📌 Capex US hyperscalers are on track to spend $806 billion this year, with annual capex projected to reach $1.234 trillion by 2027. That kind of spending points to an AI infrastructure supercycle, not just another phase of cloud growth 👉 @GoldmanSachs

Railroads CAPEX ran 6% of US GDP ... wave after wave of bond defaults. Telecom, just $1T... also everyone went bankrupt... AI data centers? ... does history repeat... that is the question.

Hyperscaler capex will 2x in 2027 to ~$1.6 trillion (up from ~$800 billion this year) per recent reports. Demand for AI compute is incredible...even though stocks aren't reflecting it currently. However, given the widening supply-demand inbalance, bottlenecks like power, memory and optics will deepen. Ultimately meaning that companies like $BE, $MU, and $LITE will report astounding growth next year. Share prices are bound to follow upwards. Time will tell!

Hyperscaler capex is set to hit $1.3 trillion in 2027 as cloud industry backlogs top $2 trillion, according to Nvidia CFO Colette Kress. To capitalize on the surge, the company is expanding its long-standing partnership with AWS to scale its compute footprint even further.

AI capex is projected to reach $765B in 2026, passing oil and gas for the first time. Compute is now getting its first futures contract. Here is what has to be true for it to work…

AI capex boom = GDP growth tailwind? Math: massive corporate spending on semiconductors/infrastructure is driving imports (Asian chip manufacturing); U.S. export benefits remain muted Result? GDP’s import/export equation shows drag despite investment boom

The capex needs to be financed at that process has bumps. Recently the two $NVDA big off balance sheet credit guarantees (open AI and data center CDO) have been examples of bumps and the road is rocky ahead. That said paying 83bp a year or 4.15% over the contract life for protection for an NVDA Default in the next 5 years seems stupidly high imho