Reviews his previous call: S&P and Mag7 up ~20%, Europe up ~18%, but Russell 2000 up 38% — so rotation into value/small caps did occur. Emerging markets up ~50% but driven by 3 Taiwanese/Korean semiconductor names, not broad EM. Onshore Chinese equities up ~30%.

He marks his own homework: 'could do better, but altogether not terrible.' The narrative of American exceptionalism persists due to AI leadership, but the actual market performance shows a rotation into cyclical/value.

Hesitates to call a top but is skeptical. Bulls argue the market has derated this year due to strong earnings revisions — unlike other bubbles, it's not expensive if earnings come through. However, those earnings are predicated on compute scarcity. If compute is not scarce (due to cheaper models, small language models, Chinese alternatives, or CPU-based supercomputers), the earnings bubble could burst.

He explains the circularity of vendor financing in the AI ecosystem: customers invest in chip companies to place orders, creating a multiplier effect on earnings revisions across the stack (model builders, hyperscalers, memory, logic). This makes the earnings growth fragile.

Explains: GPUs are advanced, powerful, energy-intensive logic semiconductors for thinking; CPUs are more basic, less powerful, less energy-intensive. A Chinese supercomputer powered entirely by CPUs just won the world ranking, showing it's not all about throwing the most leading-edge chip at problems.

This leads to two possibilities: 1) This is an earnings bubble more than a valuation bubble; 2) The huge CapEx on data centers might be too much, too soon.

Agrees. Adds that most bubbles are productive — they suck capital into something transformative (railways, dot-com). The exception is pure speculation (GFC, tulips). He thinks AI is transformative but valuations may be wrong.

He notes the tulip bubble left a legacy: the Netherlands is still a top flower exporter. So even bubbles can have lasting positive effects.

The world is changing: globalization reversing, more friction, national interest prioritized, stickier inflation. The 60/40 portfolio relied on negative stock-bond correlation. In 1974 (inflationary, positive correlation), a 60/40 lost ~30%; in 2008 (negative correlation), losses were much smaller. Sticky inflation reduces central banks' ability to cut rates in a crisis, undermining the 'central bank put'.

He cites UK CPI at ~3%, services CPI at 3.8% with 1% GDP growth — not demand-driven. US core PCE at 3.3%, with 60% from acyclical components accelerating. This suggests inflation is sticky and hard to get back to 2%.

Recommends shorter-duration cash flows — real fixed capital base (hard assets, low obsolescence) rather than intangible assets of the last cycle. This is the 'halo trade'. Finds value in smaller caps outside the US. Expects a weaker dollar, which benefits non-US equities. Bonds still have a role (e.g., UK low-coupon bonds with tax efficiency), but may not be reliable diversifiers. Sees a golden age for liquid alternatives (hedge funds, equity long/short) which benefit from higher inflation and dispersion.

He explains that higher interest rates force companies to do more remedial work, creating dispersion and opportunity for stock pickers. Returns to equity long/short improve as inflation rises.

The endowment model (1980s) offered an illiquidity premium for long-term investors. Today, the universe of private assets has grown enormously, exits are challenged, returns are diminishing, and the risk profile looks similar to public equities. In a downturn, private assets will have the same downward vulnerability, making them less diversifying.

He suggests that lack of liquidity may come with a discount rather than a premium in the future.

Still likes gold personally. Expects a weaker dollar (based on his read of Warsh's guidance as less hawkish, more flexible), which supports gold. Central bank demand for gold as a reserve asset remains strong due to de-dollarization. However, gold can become a risk-on asset at times (as seen in March), so it's not a reliable diversifier in all environments.

Amazed at the policy support crypto has received, yet it has performed poorly. Questions its utility. Still prefers gold over crypto.

He notes that maybe the poor performance makes it a good time to buy (contrarian), but he's not convinced.

Recommends 'The Wide Wide Sea' by Hampton Sides — an account of Captain Cook's last voyage. It's a reminder that modernity is only 250 years old and progress is not guaranteed.