspective, these three aren’t isolated—they form an evolution chain: AI is the demand side, crypto the infrastructure, stablecoins the middleware. AIUSD aims to merge all three into one.
Bill Sun:
My experience at WorldQuant, Citadel, Point72 Cubist, and Millennium gave me deep familiarity with traditional buyside culture—where compliance comes first. Over the past two years, the approval of Bitcoin and Ethereum ETFs, along with stablecoin legislation (including the so-called Genius Act), sends a clear signal: traditional institutions can now enter.
Once institutions can fully participate, they bring geometrically increasing new liquidity—a structural shift in crypto markets that’s often overlooked but profoundly impactful.
In this context, stablecoins and compliant custody systems (like Coinbase Custody; offshore, CEFFU) are turning “on-chain dollars” into a trusted, regulated settlement layer. I anticipate a wave of institutional-grade stablecoins emerging—possibly from big tech firms, payment groups, or even offshore exchanges. The stablecoin landscape will evolve from today’s oligopoly into a multi-party competitive-cooperative phase. This transformation will make on-chain dollars resemble an open settlement network rather than just a few companies’ products.
AI’s application in trading remains at the stage of “assisting human traders and researchers” on Wall Street. Before founding my startup in 2023, I ran a small experiment: alone with an AI analysis system focused solely on AI, fintech, and crypto stocks, I achieved ~880% returns over a year. Of course, it was a high-volatility strategy, but it proved one thing: AI doing Deep Research combined with human PM judgment is far more efficient than I ever imagined.
This experiment led me to seriously consider: can this capability become a product, enabling retail investors to access institutional-grade research and decision tools?
Regulatory changes are also pushing forward Tokenized Stocks. U.S. equity trading is moving from “extended hours” toward true 7×24 operation, and non-U.S. investors’ stock trading will gradually shift from broker systems to on-chain. Tokenized stocks will act as compliant connectors, bridging traditional finance and the on-chain world.
TechFlow: Your team has deep expertise in LLM Trading Agents and spent years developing AIUSD. Is now the ideal time to launch AIUSD? What preparations has the team made to better bring AIUSD to market?
Yao:
I believe now is an excellent window period, and we’ve prepared for this step for many years.
The strategy engine, trading system, and risk management framework behind AIUSD stem from our prior work in high-frequency trading and arbitrage. Over the past two years, we’ve run AIUSD’s prototype internally—without public exposure—achieving over $1 trillion in annualized trading volume. We’ve also deeply partnered with CEFFU in the Binance ecosystem, integrating institutional custody with MirrorX yield channels. Additionally, we’ve built comprehensive risk, compliance, and control mechanisms—from fund flow monitoring to multisig and limit management—all meeting institutional standards.
This is why we’re confident. Now is the moment for AIUSD to truly reach consumers and theagenticeconomy.
Bill Sun:
Yes, this is indeed a rare window period.
On one hand, AI is transitioning from “chatting” to “executing”: more and more agents need to autonomously manage accounts, place orders, pay, and settle, but the industry lacks a unified, scalable, and secure account and settlement infrastructure. AIUSD fills this gap perfectly.
On the other hand, crypto infrastructure has genuinely matured: stablecoin clearing is becoming standardized, cross-chain routing increasingly reliable, institutional custody robust, and liquidity markets deeper. In this environment, we can achieve security, efficiency, and compliance simultaneously—not forced to choose one over others.
We’ve reinterpreted our decade-long stack of quant trading and risk management technologies into a zero-barrier account layer. As Yao mentioned, our core engines—funding rate, microstructure execution, cross-scenario routing—have run live internally for over two years, achieving $1 trillion in annualized volume. We’ve also built risk management, compliance frameworks, and automated controls to ensure this is a capital-ready system, not just a demo.
In one sentence: we’re not simply “plugging a model into an exchange,” but packaging “exchange-level execution and clearing capabilities” into an account layer usable by both humans and AI.
Built for Live Trading: An AI-Driven Closed Loop from Intent → Strategy → Execution
TechFlow: As a flagship product targeting end-users, what differentiates AIUSD from other AI trading platforms? What feature do you think is most appealing to C-end users?
Yao:
The biggest difference is: AIUSD isn’t a strategy platform—it’s an account layer.
Many so-called AI trading platforms merely let users pick models or strategies and place orders, but rely on centralized matching, isolate funds, and treat AI only as an advisor.
AIUSD flips this: we build an “account abstraction + intelligent settlement” layer. Users holding AIUSD automatically benefit from unified liquidity and yield routing across spot, derivatives, staking, and cross-chain payments.
AIUSD isn’t trying to build a “smarter trading platform,” but to enable every platform to settle and flow more intelligently through it.
Bill Sun:
Yes, let me illustrate with specific examples.
