Meta-Driven AI Overinvestment Fears Diverge From Reality as Memory Shortage Persists
Summary
- Eugene Investment & Securities said the AI overinvestment debate is largely disconnected from reality and that ROE in AI computing rentals has risen.
- The firm said investment is continuing, with Meta expanding its AI data-center infrastructure to 14 gigawatts next year from 7 gigawatts this year.
- It emphasized that continued AI data-center investment means the memory supply crunch and memory bottleneck will be difficult to resolve quickly.
Forecast Trend Report by Period
Eugene Investment & Securities Report
Concerns that Meta Platforms Inc.'s move into AI cloud infrastructure points to excessive investment are largely disconnected from reality, according to Eugene Investment & Securities. The more important question is not whether idle equipment exists, but whether there is a market where that capacity can be turned into cash. The firm also said continued investment in AI data centers will make it difficult for the memory shortage to ease anytime soon.
Lee Seung-woo, head of research at Eugene Investment & Securities, wrote in a July 13 report that debate over Meta's external sales of surplus computing capacity had once again fueled the "AI bubble" argument. But he said that also shows returns on equity from AI computing rentals have risen enough that even older graphics processing units no longer used to train the latest models can still be sold outside the company.
"What matters is not the existence of idle equipment, but whether there is a market that can monetize it," Lee wrote.
He also said concerns in some parts of the market that excess computing resources would slow data-center investment are out of line with reality. Meta recently announced plans to begin construction of a 1-gigawatt data center in Canada that can be expanded to 1.8 gigawatts. It has also said it plans to double its AI data-center infrastructure to 14 gigawatts next year from 7 gigawatts this year.
"U.S. semiconductor shares have risen, but South Korea's top two names, Samsung Electronics Co. and SK Hynix Inc., have fallen sharply," Lee wrote. He said the debate over Meta's surplus computing capacity, combined with distorted supply-and-demand dynamics in the Korean stock market, had created an extremely confused market backdrop.
Still, he said both top-down and bottom-up analysis point to the same reality: investment in AI data centers is continuing, their cash flow remains intact, and memory bottlenecks will not be resolved easily.
Ko Jeong-sam, Hankyung.com reporter, jsk@hankyung.com
Korea Economic Daily
hankyung@bloomingbit.ioThe Korea Economic Daily Global is a digital media where latest news on Korean companies, industries, and financial markets.