The Fast and Furious Rise of Korean Semiconductors

The Fast and Furious Rise of Korean Semiconductors

Show Video

the Korean semiconductor industry is one of the most extraordinary examples of industrial ketchup in modern history despite a strong presence in electronics in 1982 just a single Korean cheboo produced their own semiconductors less than 10 years later Korean companies led the world in semiconductor manufacturing particularly in memory a remarkable achievement made possible by Blood Sweat and Technology transfer in this video we look at South Korea's Fast and Furious entrance into the semiconductor industry but before we go on I want to talk a bit about this video's sponsor blinkist blinkist offers over 6 500 titles in 27 different categories with more being added all the time in just 15 minutes you can get powerful insights into different topics while going about your everyday life learn about what is happening in the world around us or find something new to inspire you if you like what you get from asianometry I think you will like what blinkus does too the goals are the same a lot of information or nice story that you can enjoy in a single 15 to 20 minute chunk you can listen while you're doing some chores around the house that is exactly what I do I put on a blinkist book right before I start washing the dishes it helps the time go by faster today's video is about Korea's semiconductor industry but I am already assuming that you are at least a bit familiar with semiconductor history over in the United States if you need a brushing up in the topic then I recommend this blink on Gordon Moore he was one of the Giants of the semiconductor industry the more in Moore's Law you obviously need to know about this so go check it out I also want to talk about the blinka spaces feature it is like a little shared reading room you make a space for your friends or family then you put blinks you love into that space for as long as you wish everyone in the space can access the titles in it even if they don't have a premium subscription get 25 off blinkist annual premium start your seven day trial today by clicking on the link in the description below thanks to blinkus for sponsoring the channel Korea semiconductor manufacturing industry kinda begins in the 1960s I say kind of because back then it was mostly foreign multinationals using cheap Korean labor to help assemble their products the first was an American firm called Comey that in 1969 Fairchild semiconductor Motorola and cygnetics these American companies fully owned their Korean subsidiaries with little technology transfer so it is hard to call any of this a truly domestic industry during this period of time it would not be until the mid-1970s when the first chebor The conglomerates Who spearhead Korean industrial power actually entered semiconductor Manufacturing in 1974 Samsung acquired 50 of a struggling five-year-old IEC startup called Korea semiconductor Inc they bought the rest a few years later Korea semi had been the country's first company to produce chips using wafer fabrication the acquisition came at the behest of Lee Kun hee son of the Samsung founder and chairman Lee biancho Samsung first leveraged Korea semiconductors expertise to help Supply chips for their electronic watches and televisions in 1979 president was assassinated but the military General Chun du huan quickly took his place his leadership position shaky Chun sought to establish his economic credibility by Shifting the country away from its traditional Reliance on low-cost labor a 1980 task force did some research about future Industries to enter and found three semiconductors computers and Telecommunications two years later in 1982 the Korean Ministry of Science and Technology announced a nationwide r d program to upgrade Korea's technology economy semiconductors were slotted to receive the largest share of the some 500 million USD of r d dollars to be funded over the next five years it was an indication of the government's belief that semiconductors were the number one target the treble took quick notice of this and led the way on February 1983 chairman Lee Byung chor called a newspaper editor and said no matter what it takes Samsung will enter the semiconductor business so please deliver the news to the readers of your newspaper this ambitious announcement backed by 100 million dollars of Samsung's own money would later be known as the Tokyo declaration since chairman Lee was in Tokyo at the time the other Tribble followed a month after the Tokyo declaration chairman Chung Joo young of Hyundai made a similar announcement and Hyundai semiconductor was founded shortly thereafter gold star semiconductor essentially LG had been founded a few years before all of this but they ramped up about the same time but what semiconductor product to build after a lot of debate between the various product lines out there Lee and the rest of Samsung settled on memory specifically Dynamic random access memory computers use this memory as the main memory for their systems it has far more capacity than on-chip memory static random access memory or SRAM but as a result is slower the reason Samsung chose to compete in Dynamic was that it was the most widely used semiconductor product and so had the largest market they also liked that the memory structure was simple and repetitive memory chips are classified based on how many bits there are on each chip for instance 1024 bits on a memory chip and we have a 1K chip the trend is to quadruple this capacity about every two to three years Moore's Law if you recall but for these exact reasons the memory Market is very competitive the product is a commodity and its buyers computer makers and the like have freedom to switch they not only want the newest hottest thing but also at the lowest possible price lagging the market