[KH Explains] Samsung chief says he is still ‘hungry’ for foundry growth

Samsung Electronics Chairman Lee Jae-yong on Monday put an end to the long-standing rumor that the memory chip giant might spin off its lackluster foundry business, saying he is still “hungry” for growth.

“We are hungry to grow (our) business, not interested in spinning (it) off,” the Samsung chief told Reuters upon his visit to the Philippines.

Regarding the delayed operation of the company's new foundry plant currently under construction in Taylor, Texas, Lee said, "That's a little bit tough because of (the) changing situation."

His comments came after Intel recently announced it would separate its loss-making foundry business, fueling the rumor that its chip rival, Samsung, may follow suit.

Experts here say it is only reasonable for Samsung to continue expanding its foundry business because it is the “future” of the company, especially in the era of artificial intelligence.

"If Samsung decides to spin off the foundry business, it will be a huge mistake," said Lee Jong-hwan, a system semiconductor engineering professor at Sangmyung University.

"The nonmemory chip market is much bigger than the memory chip market, where Samsung dominates. Samsung knows it cannot only rely on memory chips. Abandoning the business would be the same as abandoning the future," he said.

Money-losing business

In 2019, recognizing the need to diversify the company's memory-focused business portfolio, the Samsung chair unveiled an ambitious goal of leading the foundry business by 2030, outpacing the current No. 1, TSMC.

The foundry business, which separated from the company's System LSI business in 2017, seemed to be off to a good start backed with aggressive investment. In 2022, the company became the first chipmaker to begin chip production, using the advanced 3-nanometer node with cutting-edge Gate-All-Around transistor architecture.

In recent years, however, the limitations have become clear. As rival TSMC is capitalizing on the AI boom, Samsung's foundry continues to struggle with losses, low yield rates and a lack of Big Tech customers like Apple and Qualcomm.

In the April-July period this year, Taiwan-based TSMC maintained its dominant lead with a 62-percent share, while Samsung remained a distant No. 2 with a 13-percent share, according to market tracker Counterpoint Research.

Market analysts predict that Samsung’s foundry business will log an operating loss of over 1 trillion won in the first half of this year. For the entire year, the company is expected to record an operating loss of 2 trillion won, similar to that of last year.

With the company's overall chip business grappling with poor performance in the wake of a severe downturn in the global market, the idea of spinning off the loss-making foundry business continued to resurface.

"The idea of turning the foundry business into an independent entity has been there for a long time. But the foundry business does not have the money (to stand alone)," a Samsung executive said under the condition of anonymity.

Beom Jin-wook, an electronic engineering professor at Sogang University, explained there are merits to a spin-off, but it is an "unlikely choice" for the current Samsung when the foundry business continues to struggle to make ends meet.

"The foundry is losing money for years, but it can be sustained because it has access to Samsung's coffer and the pool of chip experts working for the other parts of the company's chip business," Beom said.

As of the end of 2023, TSMC had a total of 77,045 workers. Samsung's total workforce for the entire semiconductor division was 67,000, with 60 percent of them being part of the memory chip business. Samsung's workforce for the foundry business is estimated to take about 20 percent of the total.

Flexibility dilemma

To stay competitive in the growing global foundry business, Samsung should implement a significant shift in its organizational culture which is currently focused on memory chip production and is "rigid," Lee of Sangmyung University said.

“Having grown on the success of its memory chip business, Samsung’s culture is rather rigid, given how the company is used to making standard, complete products and then supplying them to the customers," Lee said.

"The foundry market, on the other hand, is completely customer-centric. The manufacturing and work process should become flexible enough to tailor to the client's needs."

In the future, it may be ideal for Samsung to separate the foundry business, when it gains more than 30 percent of the global foundry market for high-end products, Park Jae-geun, head of the Korean Society of Semiconductor and Display Technology and a professor of electronic engineering at Hanyang University said.

By high-end chips, Park was referring to those cutting-edge semiconductors utilizing a 7-nanometer process and under.

Experts agreed that spinning off the foundry business would dispel customers' concerns about having to give their highly confidential chip blueprints to Samsung, which may be their rival in other chip sectors.

“They would be willing to give their chip blueprints when the foundry entity is independent. Under Samsung’s current structure, clients are worried that their chip blueprints could leak into Samsung’s other chip divisions, which would likely be their direct competitors,” Park said.

For now, Samsung could consider spinning off the middle-end chip production business producing legacy chips using older technologies such as the 40-nanometer process, Park suggested.

“Big investments are required to produce high-end logic device chips like the manufacturing facilities should be there and they need money for research and development and buying cutting-edge extreme ultraviolet lithography machines," Park said.

“Making the middle-end chip foundry independent is a plausible option for Samsung, especially when you look at the cost competitiveness. It is too expensive to make the middle-end chips at Samsung’s advanced chip manufacturing plants in Yongin, with high labor and infrastructure costs," he said.