Visit ai.22vresearch.com and visser-labs.com
For more information contact Mark Whaling - mwhaling@22vresearch.com
In this week's video, I push back on the bond-market panic dominating X. The 30-year sits at 5.23% and could print its highest level since 2004, but a yield chart does not confirm a crisis: bond volatility is subdued, junk spreads are the tightest ever, inflation swaps from one to ten years are all pinned near 2.5% (right where core CPI sits), and Truflation is falling fast. This is a global, debt-driven repricing that has run since 2022, and governments from Washington to Tokyo have made clear they will fight it. The economic and equity facts, meanwhile, are hard to argue with: Nominal GDP is 6%. The S&P PEG ratio is at its lowest in decades, revisions are surging globally, profit margins are expanding, and Nvidia is approaching new all-time highs while semiconductor bears call it a bubble.
The bigger investment story is the merge of AI and crypto. The best model in the world can now do essentially every knowledge-worker job, agents are operating in teams of 1,200, and the mental model has shifted from AI assistant to AI organization. Interest rates are a rounding error on inference P&L: a 200 basis point rise in 10-year yields costs roughly one point of margin, while a 10% model price decline costs three. The three companies that matter are Anthropic, OpenAI, and Nvidia, and model price compression, not rates, is the risk to watch.
As Marc Andreessen argued, AI is crypto's killer app. Agents need native payments, stablecoins, and tokenization, and they, not humans, will become crypto's biggest users. Robinhood's tokenized stock volumes, an 18x expansion in real-world assets, and perpetual futures on private names show the roadmap accelerating now. With crypto sharply outperforming AI equities quarter to date and positioning still light, I lay out how I size the bet: if you believe there is a 5% chance agents drive this ecosystem, hold 5%.
Timestamps
00:00–03:26 Empowerment through AI — Agents are breaking the old school/job/save/retire model and democratizing learning, building, and investing.
03:26–07:42 Bond fears are noise — Yields are high, but bond vol is calm, junk spreads are tight, inflation expectations are anchored, and Treasuries are only modestly lower YTD.
07:42–11:07 The equity facts — Low PEGs, rising revisions, strong PMIs, expanding margins, and broad global gains do not look like recession.
11:07–15:33 Oil, consumers & the correction — Spending is strong and the pullback looks more like deleveraging than an earnings problem; Nvidia is nearing highs.
15:33–18:33 AI portfolio stuck in the mud — Volatility across the 100-name book is falling, suggesting forced selling is fading as crypto weights rise.
18:33–20:41 Rates vs. inference economics — Model-price compression is a far bigger margin risk for Anthropic, OpenAI, and Nvidia than higher Treasury yields.
20:41–26:29 Welcome to the AGI era — Models now cover most knowledge work, agents coordinate in large teams, and businesses must redesign workflows around AI organizations.
26:29–30:21 Financial agents arrive — Agent-driven commerce, crypto buying, lending, and automated investing point toward a machine-to-machine economy.
30:21–34:28 AI is crypto's killer app — Agents need stablecoins, tokenization, wallets, and native internet payments; machines may become crypto's biggest users.
34:28–40:41 Crypto & tokenization breakout — BTC, ETH, and SOL outperform as tokenized stocks, RWAs, and private-market perps push markets toward 24/7/365.
40:41–46:21 Takeaways & sizing — If you assign a 5% probability to agents driving crypto adoption, a 5% allocation is a rational starting point.