is a good way to lose a lot of money you want to get to the head of the line you want to stuff more memory cells onto the same space than your competitors can many especially those in Japan were skeptical that the Koreans would be able to succeed in semiconductors they cited the small domestic Market little experience and large-scale Integrated Systems or LSI and insufficient capital but Samsung felt that they had an edge they already had extensive management experience in the electronics thanks to their previous work in televisions and radios and being one of Korea's richest chebour they also had plenty of money the Koreans entered the semiconductor industry quite late they had a lot of catch-up to do The Logical first thing to do if you want to catch up is to acquire technology someone else already knows how to make semiconductor so you get them to teach it to you Japanese companies had helped Samsung and the other chebo enter the electronics markets in the late 1960s but this time the Japanese turned Samsung down rightly seeing them as a future competitor this did not prevent Samsung from hiring Japan's retired engineers and Consultants however this became a major outlet for Japanese semiconductor expertise to slip away to Korea and there was a single exception Samsung managed to strike a technology transfer agreement with sharp for a memory process node this was facilitated by Dr Tadashi rocket Sasaki raised in Taiwan rocket Sasaki helped surface the idea for the Intel 4004 microprocessor research Sharps LCD screen technology gave masayoshi's son his start and lived to be 102 Legend sharp was an outsider to midi's vlsi project they weren't part of the club but Dr Sasuke nevertheless received great criticism for breaking the Japanese technology boycott to Korea with Japan largely unavailable the Koreans heavily targeted the United States the three companies set up r d centers in the United States there they aggressively recruited american-trained Korean talents with Korean nationalism and insanely High salaries in some cases semiconductor Engineers were paid salaries over three times as high as chairman Lee's gold star and Hyundai did the same too collectively the three companies invested a billion USD into their American efforts to acquire modern technology train their staff and coordinate with potential partners eventually Samsung managed to purchase a technology license for 64k dynamic memory designs from a small struggling startup in Boise Idaho called Micron they sent eight Engineers to Idaho for technology training again Hyundai and gold star did the same too from 1982 to 1986 there was a total of 53 technology transfer agreements signed by Korean companies 48 out of that 53 were related to semiconductor production and the United States was a transferring partner in 36 of those 53 nearly 70 percent Japan was a distant second with 13. another underrated part about the Korean semiconductor entry was equipment almost all of Samsung's equipment was imported many of it came from American and Japanese suppliers like Applied Materials Canon and Nikon South Korea being an American Ally and with policy focused on Japan at the time the semiconductor manufacturing equipment was largely available on the open market this High equipment import ratio continued Korea imported some 89 to 94 of their equipment deep into the 1990s but Samsung and Hyundai didn't see a lack of proprietary equipment as a disadvantage in fact they thought that casting the widest possible net for the best technology no matter where it came from would be one of their strengths over the Japanese I want to drill down a little deeper into this last point so the Korean semiconductor industry like that of Japan and taiwan's was built on top of knowledge expertise and Equipment transferred or purchased from the United States however I don't want to downplay the incredible job these companies did anyone can open a checkbook and pay big bucks for expensive equipment what Samsung LG and Hyundai did was to marry that acquired technology with a labor cost advantage against the US and Japan as well as ridiculous execution here's another way to put it selling semiconductors is like selling burgers you get good equipment use the best pure ingredients follow the recipes monitor quality and make improvements as necessary the Korean companies did this better than their competitors and that is how they got ahead simple as that in 1983 Samsung announced that it had produced its first 64k dram using micron's technology produced just a year after the Tokyo declaration Samsung had managed an incredible achievement and shocked the world they had gone into crisis mode staff and two competing teams in Korea and the United States work 24 hours a day to bring this product to the market Samsung managed to launch their memory Fab in just six months American and Japanese Fabs traditionally took three times as long as that however 64k was not new then and 1985 was a bad semiconductor year being about five years behind the Leading Edge Samsung did not sell much new product in response to Samsung's struggles LG and Hyundai decided to slow down their Investments and Daewoo you thought I forgot about them right I might as well have Daewoo Electronics fell behind after the 1985 market crash and never recovered I won't mention them again despite this macro economic uncertainty Samsung fearlessly plunged ahead to 256k this product would be completed at the end of 1984 again with micron's help Samsung had gone all in they invested three times more money in its semiconductor business than the division generated in Revenue were not for the cash flow from the group's other businesses further development could not have been possible Samsung's story is pretty well known but what about Hyundai and LG for Samsung semiconductors were everything the