Consider investing in core crypto infrastructure like Ethereum (ETH) and Uniswap (UNI), which are positioned to serve as the financial rails for future AI agent transactions. This strategy is a direct play on the explosive growth of stablecoins like USDC, which are expected to be the primary payment method for this new agentic economy. The perceived threat to payment processors like Visa (V) and Mastercard (MA) is likely overstated, as their models appear secure for the next 3-5 years before autonomous AI payments become mainstream. Conversely, be cautious with legacy tech stocks like IBM, whose service-based business models are at high risk of disruption from specialized AI. The most actionable theme is to focus on these established "picks and shovels" of crypto finance, not premature bets on decentralized AI compute.

Stablecoins (USDC, USDT)

• The guest, Haseeb Qureshi, noted that the stablecoin market has been growing at 50-60% every year.
• He referenced a projection that the market could reach $2.7 trillion by the end of the decade, which would represent roughly 70% year-over-year growth. Even at the current 56% growth rate, the market is on a path to a multi-trillion dollar valuation.
• Current demand is driven by humans, not AI. Key drivers include:
  • B2B payments
  • Remittances
  • "Backdoor dollarization" in countries with geopolitical instability, where citizens seek the stability of the US dollar.
• While AI agents will eventually use stablecoins, their demand will be a result of the underlying human demand for these assets.

Takeaways

• The growth of stablecoins is a powerful, long-term secular trend.
• Investing in the infrastructure that supports stablecoin usage (like the blockchains they run on and the applications that use them) is a way to gain exposure to this growth.
• The primary use cases driving adoption today are international and business-focused, not just crypto trading.

Payment Processors (Mastercard (MA), Visa (V))

• The podcast discussed a viral article predicting a "consumer credit crisis" where AI agents would bypass traditional payment rails to avoid the 2.5-3% credit card processing fees, opting for near-instant stablecoin payments instead.
• The article hypothetically suggested this could cause a stock like Mastercard to drop significantly.
• Haseeb believes this scenario is "AI doomer porn for finance people" and that the timeline is much longer than many think. He predicts a "two-track world":
  • Track 1: Human-Approved (Short-to-Medium Term): The vast majority of commerce will still involve a human approving a transaction suggested by an AI. Visa and Mastercard are very comfortable with this model and it will remain dominant for the next several years.
  • Track 2: Autonomous Agents (Long-Term): AI agents making purchases autonomously (e.g., while the user is asleep) breaks the current risk models of credit card companies. This is the true disruptive threat, but it's currently confined to hobbyists and startups and is considered "way too risky" for mainstream adoption.

Takeaways

• The thesis that traditional payment processors are "cooked" is likely premature. The disruption from AI-driven payments is a 5-to-10-year story, not an immediate threat.
• For the foreseeable future (3-5 years), these companies' business models appear safe as most AI-facilitated commerce will still require human approval, fitting within existing payment systems.
• Investors should monitor the development of autonomous AI agent capabilities as a long-term risk factor for this sector.

Crypto Infrastructure (Ethereum (ETH), Uniswap (UNI))

• The guest argues that the coming wave of on-chain activity from AI agents will be a massive source of new demand for crypto.
• When asked what to buy if you believe AI agents are coming on-chain, the answer was "literally buy anything" that serves as underlying infrastructure.
• He specifically mentioned that Ethereum (ETH), Uniswap (UNI), and Circle (USDC) will be major beneficiaries because they are the foundational rails and applications that AI agents will use to conduct transactions.
• The core idea is to invest in the "picks and shovels" of the AI-on-chain economy.

Takeaways

• Bullish Sentiment: The growth of AI is presented as a major long-term catalyst for core crypto infrastructure.
• Investing in established, high-utility protocols like Ethereum and Uniswap is a direct way to bet on this theme without needing to pick the winning AI-specific applications, which are still in their early, experimental stages.

Legacy Tech (IBM)

• A specific anecdote was shared about IBM's stock value taking a "gap down" after the AI lab Anthropic published a single blog post about modernizing COBOL (an old programming language common in mainframe systems where IBM is a leader).
• This illustrates how quickly AI can threaten the business models of established, legacy technology companies.
• The guest suggests that in the future, a specialized COBOL-refactoring AI agent could be available on a pay-as-you-go basis, completely disrupting the enterprise sales model that companies like IBM rely on.

Takeaways

• Bearish Sentiment: Legacy tech companies whose business models are built around "human friction" (e.g., long enterprise sales cycles, consulting services) are at high risk of disruption from specialized AI agents.
• Investors should be cautious about companies with significant revenue from services that could be automated and offered on-demand by an AI.

Investment Theme: The Convergence of AI & Crypto

• The discussion highlighted a major shift in how investors are thinking about the intersection of AI and crypto.
  • The Old Thesis (Decentralized Infrastructure): Two years ago, the focus was on projects trying to decentralize GPU marketplaces, AI model training, and inference. The guest stated that these have seen "very low demand" and have not been successful investments so far, as AI developers prioritize performance over ideology.
  • The New Thesis (Agentic Payments): The current, more promising theme is that AI agents will become major users of crypto's financial infrastructure, primarily stablecoins on public blockchains, for payments and commerce.
  • The Future Thesis (Censorship Resistance): The "decentralized AI" thesis is not dead, but dormant. The guest predicts it will become highly relevant when centralized AI models become heavily censored and controlled by corporations and governments. As this happens, demand will surge for decentralized, censorship-resistant AI models, similar to how Bitcoin emerged as an alternative to state-controlled money.

Takeaways

• The most actionable investment thesis today is in the "picks and shovels" that will support AI agent commerce (e.g., stablecoins, L1/L2 blockchains, DeFi protocols).
• Investments in decentralized AI compute and infrastructure are likely premature. The catalyst for their adoption—widespread censorship of centralized AI—has not happened yet but is a key trend to watch.
• Long-term investors could monitor the global regulatory and corporate landscape around AI. Increasing restrictions and "locking down" of models could be the signal that the "decentralized AI" thesis is about to have its moment.