Syz the moment

$ORCL has now plunged 48% since its all-time high on September 10, a total market cap loss of $475 Billion 📉📉 Note that the stock is UP 6% after-markets on TikTok deal + OpenAI securing $100B in funding from UAE sovereign wealth fund. Source: Barchart @Barchart

Coreweave in blue Oracle in red Source: www.zerohedge.com

The chart from Epoch.ai GPU clusters reveals a 16x surge in Data center power demand. With 94% of infrastructure not yet built and power constraint ahead, hyperscalers and start-up race to control the future of AI. The US might lose the race to AI not because of the access to talents or capital but because of a lack of power capacity. Source: Markets & Mayhem

Amazon is in talks to invest $10B+ in OpenAI, pushing its valuation north of $500B. But this isn’t just about capital. It’s about control of the AI stack. 🔹 OpenAI would use Amazon Trainium AI chips 🔹 Rent massive AWS data-center capacity 🔹 Deepen infrastructure dependence beyond Microsoft This comes right after OpenAI restructured its relationship with Microsoft, unlocking deals with rival cloud providers. 💡 What’s happening behind the scenes: OpenAI already committed $38B over 7 years to Amazon servers Has $1.5T (!) in long-term infrastructure deals with Nvidia, Oracle, AMD & Broadcom Nvidia alone plans up to $100B in a multi-year partnership 🤔 Investors are uneasy. Many of these deals are circular — suppliers invest in OpenAI, OpenAI buys their hardware, and sometimes takes equity in return. Meanwhile, rivals aren’t standing still: Anthropic has raised ~$26B from Amazon, Google, Microsoft & Nvidia Amazon alone has put $8B into Anthropic since 2023 ⚠️ Key limitation: Amazon still won’t get rights to OpenAI’s most advanced models — Microsoft keeps exclusivity until the early 2030s. 🛒 Bonus twist: Amazon and OpenAI are also discussing e-commerce integrations, as OpenAI expands beyond chat into platforms like Etsy, Shopify & Instacart. 📌 Big takeaway: The AI race is no longer about models. It’s about chips, clouds, capital, and distribution — and Big Tech is locking it all down fast.

We have experienced swings on which models are gaining the most attention as Goldman's Jim Schneider illustrates the chart below... Source: zerohedge

Source: RBC, Bloomberg

Source: Conor Sen

Forget the hype. Look at the numbers. The AI sector will spend $400 BILLION this year. Revenue? A measly $60 BILLION. That $340 BILLION gap? It's filled with circular financing and off-balance sheet debt 💣 🔴 The CoreWeave Trap: Circular Financing 🔄 The Play: CoreWeave uses NVIDIA’s money to buy NVIDIA’s chips, then rents them back to NVIDIA. The Math: They are spending $20 BILLION to make $5 BILLION in revenue. The Debt: They have $14 BILLION in debt due next year and a staggering $34 BILLION in lease payments starting in 2028. 🔴Debt & Leverage 🧊 OpenAI's Burn Rate: They make $10B but need $50B just for their Oracle deals! They're projected to lose $15 BILLION this year. The only one making money? NVIDIA. Everyone else is buying chips on credit, praying for a future payoff. 🙏 🔴SPVx Meta hid a $27 BILLION data center build off their balance sheet using Special Purpose Vehicles (SPVs). 🔴The New CDO? Companies are taking GPU-Backed Loans, posting chips as collateral. What happens when the chip bubble pops? It’s a cascade risk Will we see something similar to the housing collateral crisis, but with sthis time with silicon ??? 📉 🔴Private Credit: The $1.25 TRILLION Blind Spot 🚨 The riskiest part is happening in the shadows: Private Credit. Private Equity firms have already lent $450 BILLION to tech and plan to lend another $800 BILLION in two years. Zero Disclosure. They operate outside of traditional banking scrutiny. Life Insurers (who hold your policies!) have $1 TRILLION tied up in this private credit gamble. If AI loans fail, private credit fails. If private credit fails, banks and insurers are at risk because everyone is connected. To make a long story short, the leverage is building... Source: Hedgie on X