Hyundai and LG had other businesses to tend to pulling resources away from their semiconductor divisions Hyundai's semiconductor program was founded in 1983 already a few years behind Samsung worse yet the company shot itself in the foot at the outset the company initially focused on SRAM a special type of memory that you can find on CPUs and such Hyundai did this because they did not want to compete head to head with the big Japanese conglomerates but as it turned out SRAM is far more sophisticated and harder to do than dram in December 1984 Hyundai set out to produce a 16k SRAM but failed to achieve good yield due to design flaws they switched to Dynamic in 1985 but they were by then very far behind Samsung they tried to catch up with their own 64k and 256k dynamic memory products produced with designs transferred over from an American startup called vitalik but vitalik's design sucked and Hyundai's own branded products again failed to hit good yield FYI vitelica still around today has the Taiwanese analog chip company mosul vitellic ultimately Hyundai saved itself in the early days by being a Second Source OEM producer for companies like Texas Instruments and general instrument TI was hurt bad by the 1986 downturn and wanted to spread out some of their Manufacturing this memory Foundry model allowed Hyundai to learn how to improve their yields it kept the company alive the Hyundai semiconductor did not make a single Dollar in revenue for its first two years and as late as 1987 the company spent far more than it generated in Revenue gold star LG basically was a bit more cautious but they had an ace up their sleeve in 1982 LG acquired its dynamic memory technology from AMD and at T paying about 1.8 million dollars and 6.5 percent of their chip sales for the process two years later in 1984 they struck a deal with LSI Logic the Asic producer for gatorase though this fell through due to yield issues so they largely focus on things like transistor transistor logic TTL ROMs 4-bit microprocessors and other non-memory Specialty Products decent enough but the markets were small LG remained a distant third place the mid-1980s turned out to be a boon year for Korea with two favorable economic developments first the 1985 Plaza Accord this agreement revalued the Japanese yen in a way that greatly favored the Taiwanese and Korean currencies second Japan and the United States signed the 1986 semiconductor Trade Agreement here the Japanese government agreed to self-restrictions on memory exports to America for the next five years the intention was to give the American memory makers some breathing space but it also had the unintended side effect of opening the door for the Koreans in 1986 Samsung shocked the Japanese with their one megabit dram chip closing the Korean Japanese semiconductor Gap from five years to just one two years later in 1988 they brought out their four megabit chip so when the computer Market started growing again in 1987 and 1988 Samsung had Leading Edge product and they sold every memory chip they made turning very good profits it was the first validation of the prowess of the Korean semiconductor manufacturers thus far the big three chebel worked largely alone but in 1986 the Korean government tried to get them to work better together a collaborative project kind of like Japan's vlsi project in the 1970s the immediate goal was to produce a four megabit dram chip by 1989. the companies would get subsidies to share research on a new Innovative memory cell design known as a trench capacitor it would take the design into the third dimension allowing for much greater density the turbo were already skeptical about the project they figured the Americans and Japanese would already be done with 4 megabit by 1988 but they went along anyway for the subsidies the project had some pluses for instance Samsung probably would have taken longer to develop 4 megabit without it and Hyundai and LG benefited from seeing how the way Samsung did things kinda like how LeBron and Carmelo improved from being with Kobe during the 2008 Olympics and seen his work ethic the government arranged a 16 and 64 megabit project to follow up which the table also did participate in but ultimately The Big Three prefer to work on Next Generation products on their own and as those memory divisions became increasingly profitable they no longer needed the subsidies in 2017 Samsung's r d investment alone was equal to half that of the entire Korean governments Hyundai had slow played their semiconductor efforts in the mid-1980s watching Samsung to see how it would shake out for instance They delayed a technology transfer agreement with AMD for their 256k dynamic memory for this reason they had also been late to the market with their one megabit generation product and so they missed out on Samsung's big profits losses were piling up they had to get things right in the Next Generation 4 megabit product or it was over yet even as 1990 rolled around Hyundai had yet to produce a working die things seemed grim desperate Hyundai hired Dr minwee sick away from Intel to serve as general manager and take over 4 megabit development he kicked the team into shape bringing world-class work processes in order to cut cycle time and improve yield to decide whether or not to use trench capacitors for 4 megabit memory he set up two teams and had them compete with one another they eventually decided to move forward with trench capacitors Hyundai Electronics pulled off a stunning turnaround cutting development time for the 4 megabit from three years to just one their product was released in April 1991 just six months behind Samsung and it was so good that Fujitsu even licensed it Hyundai was back in business in 1989 LG sold about 52 million dollars in chips while Samsung did 600 million LG's distant third place was too much to bear LG wanted to get into the memory Market but how their products were at least two years behind Samsung and Hyundai enter Hitachi the ace up their sleeve LG and Hitachi have long had good relations they worked on VCRs together before now Hitachi needed additional trailing Edge capacity to compete with their Rivals Toshiba and NEC so in 1989 the two companies struck a rare technology transfer deal for a sophisticated Fab in the Korean city of chengju it would be a complete duplicate of the Hitachi plan it would cost 2.5

