Ruchir Sharma's column: AI 'boom' should not prove to be a 'bubble' due to rising debt burden

Warnings about America's growing debt have been made since the 1970s, but none of the warnings proved to be true. However, now the unbridled debt is beginning to matter. Last month, it started selling government bonds around the world. And its first tangible effect may be that rising interest rates on US bonds put the brakes on the AI boom. If you look back at history, you will find that every major economic-bubble came to an end when the cost of credit for the companies at its center increased significantly. This includes the railway crisis that unfolded in the 1800s. In the last century, all the big bubbles burst in the era of central banking, when central banks increased their short-term lending rates sharply. But this time the situation is different. Normally, during a market frenzy, companies start taking heavy loans to invest more money in the increasingly popular investment sector. A bullish environment causes inflation to rise, interest rates to go up and the bubble bursts. Governments then step in and take responsibility for the debt of the affected companies and take loans themselves to speed up the weakened economy. In recent decades, the role of the government has changed to provide continuous stimulus even in good economic times. Even long after 2008, the government continued to increase debt rapidly. The US budget deficit in this decade has been around 6 per cent of GDP – more than double the average of the decades before it. In this period, companies avoided taking on too much new debt. Only in the last one year, large hyperscaler technology companies have started raising huge loans to finance the expansion of AI infrastructure on a large scale. The impact of overborrowing has been on government accounts. And the problems start from there. A serious warning comes from the interest rate on public debt, which has more than doubled in the last five years to above 3 percent of GDP. This is a new record for the US and the fastest increase in interest payments in any major developed economy. Rising concern about government finances — combined with other factors, including rapidly rising energy prices — has pushed government bond yields up around the world. This environment of high interest rates is also driving up the borrowing costs of AI companies. There are many hallmarks of a bubble in the AI boom. The bubble will continue to inflate until interest rates reach a level where borrowing becomes extremely expensive. My research shows that the yield on ten-year US Treasury bonds is the global benchmark for long-term borrowing costs and is currently at 4.8 per cent. When it crosses the 5 per cent level, the AI bubble may burst. This will signal the beginning of a tighter monetary era, in which it will be difficult to raise money for large projects of AI. When big tech companies have to compete for capital with a government that is paying more than 5 percent yield on its bonds, many companies may be out of the debt market. This year, the estimated annual revenue from the use of AI is about $ 200 billion, while companies are spending more than $ 1 trillion on data centers and other infrastructure. To bridge this gap, companies are relying on the issuance of new bonds and equities. It is feared that the rising debt burden on the US could weaken its superpower status and take away the crown of the dollar as the world's reserve currency. But America's main rivals are also struggling with similar problems on the debt front. At the moment, what needs to be monitored is how fast the ten-year Treasury yield crosses 5% and how much of a threat it poses to the AI boom. (These are the author's own views)