First, at AIUSD, natural language iscommand: users don’t need to click buttons, switch chains, or calculate gas. For instance, saying “Buy some ETH with 1000 AIUSD, put the rest in stable yield” allows the system to automatically decompose this into a series of trades, routing, and settlements executed in the background. To the user, it’s just a sentence; to the system, it’s a full microstructure-level task.
Second, at AIUSD, yield and safety are guaranteed together: AIUSD itself is 1:1 pegged to USDT, while yield is carried by sAIUSD through delta-neutral funding rate strategies. This strategy has run live internally for two and a half years with no monthly drawdowns. This structure makes stability and yield non-mutually exclusive.
Another crucial aspect is AIUSD’s truly unified wallet experience: users don’t need to care about which chain they’re on, switch wallets, interact with bridges, or learn cross-chain syntax. Cross-chain, spot, perpetual, and payments all happen within a unified account semantic—this is our emphasis on Machine-Native architecture: money is unified, execution is unified, so the experience is naturally unified.
In this model, users aren’t passively “selecting strategies.” They own money that thinks, executes, and manages—backed by a personal AI finance team: an Analyst for deep research, an Execution Trader for order placement and microstructure handling, and a Wealth Manager for position management and allocation.
We call this Vibe Trader or Vibe Coding: users express intent through semantics and intuition, and AI completes the full loop from “intent → strategy → execution,” turning the entire process into a seamless financial interaction.
TechFlow: “Stablecoins usable by AI” is a key concept for AIUSD. Why is enabling AI-driven stablecoin trading important? What innovations will this unlock for on-chain finance?
Yao:
Because the next generation of financial actors may not be humans.
AI agents will hold assets, trade, settle profits, and reinvest. But today’s stablecoin systems are designed for humans, not agents. AIUSD aims to upgrade stablecoins into “AI-native Money.”
Bill Sun:
We believe this unlocks several new possibilities:
First, AI can actively manage cash flows—e.g., automatically switching between AIUSD and sAIUSD based on market volatility, then back to liquidity accounts.
Second, autonomous financial networks can form between agents—e.g., one AI hiring another for data analysis and paying compensation, all settled on-chain.
Third, finance shifts from “human placing orders” to “intent execution.” Users state goals; AI uses AIUSD to search paths, execute, and settle.
Long-term, AIUSD won’t just be a stablecoin—it’ll be the “settlement hub” for the entire AI financial ecosystem.
TechFlow: Recently, the trading competition among the world’s six top AI models sparked widespread discussion. What’s your take on this live test? If AIUSD participated, would it perform better?
Yao:
My focus on this match differs. How well models perform is secondary. The key issue is that AI’s capabilities are limited to “generating judgments” without a complete execution environment. AIUSD’s system is the opposite—designed for live trading from day one.
AIUSD’s Agent doesn’t just give opinions—it’s directly linked to real liquidity, funding rates, cross-venue execution paths, and clearing systems. Our internal execution and routing engine has run live for over two years, achieving trillion-dollar annualized volume, with profits derived from structural capital efficiency gaps—not directional bets.
So if future competitions allow real execution, I’d gladly participate—because that’s where “AI financial systems,” not just “AI models,” can truly shine.
Bill Sun:
I believe this match does reveal one truth: models can now interpret market signals and form trading logic—this is AI’s most visible progress in finance over recent years.
But my immediate reaction was that such competitions still face a fundamental gap from real trading. These AIs are “can think but cannot act”: they can’t move funds cross-chain, place real orders, or connect to clearing systems—so they remain stuck at paper trading.
In real markets, the challenge has never been predicting prices, but execution, risk control, capital routing, and clearing—the transaction infrastructure. If future matches evolve into “AI + live execution” contests, the significance would be entirely different. That would test whether a system can survive in real markets—not just whose model outputs prettier probabilities.
TechFlow: To date, what key metrics has AIUSD achieved?
Yao:
We’ve run the execution and settlement pipeline live for over two years: the core strategy stack has sustained an annualized trading volume in the trillions of dollars—not a polished cumulative figure, but a repeatable, backtestable, auditable live execution scale. The core combination of funding rate and microstructure strategies has delivered a historical Sharpe ratio around 22, with no monthly drawdowns over the past 2.5 years. On stability, AIUSD maintains 1:1 USDT backing, with sAIUSD yields generated at the staking layer—two separate ledgers ensuring redemption and interest calculation don’t interfere. We’ve also deeply integrated with CEFFU on custody.
User experience is also improving: natural language intent parsing now covers cross-chain transfers, micropayments, and composite spot/perpetual instructions. Users manage collateral within one account, eliminating the need to shuttle assets across venues. To me, the real achievement isn’t a flashy number, but that this system hasn’t failed—even during the most volatile market conditions.
