Thank you Lufei for noticing my warning yesterday about tech giants' capital expenditures rising from cash flow statements to balance sheets. Today, I saw Oracle issue $15 billion in corporate bonds, but the main issue is that Oracle's financial data is actually quite average. To make a rough comparison, the top three US tech giants with the highest capital expenditures, Microsoft, Google, and Meta, all have net profits of over $20 billion per quarter, strong free cash flow, and debt to asset ratios of 44%, 27%, and 33%. The low debt ratios of Google and Meta are exaggerated. Oracle's net profit for the most recent quarter was only $2.9 billion, with negative free cash flow and a debt ratio of as high as 86%. The foundation is not as rich as the other three companies, and they need to invest heavily to seize a competitive advantage, so they can only use financial instruments. Of course, it is only the first giant to issue bonds to maintain capital expenditures at the beginning. It can be seen that the financial situation of the other giants is very healthy, with their annual net profits fully covering their capital expenditures. When it comes to vigilance, we can see how this situation evolves in the future. Of course, the core is whether the money spent can be earned back in the future. Lufei here https://(x.com)/lufeieth/status/1970832250309617 uses GPT to analyze the monitoring list, and the core is to look at the actual business data performance of the last two items in the future. 1. Company by company: CapEx/CFO, FCF, Net Debt/EBITDA, Interest Coverage Ratio 2. Frequency and scale of external financing (including project financing/joint venture SPV) 3. AI data center utilization rate (GPU available hours, job queue latency), unit computing power gross profit 4. Price/supply side signals: cloud AI pricing, GPU rental rates, shelf life, and speed of implementing concurrent large models