Publikuj

#AIInfraEarningsWatch I think AI infrastructure is entering the part of the cycle where revenue growth alone stops being enough. That matters for names across compute, data centers, networking and power. The first stage of the AI trade rewarded almost anything attached to capacity expansion. The next stage should be much more selective. I’m watching three things now: How much contracted demand actually converts into revenue. How much capital has to be spent to deliver that revenue. And whether pricing stays strong once more capacity comes online. That distinction is important. A company can have enormous AI demand and still destroy shareholder value if every new dollar of revenue requires an uncomfortable amount of debt, dilution or capex. So I’m becoming less interested in: “Who has the biggest AI backlog?” And more interested in: Who converts AI scarcity into durable free cash flow? That is probably where the next separation inside the AI infrastructure trade happens. #CPIPPIEaseFedSplit $BTC $ETH $SPCX