This one is especially for AI Infra folks $IREN $NBIS $MRVL and even $META We are no longer seeing a clean selective capital formation scenario and thats probably why we have seen so many macro guys that were quite for so long, suddenly rushing out the door claiming the king is naked The market has rotated toward what we can call a scarcity crowding out regime: geopolitical + logistical bottlenecks are raising the value of existing physical capacity while the high cost of capital is beginning to suppress leadership in the next layer of AI and industrial capacity expansion. Macro folks spot this because they are always looking at commodities, rates and they never lose sight of tech (especially the high flyers). What they are seeing now is that the AI physical capital cluster has broken relative leadership, while barrels exist but moving and refining the barrels is now very hard, and rates keep testing Bessent (also keep a look at JPY - not great). All in all, it does not surprise me that they are now embarking on two thesis archetypes: i) The technology can be transformative while the people financing the infrastructure destroy capital. He compares data centers to shale's enormous spend, but terrible ROIC, and says the buildout can continue long after the economics stop making sense. ii) Assumption invalidation: compute scarcity, unobstructed infrastructure deployment, and abundant financing. The argument is now that planned capacity is huge relative to estimates of 2030 demand, while bottlenecks mean much of the announced pipeline may never get built (which is actually bullish I know). Some of this are actually sensible questions, like whether the data center itself generates enough AI revenue to justify its depreciation and capital cost. But the real point here is that they are watching macro tipping, and they know this is the time to ring the bells and be the next macro oracle: "the guy who saw it coming." Obviously, the AI buildout contains characterics of previous bubbles and it is happening in a cyclical sector, which raises a lot of eyebrows, but what we are seeing now from the macro crowd is simply a great risk/return statement opportunity: they speak now because the clouds are getting darker - it helps if your bubble thesis has help from broader macro factors. Their risk is basically if that macro squeeze reverses before hyperscalers actually cut spending, the whole construct can rerate hard. Or even if freight and product tightness ease, inflation pressure fades, the 10 year finally falls and suddenly marginal AI projects that looked questionable start clearing the hurdle again.