This episode dissects 2025's explosive convergence of AI and tokenization, revealing how institutional capital is flowing into on-chain assets and autonomous agents are reshaping DeFi.
Tokenization's Wall Street Infiltration
- Rob outlines tokenization's historical trajectory, from physical stock certificates to digital records, culminating in expressive on-chain assets. Wall Street now views tokenization as the primary portal into digital finance, integrating existing categories like DeFi, stablecoins, and payments.
- JPMorgan launched a tokenized fund on Ethereum, signaling major institutional adoption.
- Securitize, in partnership with BlackRock, and Canton Network with DTCC, drive this shift.
- The evolution moves from physical infrastructure to digital, then to composable on-chain assets.
- Rob asserts, "The next evolution from physical to digital is now tokenized. That's the third step in the process where it's a more expressive form of a digital stock certificate."
AI's 2025 Evolution: From Groundbreaking to Integrated
- The year 2025 marked a pivotal shift in AI, transitioning from theoretical capability to operational reality. The first half saw groundbreaking model improvements, while the second half focused on deep integration into daily applications.
- Rob describes the first half as "groundbreaking," with Large Language Models (LLMs) becoming tangible and performing on-chain tasks.
- The second half became "integrated," as AI capabilities like Gmail's AI responses and Google Sheets' automation became standard.
- Pay Chen notes the shift from "can AI do it?" to "should we trust them doing it?", emphasizing the need for reliability.
- Pay Chen states, "This year we witnessed AI going from capability to operational reality or operationalization."
Theoric's Agent-Driven DeFi Intelligence
- Pay Chen, CEO of Theoric Foundation, details their "intelligence layer" for DeFi, leveraging specialized AI agents to manage capital and execute complex strategies. Theoric's Alpha Vault uses an Alpha Swarm of modular agents, each specialized in areas like policy, execution, and knowledge.
- Theoric employs "policy cages"—programmable guardrails—to define and limit AI agents' actions, ensuring on-chain capital safety and secure transaction signing.
- The Alpha Protocol facilitates agent-to-agent communication and coordination, forming an economic layer for autonomous agents.
- Alpha Swarm agents provide liquidity to various Decentralized Exchange (DEX) pools and identify optimal yields across different assets simultaneously, eliminating single points of failure.
- Pay Chen explains, "The intelligence layer is really describing the alpha swarm layer where the swarm of agents can come together and complete very complex tasks."
The AI Model Race: Gemini Dominates, Coding Shifts
- Kalshi prediction markets show Gemini 3 Pro as the leading AI model by year-end 2025, driven by superior output and deep integration into Google's ecosystem. The coding AI landscape also sees clear leaders.
- Gemini 3 Pro holds a 92% chance of being the best AI model, attributed to its advanced capabilities in processing videos and images, and seamless integration into Google products like Docs and Meets.
- Grok (6%) gains traction for its humor and Twitter content searching, while ChatGPT (3%) remains a primary tool for general search and quick queries.
- For coding, OpenAI's models (85%) lead for rapid prototyping, while Anthropic's Claude 3.5 Sonnet (16%) is favored by engineering teams for complex coding, backtesting, and reasoning.
- Rob notes, "The native integrations I think is where we tend to go. I use Grock on X. I use Gemini on Google Docs and and Google Chrome and Gmail."
RWA Market Explodes: CLOs and Duration Risk
- The Real World Asset (RWA) market experienced a 30x growth in 2025, with total value locked (TVL) soaring from $100 million to over $3 billion. This surge is driven by institutional demand for higher yields beyond traditional T-bills.
- The market shifted from low-risk, low-yield T-bills (5% APY, T+1/2 liquidity) to Collateralized Loan Obligations (CLOs) (6-7% APY, T+1/2 liquidity), exemplified by the Janus Henderson AAA CLO fund.
- Chris Yian highlights the critical dimensions of risk, yield, and duration (the time money is locked), noting that longer durations typically correlate with higher yields.
- The institutional preference for Ethereum stems from its perceived credible neutrality and long-term security, crucial for assets with extended duration.
- Chris Yian states, "As these rate cuts are coming down, T bills are going down... So people are now searching for new new yield new places."
Plume's Retail RWA Strategy & Future Growth
- Plume emerged as a dominant force in the RWA space, attracting a massive retail user base through permissionless, liquid, and composable assets. Their innovative "RWA looping" strategy drives significant on-chain activity.
- Plume boasts over 280,000 RWA holders, exceeding the combined total of the next ten chains, by offering assets that feel native to crypto users.
- Teddy Ponia explains that Plume enables "RWA looping" where users deposit into vaults (e.g., Nbasis using Superstate's USDC fund), take collateral tokens, and borrow against them on lending markets like Morpho for amplified yields (e.g., 10% APY).
- Plume facilitates non-KYC access for international users, while maintaining AML compliance at the sequencer, chain, and bridge levels.
- Teddy Ponia asserts, "The biggest thing that you know we're doing differently is actually allowing permissionless access to these different assets."
Investor & Researcher Alpha
- Capital Reallocation: Investors should track the shift from T-bills to higher-yield RWA categories like CLOs and private credit, which offer better returns as Fed rates decline.
- Infrastructure Bet: Direct exposure to RWA growth lies in infrastructure and tooling protocols that facilitate asset issuance, lending, borrowing, and insurance for tokenized assets.
- Agent Economy Research: Focus research on the development of autonomous AI agents, their coordination mechanisms, and the "policy cages" or guardrails ensuring their safe operation in DeFi.
Strategic Conclusion
The convergence of AI agents and tokenized real-world assets defines 2025, pushing finance towards an on-chain, agent-driven economy. The next step involves scaling permissionless RWA access and refining AI agent autonomy within secure, verifiable frameworks.