Bill Sun:
Our current system evolved from Alpha.dev—an AI platform for crypto news, sentiment, and trading signals, now serving around 3.5 million users with over 60 million total interactions. This data reflects a simple but real demand: people genuinely want AI to help read information streams, filter noise, generate insights, and even find trading opportunities.
Our quant engine is a separate track—the underlying execution system, achieving an annualized trading volume equivalent to one trillion RMB. Our core strategy is fully delta-neutral, fully hedged funding rate arbitrage—profiting purely from structural capital efficiency differences, not directional bets. Over the past three years, this strategy has delivered average annual returns above 20%, a Sharpe ratio around 22, and no monthly drawdowns. No matter how volatile the market, our return curve remains smooth.
Unlike many “rebalancing” approaches, we do dynamic funding rate optimization—tracking, switching, and reallocating capital among the top 20 market-cap coins. This isn’t a superficially stable strategy, but one rigorously tested under high volatility.
Another noteworthy aspect is our AI analysis and execution framework, which can close the loop from “proposing research hypotheses” to “automatically searching data for validation,” then “generating and executing strategies,” and even “automated backtesting.” As mentioned earlier, I applied this architecture to the stock market using an “AI does deep analysis, I make final PM decisions” approach—my personal account returned ~880% over the past year, validating this method across asset classes.
The Inflection from 1 to 10: Becoming the Universal Settlement Layer for the AI Economy
TechFlow: With “AI-driven financial intelligence” as the core vision, where do you see AIUSD in its journey from 0 to 1 to 100? What challenges lie ahead?
Yao:
I believe we’ve successfully completed the 0-to-1 phase: 0 was proving technical feasibility, 1 was creating an account layer anyone can use—speak, and the system routes and settles.
Now, we’re at the 1-to-10 inflection point, aiming to reliably deliver this same infrastructure to a broader consumer and developer ecosystem.
Bill Sun:
I categorize future challenges into three:
First, scalability uncertainties—including funding rate congestion and capacity limits. The larger we grow, the more restraint we need—we’d rather sacrifice some yield than allow drawdowns.
Second, regulatory puzzles across jurisdictions. Boundaries for stablecoins, payments, brokerage, and yield must be clearly defined—requiring time and patience.
Third, agent ecosystem usability—intent expression, permission granularity, rollback capability, audit logs—must be plug-and-play for developers. I prioritize refining these invisible technical details until they feel seamless.
For example, we liken our work to Stripe: building bridges and roads for digital finance, consolidating fragmented rails into elegant APIs. AIUSD aims to become the universal settlement layer for the AI economy.
TechFlow: Stablecoins and RWA are central themes of this cycle. Does AIUSD have further expansion plans around stablecoins and RWA? How might their continued development transform global finance?
Yao:
Let me outline our strategic direction.
AIUSD’s positioning remains “universal stablecoin infrastructure,” not “passive mapping of a basket of assets.” Thus, on the stablecoin side, we maintain a two-layer structure: AIUSD itself stays 1:1 backed by USDT, with a simple, transparent redemption path; yield resides in optional sAIUSD, powered by delta-neutral funding rate strategies to deliver “alpha,” without exposing principal to external credit or duration mismatch risks. We’ll deepen this line further, including finer-grained risk limits and dynamic fallback mechanisms.
Regarding RWA, we’ll proceed cautiously. I don’t oppose bringing real-world cash flows on-chain, but we must honestly address three issues: liquidity stratification, verifiability of valuation/pricing, and legal clarity on “beneficial ownership vs. legal ownership.” Where these can be resolved—for example, shortest-duration receivables or highly liquid treasury opportunities—we’ll pilot small-scale, retractable, low-correlation experiments—never compromising AIUSD’s core redemption certainty.
Macroscopically, the ongoing development of stablecoins and RWA will elevate the baseline for “clearing speed” and “ownership proof” in financial markets. As global capital becomes more programmable, pricing will accelerate, mispricing will be costlier, good risks will be priced more accurately, and bad risks harder to hide off-balance-sheet. I believe this will force finance to return to fundamentals: quality assets and compliant capabilities.
Bill Sun:
Stablecoins correspond to the M0 layer—the global cash layer—solving the most basic question: can dollars truly become globally accessible and instantly settle on-chain, independent of local banking systems?
RWA builds atop this—it’s the asset layer—enabling high-quality dollar assets, treasuries, bills, even private credit to become 7×24, sliceable, globally tradable assets.
Without RWA, stablecoin adoption will likely remain confined to high-inflation, relatively open-capital economies. But once RWA opens, institutional-grade assets can truly reach global retail—transforming the significance entirely.
AIUSD’s role isn’t to create another on-chain stablecoin, but to turn interactions between different stablecoins into a platform layer—letting users buy any RWA on any chain—whether Ethereum, Solana, Base, Sui, or Tron. Underlying VM and gas differences are automatically abstracted by the system.
