"You kill and save my stock account and you're doing it alone"...AI version mixed with good performance and debt

사진 확대

Artificial intelligence (AI) is making the global stock market cry and laugh. Demand for AI semiconductors is still strong and some key companies are performing at record-high performances, but the question of whether the AI boom is really money is growing at the same time as the investment scale and debt snowball. The atmosphere in which stock prices have been jumping just under the name of "AI" for a while is changing. Investors have now begun to weigh not only growth potential but also performance, debt and cash flow.

The stock market clearly showed this last week. According to Reuters, Nvidia shares fell about 8% and TSMC, the world's largest foundry company, fell about 4% in the week to the 15th. The fall of so-called "NeoCloud" companies, which lend graphic processing units (GPUs) to AI data centers, was even greater. This is due to higher-than-expected prices in the U.S. and increased concerns over prolonged high-interest rates, and the spread of vigilance over the profitability and financing structure of AI infrastructure investments.

However, it is difficult to say that demand for AI semiconductors has cooled. TSMC's third-quarter net profit announced on the 15th was 763.1 billion dollars, up 68.7% from the same period last year. It is the highest net profit ever for six consecutive quarters. TSMC expects sales to increase by about 44% this year from the previous year, and it also maintained its annual facility investment at $60 billion to $64 billion. Demand for advanced semiconductors for AI and high-performance computing is still boosting performance.

Orders from semiconductor equipment manufacturers are also solid. Dutch ASML posted 11.7 billion euros in revenue and 4.96 billion euros in net profit in the third quarter. ASML plans to expand its production capacity by 30% in 2027 and 30% again in 2028, but the company believes that this will not be enough to keep up with demand. Strong orders from ASML have somewhat allayed concerns that investment in AI semiconductors is sharply declining.

On the other hand, things will change if you go down to data centers and cloud companies. Coreweave, an AI cloud company invested by Nvidia, has recently been at the center of market instability due to huge data center investments and borrowings. Reuters reported investors' concerns that Coreweave's debt is worth about $50 billion and that the data center construction boom is largely dependent on credit. If U.S. interest rates remain high, the interest burden can increase faster for companies that build data centers through borrowing.

Investors' standards for AI companies are also changing. Investors are starting to see how much real cash they can earn while covering not only growth rates but also enormous computing and power costs. The fact that some companies' listing plans, which had been pushed forward amid the craze for investment in AI and data centers, have been postponed or withdrawn one after another shows this atmosphere.

There is another variable in the Asian stock market. It is the pursuit of technology by Chinese semiconductor companies. Global semiconductor stocks, including South Korea, were shaken this week by reports that Chinese memory company Changsin Memory Technology (CXMT) has made progress in new memory technology. This is because it is difficult to guarantee whether memory companies' high profitability will continue even if AI demand increases if Chinese companies' pursuit accelerates.

It is difficult to say that the recent AI investment boom has ended just by adjusting. Semiconductor orders and performance remain strong. However, the market no longer sees AI companies as a lump. Stock prices of semiconductor and equipment companies with strong cash flows and companies that expand their data centers with borrowings are divided. AI is still a key theme in the stock market, but "how much you earn and how much you owe" is now more important than "Is it an AI company?"