Nvidia just spent $20 billion to buy a company that missed its revenue target by 75%. This panic acquisition of Groq signals the beginning of the end for the AI bubble - here’s why the money is running out in 2026.
Timestamps:
0:00 - Nvidia’s $20 Billion Panic Buy
0:43 - What is Groq? (Not Elon’s Grok)
1:21 - How LPUs Work vs GPUs
2:09 - The Grocery List Metaphor
3:24 - GPUs vs LPUs Explained
3:54 - Groq’s Product Market Fit
4:50 - Formula One Use Case
5:34 - Groq’s Insane Financials
6:03 - $1.5 Billion Saudi Investment
7:25 - 75% Valuation Haircut in 4 Months
8:27 - Nvidia Buys Groq for $20 Billion
9:16 - Why This is Bubble Behavior
10:18 - The AI Hardware Crisis
11:19 - Nvidia’s Monopoly Problem
12:01 - The Power Grid Bottleneck
13:13 - Electricity Cost Increases 250%
14:52 - Data Center Deals Explained
16:11 - Who Pays for AI Power Usage
17:01 - Senate Investigation (Won’t Matter)
17:51 - Unit Economics Breaking Down
18:02 - The Circular Money Flow
19:37 - Nvidia’s Vendor Financing Scheme
20:30 - OpenAI’s $100B Nvidia Investment
22:17 - 24x Return on Investment
23:26 - OpenAI is Not Profitable
24:32 - $75 Billion Annual Burn Rate
26:05 - Needs $200B Revenue to Break Even
27:01 - The Grok Lifeboat
28:08 - Labor Displacement Reality
29:29 - H-1B Visa Strategy
31:04 - Megacorp Layoff Pattern
33:14 - Salesforce CEO Walks Back Claims
34:34 - AI Tracking in Performance Reviews
37:01 - MIT Study: 95% Zero ROI
38:15 - Nvidia Crashes 3.5%
39:08 - Signs of the Bubble Pop
40:01 - Timeline: Q1-Q2 2026
41:21 - AI Isn’t Going Away
42:01 - Final Thoughts & Substack
