PodBrowser
Forward Guidance

How To Trade The AI Productivity Boom | Weekly Roundup

Friday, 29 May 2026 · 4 min read · Listen to the episode ↗

This week's discussion centers on how to position around the AI productivity boom and a fragile equity market sitting on historically cheap hedges. The hosts argue that AI infrastructure spending is real but the gains are accruing to a narrow group, with capital being pulled out of crypto and into data centers in a dynamic reminiscent of how Silicon Valley drained gold miners around 2012.

The Fed is described as running wartime monetary policy driven by the need to service roughly one trillion dollars in annual interest expense rather than by inflation-fighting orthodoxy. Taylor rule models suggest the funds rate should be higher given that inflation has exceeded the 2 percent target for more than 60 months. The long-term policy direction is framed as engineering permanently negative real rates to fiscally grow out of the debt problem, with balance sheet operations including RRP management characterized as more stimulative in many cases than rate cuts.

The VIX dropped from around 40 to sub-16, and the administration is described as holding extreme information asymmetry that allows trade headlines, real or fabricated, to reliably compress volatility and push equities higher. Semiconductor call positioning is in the 98th percentile with no downside protection being purchased, and S&P one-month 25-delta put skew relative to calls is in the fourth percentile, meaning hedges are historically cheap. The dispersion and low-VIX dynamic is flagged as capable of unwinding sharply, as seen during the prior year's carry trade unwind. A massive sectoral rotation is anticipated given how compressed credit spreads currently are, with a baton passing away from semiconductors expected over the next few weeks without necessarily implying a broader market downturn.

Personal disposable incomes have turned negative for the first time since 2022, forcing consumers to draw down savings to absorb energy costs. The US strategic petroleum reserve is being drawn toward zero and oil inventories are below the five-year seasonal range, with Iran described as incentivized to drag out the current geopolitical situation given time pressure on US supply buffers. Energy supply shocks are described as a nightmare for central banks because they raise headline inflation while simultaneously crushing incomes. Large tax refunds from the Big Beautiful Bill temporarily supported consumer spending but are described as running out over the next couple of months, and without trade deal resolution the speaker predicts conditions will turn stagflationary within that same window.

The US is approximately four months from midterm elections, which constrains the administration's ability to re-escalate trade conflicts. One speaker predicts Trump will cut a trade deal that ends up worse than Obama's because he cannot sustain another prolonged conflict given current gas prices and consumer health. The lower half of the K-shaped economy is described as doing very poorly going into midterms with zero policy consideration, and the speaker expresses doubt that poll numbers can be meaningfully improved in five months while leaving lower and middle income classes behind. The prolonged concentration of wealth is predicted to eventually transmit volatility to the social contract and produce more leftist redistributive political outcomes, with AOC and Mamdani-style politics described as more likely than hands-off conservative government if inequality is not addressed. Peter Thiel relocating to Argentina is framed as a signal of the beginning of a secular shift rather than something to fade, citing his track record of being early to major trends.

The AI productivity boom is described as real and contributing to a managed policy environment spanning geopolitics, monetary policy, and treasury issuance, but the windfall is characterized as accruing to a small cadre of people. Capital is described as being sucked out of crypto into AI infrastructure in a dynamic analogous to how Silicon Valley tech growth drained capital from gold miners around 2012. The AI data center infrastructure trend is expected to remain significant for the next two years, with Nebius identified earlier in the year as a major opportunity in that space. Mike Green's argument is cited that the shift from defined benefit to defined contribution pension plans created multi-trillion dollar flows that no longer require duration hedging, making markets more synthetic and flow-driven, and that wealthier individuals compound capital exponentially better because they can tolerate drawdowns while lower-income individuals cannot.

Bitcoin is described as currently in a bear market with unfavorable macro conditions. Entities described as DATs bought approximately 15 billion dollars of Bitcoin and ETH this year while both assets are down significantly. David Hoffman from Bankless is cited as having recently sold his Ethereum holdings, stating he is bullish on the network but considers the ETH asset uninvestable, and the speaker notes that Ethereum network usage does not correlate to the price of the ETH token. Productive crypto applications such as stablecoin adoption and improved payment rails show real adoption but no value capture in associated tokens, and the speaker states 99 percent of the token landscape has nothing useful.

A thematic investing framework attributed to former boss Mark Hart uses access, awareness, patina, total addressable market, and use as collateral as key indicators, with concentric circles of adoption tracking when new pools of investors gain access to a theme through events such as ETF launches. A regulatory change that restricts access is treated as a signal to exit a thematic position. Critically attractive entry points are described as essential because they allow the investor to hold through volatility using house money, whereas chasing entries out of fear of missing out leads to poor fills on wicks that reverse and shake out the position.

This summary was generated from the episode transcript and can contain mistakes.