deafturtle

God dammit Clement 🫠 Druck shouted this book out speaking with hedge funds last month. I rushed to read it, as I see my friend did here as well. Great read, very unfortunate consistencies between now and then. Hope it's all wrong, but what can we do other than read the tea leaves? The whole book looks at the relationship as if Data Centers now are the railroads back then. Railroads in the 1860s are the AI infrastructure of today, transformative technologies that require enormous amounts of capital upfront. Their finance mechanism of choice? MASSIVE BOND ISSUANCES to fund construction. When investors began questioning whether railroads could generate enough revenue to service their debts, financing dried up. The big red flag in similarities is the building of infrastructure ahead of demand as they bet the building of said infrastructure would create demand. They were developing railroads in sparsely populated areas thinking they'd attract settlers. Sometimes it worked like a charm, other times companies went bankrupt before the anticipated economic activity came through. The analog works as such: -Hyperscaler builds a massive data center assuming AI inference demand will explode. -It purchases GPUs based on the expected utility several years in the future. -Signs long term power contracts and infrastructure agreements. -Financial projections assume customers will pay enough to justify the enormous investment. WHAT BROUGHT DOWN THE WORLD ECONOMY BACK THEN: the anticipated economic activity failed to justify the lofty financial projections and the enormous debt accumulated to fund them. The 1873 crisis wasn't necessarily about railroad stocks collapsing due to lack of growth, it was more about the railroad bond market losing its ability to finance the continued expansion. The railroad businesses were booming! They were simply not growing enough compared to the aggressively lofty expectations. The aspect most striking to me was the net positive that railroads did for society and still lead to the market influx. The railroad boom was absolutely correct in its future utility. They DID transform commerce as they connected national economies, reduced shipping costs and overall helped create modern capitalism. Just like AI WILL cure diseases and all these other amazing things. So the technology was ABSOLUTELY evolutionary, but the capital allocation is what lead to the market top. THE BIG DIFFERENCE THOUGH, between now and then, were the financiers. Hyperscalers got a lot of cash and generate a whole lot more whereas the financing in the bond offerings simply dried up back then when investors began to doubt the growth projections. Jay Cooke got cooked back then but we know guys like Zuck will spend forever. They are all in and will keep spending to keep this trade in motion. I have no idea what happens next, I am not calling anything. I have been on margin long for weeks until today. The one thing i DO KNOW, is that memory looks like absolute shit after today. $DRAM here was looking like a base for weeks and with today's failure is resembling much more of a stage 3 top. Again, I am not saying this is set to happen. I am simply explaining @Clement_Ang17 's post. If this post triggers you in any sort of way ... BLAME CLEMENT!