Interest rates pushed up by AI: can AI withstand them? (Park Jong-hoon's Knowledge in One Shot)

The speaker focuses on AI among five causes of the rise in US Treasury yields to 5.2% for the 10 year maturity (the highest since 2007) and 5.5% for the 30 year maturity (the highest since 2004), arguing that Bessent's and Warsh's optimism about AI has proved misplaced and that the Federal Reserve is caught in a dilemma of raising its policy rate to follow the market. The speaker reports that Bessent changed his position to acknowledge a short-term scramble for capital, while Warsh pointed to overheating in corporate bonds at the Jackson meeting on August (month 8) 28 and then joined the rate increase on September (month 9) 16. Citing projections of 4,200 hundred-million dollars in big-tech corporate bonds in 2027 (a 60% increase from the previous year), 4.1 trillion dollars in AI debt by 2030 (JPMorgan), and data-center investment equivalent to 2.5~2.7% of GDP, the speaker describes this as the largest investment wave since railroads and compares it with the much smaller 1% during the IT bubble. The speaker explains that the rise in the neutral interest rate from 2.5% to 3.2% has made high interest rates the new normal, and that AI pushes up prices through intermediate goods and electricity charges (a House vote of 417 to 3). The speaker notes that the spread on AI corporate bonds is 1.15 percentage points (versus 0.78 percentage points for investment-grade bonds overall), and that the S&P500 earnings yield matches the Treasury yield of 5.2%, leaving the compensation for equity risk effectively at 0. The speaker believes the bond market is reacting first, citing Alphabet's quarter 2 net income of 1,121 hundred-million dollars as equity valuation gains while its cash flow is -59 hundred-million dollars; uncertainty over converting Oracle's 6,640 hundred-million dollars in orders into cash, electricity disruptions in New Mexico, and refinancing burdens; and Nvidia's CDS rising from 0.4% to 0.8%, with trading increasing from 6.4 hundred-million to 69 hundred-million dollars. The speaker characterizes calls to moderate the pace as bluffing and argues that, given the game-theory and chicken-game structure, the vicious cycle will continue until monopoly or oligopoly emerges. Productivity, the way out, would have to rise from 2.2% to at least 4~5%, and the speaker contrasts a repeat of 1996 with the risk of a scenario like 1847, depending on the direction of oil prices and developments in the Strait of Hormuz. The speaker says the Korean stock market is more sensitive to interest rates than the US market because of the HBM–data-center link, and that the exhaustion of share buybacks in mid-October (month 10) will be a test. The speaker presents whether AI can withstand the interest rates it has created, whether productivity improves first, and whether the Hormuz situation is resolved as points for investors to consider. All directional outlooks are the speaker's conditional judgments, not definitive predictions of short-term prices.