Macroscopically, this redefines the financial system’s baseline: clearing speeds reach real-time, ownership proof becomes default; good risks are precisely priced, bad risks harder to conceal; the entire system converges toward “high-quality assets + high transparency + high liquidity.”
If I summarize this in one sentence: stablecoins are the global projection of the dollar; RWA is the global projection of high-quality non-dollar assets. AIUSD aims to be the intelligent connector between them—making the flow of money and assets simple, smart, and automated.
Our ultimate vision is straightforward: a stablecoin-based money system + a global exchange based on RWA. Just talk to money, and AIUSD helps you buy any tokenized asset worldwide. These assets themselves are programmable—interactable via semantics for humans, autonomously callable, rebalanced, and settled by AI.
TechFlow: Beyond AIUSD, do you have plans for a broader product matrix to build an AI-driven financial intelligence future?
Yao:
Yes, but we won’t expand by stacking product lines. We’re more like building a stablecoin highway and gradually adding essential service stations.
Short-term, we’ll enhance “account abstraction + collateral credit” capabilities—allowing the same collateral to efficiently serve multiple venues while hardcoding liquidation paths and extreme scenarios into contracts.
Mid-term, we’ll consider productizing yield curves—offering different duration and volatility tolerance tiers atop sAIUSD, letting users and AI agents select levels by “target drawdown/target volatility” instead of accepting a single averaged yield line.
Long-term, I don’t envision building N apps, but a stable, transparent, accountable financial operating system.
Bill Sun:
Yes, we won’t horizontally stack more product lines. All capabilities will converge into AIUSD as the core hub. It starts as a stablecoin hub, then evolves into an intelligent brokerage: regardless of which stablecoin users hold, they can use, trade, and allocate them across assets within a unified experience.
We’ll integrate Alpha.dev’s AI news, event detection, and research capabilities—so the system doesn’t just read information, but understands your preferences and risk habits, proactively recommending opportunities you might like and can directly trade.
Another track connects TradFi and CEX. Many people’s funds remain in Interactive Brokers, Robinhood, or centralized exchanges. For real trading, you can’t bypass these. So we’ll directly integrate their APIs into AIUSD, unifying identity, keys, permissions, and order placement into one account layer—eliminating asset fragmentation and system switching.
For developers, we’ll standardize the “intent layer + permissions/quotas/whitelists/rollback/audit” into SDKs, and offer B2B white-label routing and settlement. In short, we aim to transform financial interaction—from multiple entry points, accounts, and logics—into one unified operating system.
Post-AGI Era: Humans as Narrative Producers, AI as Execution Machines
TechFlow: As AIUSD progresses toward its vision of an “AI-driven financial intelligence future,” could you describe what trading looks like in the Post-AGI era?
Yao:
In my view, Post-AGI trading won’t feature flashier interfaces, but quieter infrastructure. Price discovery will accelerate, and prediction itself will no longer be the main profit source. True alpha will come from orchestrating capital, compute, data, and settlement paths with minimal friction. Trading shifts from “humans watching screens” to “intent governance”: you declare goals, constraints, and trust boundaries; the system executes, reviews, and corrects within verifiable rails. Human roles will resemble governors and narrative producers, setting rules and boundaries, defining acceptable risk and desirable returns, then delegating most execution to machines. Capital will move faster, yet more systematically—every movement carrying auditable rationale and attributable signatures. This is my ideal trading world: efficiency maximized, but discipline and explainability embedded in the protocol.
Bill Sun:
My definition of AGI is pragmatic: essentially, it systematically replaces white-collar jobs, merging professional skills within a single model—not a sci-fi “singularity” moment. Given current progress, I believe the Post-AGI phase may arrive within three to four years.
At that stage, a critical need emerges: AI requires a form of “money” it can truly understand and directly use. That’s the direction we’ve set for AIUSD—to become M2M, Machine-to-Machine payment’s foundational currency. This allows AI to incorporate “money” into its world model: paying for compute, data, services, investing, reinvesting—treating finance as action.
Simultaneously, I foresee two key agent forms: one, the AI Scientist, using AIUSD to buy compute, data, RL environments, energy—forming a self-improving loop capable of self-iteration; and two, a free market among multiple agents. We won’t enter a “single supermodel planned economy,” but a world where countless independent AGIs trade, compete, collaborate via AIUSD, allocating resources through price discovery.
In such a world, AIUSD holds different yet complementary meanings for humans and AI. For humans, it’s a Financial OS—managing wealth through intent governance, needing no technical knowledge, just stating goals, while the system handles strategy, execution, and settlement. For AI, it’s Agentic Money—granting true financial agency, enabling agents to operate as active participants, not passive observers, within the economic system.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
深潮TechFlow
