SpaceX’s record IPO plan shows investors are keen to keep pouring money into AI in general, despite alarm bells ringing across the economy.
That leaves analysts wondering, “If soaring inflation hurts growth, where will the money come from?” Or if the implementation of artificial intelligence turns out to be less profitable than expected?
historic influx
Investments by AI institutes are at historically “unprecedented” levels, with projected spending by 11 of America’s top companies over the next 12 months equivalent to nearly 3% of U.S. GDP, said Raphael Gallardo, chief economist at Paris-based wealth management group Carmignac.
Earlier this year, chipmakers and other high-tech hardware companies hammered stock markets around the world, shaking confidence in their spending spree. But “so far, those concerns have been largely ignored by the market,” said Adam Sirhan of 50 Park Investments in New York, following a reassuring profit report despite ongoing wars in the Middle East.
“If you look at the actual returns, those fears are unfounded and many companies are actually increasing their investment in AI,” Sarhan told AFP. Google, for example, announced this week that it will raise up to $80 billion to massively expand its AI infrastructure.
The company said there are “computing constraints in the short term.” In technical terms, it means not being able to build the necessary infrastructure fast enough to meet demand. Meanwhile, SpaceX is aiming to raise $75 billion in an initial public offering (IPO) scheduled for next week, making it the company’s largest IPO to date.
Rivals OpenAI and Anthropic, behind ChatGPT and Claude respectively, are expected to follow suit in the coming months, with both companies valued at a whopping $1 trillion.
devour chips
Beyond U.S.-based chatbot makers, companies around the world are benefiting from the AI rush, especially chip makers that provide computing power.
South Korea’s benchmark Kospi stock index, for example, has nearly doubled in value since January, driven by chip makers Samsung Electronics and SK Hynix, both now trillion-dollar companies.
These two companies alone account for half of Kospi’s market capitalization.
“The fact that these two companies hold such a large portion of the market highlights how concentrated their dependence is, and that is the biggest risk factor,” said Kim Dae-jong, a professor at Sejong University.
In Taiwan, TSMC, a supplier to AI chip specialist Nvidia, accounts for 40% of the Taipei stock market, while Japanese technology investor SoftBank this week overtook Toyota to become the country’s most valuable company.
In the U.S., hot demand for Micron and Intel chips has more than doubled stocks so far this year, while European stock benchmarks have soared, thanks in large part to Infineon and STMicroelectronics.
too hot to be comfortable
However, there are signs that market expectations are outpacing companies’ ability to respond.
This week, shares of U.S. chip specialist Broadcom fell sharply after forecasts for more than 200% growth in third-quarter chip sales fell short of expectations, even though second-quarter profits nearly doubled to $9.3 billion.
“The support from the huge capital inflows into AI and semiconductor stocks is waning, exposing the often extreme valuations in these sectors,” said Andreas Lipkow, an analyst at CMC Markets.
“In the best case scenario, investors will take profits and the market will have time to consolidate ahead of the summer break,” he said, especially if investors sell tech stocks to buy new SpaceX shares.
“Otherwise, a significant correction in international stock markets in the short term remains likely,” he said.
“These companies are cash cows and we are in one of the biggest investment cycles in history,” said Frédéric Ducrozet, head of macroeconomic research at Pictet Wealth Management in Switzerland.
But so far, none of the big three AI companies — SpaceX, Anthropic and OpenAI — are profitable, so “we need to be more cautious,” he said.
AI VS stagflation?
Analysts and policymakers worry that enthusiasm for AI will not escape the pull of rising energy costs (data centers consume large amounts of electricity) and slower overall growth.
Carmignac’s Gallardo said that in the U.S. alone, AI investment now accounts for nearly nine-tenths of overall GDP growth, overshadowed by weak consumer demand and rising costs for small and medium-sized businesses.
ING economist James Smith added: “AI-related spending is a big part of the US growth story. The same handful of companies are raising money, buying chips, leasing computing, and reaping the revenue from each other.”
“However, if you remove AI, the fact remains that the rest of the country’s private nonresidential investment has declined year over year for six consecutive quarters,” he said.
And the situation could get even worse if the U.S. Federal Reserve, European Central Bank and other central banks raise interest rates to curb energy-driven inflation, which many analysts believe is inevitable.
