When Japan decided to pick a fight with South Korea last month, it dug deep into the supply chain to impose sanctions on three obscure materials made by a handful of Japanese companies few have ever heard of.
The most powerful weapon in Tokyo’s campaign against its neighbor turned out to be a half dozen or so niche firms with names like JSR Corp., Shin-Etsu Chemical Co. and Tokyo Ohka Kogyo Co. They make fluorinated polyimide, hydrogen fluoride and photoresist: essential ingredients for the manufacture of the displays and semiconductors that go into every piece of modern consumer electronics, from Apple Inc. iPhones and Dell Technologies Inc. laptops to the full range of Samsung Electronics Co. devices.
How did they become so indispensable? And how did they manage to stay on top even after their Japanese clients ceded the chip and display markets to Taiwanese and South Korean rivals? The answer lies in a series of well-timed investments decades ago, combined with a willingness to explore foreign markets and an unceasing refinement of manufacturing standards too exacting for anyone else to try and match.
“JSR is an interesting case in that they became big in photoresists because they succeeded overseas first,” said Damian Thong, an analyst at Macquarie Group Ltd. “And much of this success was because of the strategy of one man — Mitsunobu Koshiba.”
The JSR chairman’s story shows just how hard it would be for a newcomer to fill the shoes of one of these suppliers. Koshiba spearheaded the company’s pivot into photoresists, a light-sensitive liquid used to imprint circuits as narrow as a few strands of DNA onto silicon wafers in a process called lithography. Gadgets keep getting slimmer, more powerful and cheaper because chip companies are able to etch ever smaller circuit patterns onto silicon. When it comes to the most advanced chip processes, JSR is one of the few that can deliver the goods.
When the 25-year-old Koshiba joined JSR in 1981, the company’s biggest business was still tire rubber. (The name is an abbreviation of Japan Synthetic Rubber.) As luck would have it, photoresist at that time used resins that JSR had access to for its existing business, and the company saw an opportunity to break into a new growth industry. Japanese semiconductor makers were just beginning their rise to global dominance, and suppliers were positioning themselves to go along for the ride.
The problem for JSR was it didn’t belong to any of the local keiretsu, a grouping of suppliers that receives preferential access to contracts. And the company was also up against Tokyo Ohka, or TOK, the first in Japan to manufacture photoresist. By the mid-1980s, TOK controlled as much as 90 percent of the domestic market.
“As a neutral company without keiretsu affiliations, we had to look outside Japan,” Koshiba said in an interview, outlining JSR’s decadeslong rise but declining to talk in detail about sensitive trade negotiations now underway between Tokyo and Seoul.
JSR’s decision to get into that market was bold, but Koshiba seemed like the right person for the job. He’d spent two years studying materials science at the University of Wisconsin-Madison on a Rotary Club scholarship, was one of the few English speakers at the company and was eager to work abroad. In 1990, JSR sent him to Belgium to set up a photoresist joint venture with biopharmaceutical giant UCB SA. The goal was to target the American market.
As timing would have it, JSR was going overseas just as Japan was approaching the peak of its semiconductor prowess. That same year, NEC Corp., Toshiba Corp. and Hitachi Ltd. were the world’s biggest chipmakers, pushing aside Intel Corp. and Texas Instruments Inc. Japanese firms occupied six spots in the industry’s top 10 ranking by revenue, a level of concentration that hasn’t been matched by any country since, according to IC Insights.
Japan’s seemingly unshakable control of the computer memory market prompted right-wing politician Shintaro Ishihara and Sony Corp. co-founder Akio Morita to pen “The Japan That Can Say No,” a book arguing for a more muscular foreign policy. In an eerie echo of recent events, the authors contended that the Japanese government had the power to determine the outcome of the Cold War just by directing its national companies to sell the chips used in intercontinental ballistic missiles to the Soviets instead of the U.S.
But the Cold War ended before that theory could be tested. Over the following decade, personal computers overtook ICBMs as the primary destination for chips and demand shifted to prioritize low unit costs over military-spec quality. By 2006, Samsung had risen to No. 2 on the list of the world’s biggest chipmakers, with South Korean compatriot SK Hynix Inc. ranking seventh and only three Japanese names remaining among the top 10.
For JSR, the turning point came in 2000. Koshiba, who was based in California at the time, recalls being dragged into an emergency meeting on a Sunday wearing a T-shirt and shorts. Word was a rival company was about to clinch an agreement with IBM for joint research on a next-generation photoresist material. “Get it back,” he was told. Koshiba leaned on the network of American industry contacts he had spent a decade building, people who had known him through the worst of U.S.-Japanese trade tensions. Within a month, IBM signed with JSR.
“Without that deal, we wouldn’t have gotten to No. 1,” Koshiba said.
In lithography, the formula for shrinking transistors has only two levers: increase the light power or use a lens that lets more light through. Every time the chip process shifts to a higher-energy band of light, resist makers have to go back to the drawing board, opening up new opportunity. The research partnership with IBM ushered in the fourth such shift since integrated circuits replaced vacuum tubes in the 1970s, and JSR rode it all the way to the top.
The company now commands about 40 percent of the market for the latest generation of resist used in mass production. It also supplies more than 30 percent of the photoresist for 3D NAND, the most advanced flash memory chips, which are among the few product lines where Japan still competes with Korean rivals. In 2019, JSR is expected to generate about three times the revenue and five times the profit it did in the early ’90s.
What makes this business inaccessible to newcomers is the extreme degree of purity and quality demanded by customers. TOK says a single drop of coffee in two Olympic-sized swimming pools would be considered an unacceptable defect. JSR’s analogy is to a handful of tainted golf balls being enough to spoil a batch the size of the entire Japanese archipelago.
In addition to being technically challenging, the markets these companies operate in are small and don’t promise fantastic growth. According to research firm Fuji Keizai Group, the industry’s sales rose just shy of 8 percent last year to $1.3 billion. Koshiba jokes that even the market for ramen noodles is bigger than that.
“To re-create JSR, you basically need to spend as much as they did in the past 20 years on R&D and relationships, and also rebuild their reputation,” Macquarie’s Thong said. “These materials are used in such moderate quantities that to rebuild the whole infrastructure is probably not worth the investment.”
And that’s the irony of the current situation. By stoking trade tensions, Japan may encourage its neighbor to subsidize competition to JSR and TOK that wouldn’t make sense under normal market conditions. It’s a matter of survival: Korean corporations now depend on Japan for over 90 percent of all the fluorinated polyimide and resists they need, and 44 percent of hydrogen fluoride requirements, Societe Generale estimates.
For the time being, JSR and TOK retain dominance over one prized material that keeps the consumer electronics industry ticking. According to South Korean Prime Minister Lee Nak-yon, Japan has approved exports of photoresist for the next-generation of lithography currently under development by Samsung and Taiwan Semiconductor Manufacturing Co. But one of Japan’s last strongholds of tech industry domination may be under threat.
“They have the engineers, and once national pride is involved they can possibly make it even if it loses money,” Koshiba said. “We don’t have an impregnable wall.”