I disagree with the forecast too, but your critique is off-base. The assumption that exponential cash is required assumes that subexponential capex can't chug a

I disagree with the forecast too, but your critique is off-base. The assumption that exponential cash is required assumes that subexponential capex can't chug along gradually without the industry collapsing into mass bankruptcy. Additionally, the investment cash that the likes of Softbank are throwing away comes from private holdings like pensions and has little to nothing to do with the sovereign holdings of OPEC+ nations. The reason that it doesn't hold water are the bottlenecks on compute production. TSMC is still the only supplier of anything useful for foundation model training and their expansions only appear big and/or fast if you read the likes of Forbes.