Marc - congrats on the year. 96K+ subscribers and maintaining that depth of research is real execution.

Your "Financing the AI Boom" piece in January calling $288B capex that ended up at $400B+ is the kind of early-cycle call that matters. I've been digging into the back half of that story - the capex bill is here, but the FCF justification isn't showing up yet in most hyperscaler financials. Amazon's burning through everything it generates. Meta's $70B capex vs $45B FCF. The math only works if AI monetization materializes before the treadmill breaks them.

Your "Bubble Trouble 2" point on the H100 Rental Index as a coordination mechanism is sharp. I hadn't thought about it as the ABX equivalent, but you're right - once price discovery becomes transparent and tradeable, mispricing can unwind fast. The circular funding structure (banks lending to GPU buyers who rent to AI companies who raise from the same banks) has that 2006 feel.

On private markets eating public: your Apollo and KKR coverage is ahead of where most people are. The "mini Berkshire" framing isn't cute - it's the actual playbook. The question I'm watching is whether permanent capital structures survive the next credit cycle, or if they're just getting away with illiquid marks because nothing's forcing a bid yet.

Looking forward to 2026. You're one of the few doing the plumbing work that actually matters.