Louis Gave

There's Going To Be One Hell Of A Hangover When The Market Party Ends | Louis Gave

• Top takeaway He argues the AI buildout is financially unjustifiable on current assumptions: the projected $6.7 trillion of AI capex between 2026 and 2030 would require roughly $2 trillion a year of AI revenue — about twice the entire global advertising industry — and he calls the numbers very stretched.
• He lays out a four-scenario framework (deflationary and inflationary boom and bust) and judges the deflationary bust least likely given US deficits around 7% of GDP and bank shares not flashing red, leaving an inflationary boom as his base case — favoring value stocks, metals, and financials.
• He reiterates that the 60/40 portfolio is dead and recommends roughly 60% equities, 20% energy and 20% metals instead, arguing bonds stay dead because regulation and financial repression trap savings — pointing to the Fed missing its 2% inflation target for 64 straight months as evidence it is deliberate.
• He flags Japanese repatriation of its roughly $3.5 trillion of US assets as a live risk for the next 12 months — citing Korea's tax-free repatriation program as precedent and BOJ tightening — and recommends cheap out-of-the-money yen calls as a hedge that also pays off if an AI bust sinks the dollar.
• He admits he has been wrong to stay wary of semiconductors, losing the internal debate to Asian Opportunities Fund manager Alfred Ho, but sees better value in uncrowded areas — Latin America after its rightward election wave, global financials, and China's newly world-class exporters — while warning that DeepSeek-style efficient models threaten the US capital-intensive 'Ferrari' approach.

Gave argues AI capex of $6.7 trillion needs about $2 trillion in yearly revenue to pencil out, and current valuations ignore that math. A four-scenario framework (deflationary or inflationary, boom or bust) points to an inflationary boom favoring value stocks, metals, energy, and financials. He declares the 60/40 portfolio dead, replacing bonds with 20% energy and 20% metals, because regulation and fiscal dominance keep bond money trapped. Japanese repatriation of roughly $3.5 trillion held abroad is a top risk for the next 12 months; cheap yen calls are a dual-purpose hedge. China's state-backed industrial push is producing world-class companies, and DeepSeek-style efficiency threatens the revenue assumptions behind US AI spending.