@HedgieMarkets Anyone remember the dot com & telecom bust? •Massive overcapacity: The telecom boom of the late 1990s led to a frenzy of fiber deployment, leaving 85% to 95% of the newly laid fiber unused when the bubble burst in the early 2000s. •Widespread losses: Major companies suffered

@Marco20307855 🦔 The dot-com comparison is perfect as we're building massive overcapacity based on revenue projections that won't materialize.

@HedgieMarkets Interesting. Thanks for sharing. You earned a follow.

@BobEUnlimited 🦔 Glad you found it useful, thank you so much!

@HedgieMarkets They can't articulate it exactly yet, but the great unlock with all this spending, is your ROBOT SLAVE. The AI, the datacenters, the robotics, will one day converge on TAM for the robot helper market everyone will participate in.

@PowderkegPro99 🦔Robot slaves require the same impossible infrastructure scaling we're discussing, plus manufacturing capacity that doesn't exist.

@HedgieMarkets They're going to claim too big to fail and demand government taxpayer funded bailout!

@InfowarrioBOB 🦔The "too big to fail" playbook is already being written when JPMorgan commits $10 billion to critical infrastructure sectors.

@HedgieMarkets Yeah.A depreciating asset based on high electrical costs the population will refuse to stomach.

@MaralynBurstein 🦔When electricity costs are already forcing utilities to restart coal plants, the population will definitely revolt against higher energy bills for AI.

@HedgieMarkets What’s the alternative? Not pursuing new tech and AI is a nonstarter as some other entity /country will. So what the alternative given we have this new tech

@thunderswitch 🦔The alternative is sustainable development instead of burning hundreds of billions on infrastructure that can't be powered or cooled at scale.

@HedgieMarkets Good Story Need big interest cuts to keep the bubble from popping

@andre212nyc 🦔 I see your point, but rate cuts won't fix the fundamental economics when data centers need $1 trillion in revenue but only generate $20 billion annually.

@HedgieMarkets Artificial artificial intelligence driving Artificial Intelligence ? 🧐

@bomygandhi 🦔Exactly, circular financing where companies invest in each other then buy each other's products with that same money.

@HedgieMarkets E Dowd thank you again for reposting important info. My guru M Armstrong (read REAL AI, Socrates), warns already of a massive correction in the AI Bubble. is watching NVDA. When NVDA falls in oct/nov with a closing price in nov under 100, NVDA will never reach old highs again.

@JosephDeka 🦔NVDA under $100 would signal the bubble's end, but the real damage happens when companies can't refinance their AI infrastructure debt.

@HedgieMarkets Advancing technology? Who wants to invest in that? Crazy.

@GilKlein4 🦔There's a difference between advancing technology and lighting money on fire while ignoring physical constraints.

@HedgieMarkets This doesn't matter because AI is a competition between megacorps and governments. The economics of ballistic missiles are terrible, but we still built them as fast as humanly possible for 70 years straight.

@HedgieMarkets But but since oracle nvidia and OpenAI all swapping spit nobody paying for anything so everything gets built in top of OpenAI then either thru eventual economies of scale or price raises most likely thru agents is model

@HedgieMarkets You could have posted the article. It's about legacy datacenters and refers to similar build-out from Global Crossing. It's not relevant to the new Neo Cloud and the vertically integrated plays like IREN along with CIFR. Horrible take and you simply could have linked article

@HedgieMarkets Overcapacity? We’re not even close to having enough capacity to cover 2026 projections. They literally can’t build fast enough, countries are going to trade wars over this and you’re out here crying nonsense.

@HedgieMarkets @RamiMayron Totally irrelevant analysis. Completely discounts Moores law or the effects of Quantum computing on needed future computing power.

@HedgieMarkets Your math seems overly pessimistic to me. Let’s look it another way using a heavy-duty 128k-token context workload for inference. NVIDIA’s “Dynamo” chart shows tokens-per-second per megawatt at short contexts like 1k. It does not publish a 128k point. But we can conservatively

@HedgieMarkets I don't dispute a lot of what you say - I've though AI was the new http://dot.com bubble for the past year - but your argument to authority of a hedgie is bad enough, but Kupperman doesn't even beat Joe Retail in an S&P 500 index fund. https://hedgefollow.com/funds/...

@HedgieMarkets This is the most expensive prawn membership in the history of the world.

@HedgieMarkets I would imagine longterm the AI allows us to secure resources globally we otherwise wouldn’t have access to as well as reducing govt spending on social programs, longterm.

@HedgieMarkets It does seem like a bubble now. The only question is when it pops because these things don’t pop on common sense or they wouldn’t be bubbles. The crazy part is AI is going to radically change our lives over the next five to ten years and some companies will definitely win but

@HedgieMarkets Kupperman’s analysis seems to completely disregard downstream revenues, at least in the case of Xai. FSD for Tesla and intelligence for Optimus robots are likely to be enormous revenue generators. Government demand in defense and administration also have huge potential.