billion dollars by 1997. LG would produce one megabit dram and 256k SRAM Hitachi would purchase some of the supply LG would be free to sell the rest with Hitachi receiving a handsome 6.5 royalty of course the 1990s were the decade in which the South Koreans moved ahead of the Japanese in 1991 the Japanese real estate Bubble Burst and the country entered its long economic malaise one by one the Japanese memory makers would pull out of the dynamic memory business the year before in 1990 Samsung released the 16 megabit dram chip they're already working on 64 megabit and 256 megabit by now they were doing their own r d in their 26 research institutes no more need to borrow anything at the same time the one megabit and later 4 megabit markets boomed as Hardware makers upgraded to accommodate more Graphics intense Microsoft Windows enabled computers the Korean memory makers were the dominant beneficiaries of this trend Samsung simply overwhelmed the Japanese their dynamic memory business was now throwing off incredible profits and few companies can keep up with the spend there were like bullies at the poker table in 1993 Samsung was the first Memory Maker to upgrade their 16 megabit dynamic memory to 200 millimeter Wafers this made their semiconductor production far more efficient granting a significant cost advantage and 13 total market share Samsung Electronics thus sought to expand into other areas like logic ROM and LCDs the years 1995 and 1996 saw the Korean dynamic memory industry at a fever high they had 30 percent of the total memory Market but LG's entry into the market had added a lot of capacity from 1991 to 1995 Korean memory production had grown from 1.8 billion dollars to a staggering 15 billion that is 8.3 times growth

50 plus new lines started up between 1995 and 1996 alone not to mention all the new semiconductor startups in 1995 the mid-ranked chebo ilgen group entered semiconductor fabrication and in 1997 dongbu previously the Korean government could have been counted to step in but in 1993 Kim jong-san came into power and he followed an economic liberalization policy that saw the government lose many of its tools for doing so memory demand remained quite strong the secular Trends from the rise of the personal computer and other Technologies were still there but as we entered the 1996 year the market just had too much of a commoditized product for additional information about the memory over capacity situation I recommend this write-up from Doug of fabricated knowledge all this new capacity was funded by Korean debt the table have always heavily relied on debt as we entered 1997 all the chebble had debt to equity ratios exceeding five to one this left them dangerously vulnerable a danger in part obscured by opaque corporate structures Hyundai's official debt to equity ratio was reported at 341 percent but if you were to Value the assets at their actual market value the ratio shoots up to a stunning 1 378 percent the Asian financial crisis began in Thailand and then spread into Indonesia from Indonesia it reached Korea as many Korean Banks had lent to Indonesian Banks Hong Kong raised their interest rates putting pressure on the Korean won the bank of Korea eventually had to float the one and it crashed causing lenders to shut off their lines of credit Korean Banks found themselves unable to turn over their short-term debt for several weeks South Korea tried to pull itself out of the crisis situation these efforts failed and the IMF stepped in with the 57 billion dollar bailout things were just fine for Samsung electronics though net profits fell from 2.8 billion dollars in 1995 to just 87 million in 1997. the company simply changed its focus to producing more consumer items like TVs fridges and the like cheaper chips meant better margins for those end user items sales of these consumer products jumped 31 to 15 billion dollars more than making up for whatever was lost on the semiconductor side of the business furthermore the larger Samsung group generated its core revenues and profits from technology rather than debt laid in heavy industry so Samsung's top managers jettisoned many of their underperforming segments without drama mostly and emerge from the crisis a stronger and more streamlined company good for them but it was a different situation for LG and Hyundai first the two were semiconductor followers they didn't have Market leadership nor were they as strong and exported consumer electronics as Samsung was Hyundai had it the worst thanks to Falling memory prices Hyundai Electronics turned to 146 million dollar net loss in 1997 and then another 112 million dollars in 1998 debt skyrocketed ah yes the debt thing the larger Hyundai group had good businesses and Automobiles and ships but also suffered significant over capacity in those areas this suffering prevented the siblings from bailing out the electronic subsidiary very significantly Hyundai's leadership did not understand the political and governmental changes brought about by the 1997 crisis the table brought this about the government had to do something the bailout came as Korea elected a president in December 1997 Kim Dae Jong unlike his predecessors this president had few ties with Korea's big businesses heralding a new far more hostile business environment for the chibor heading into office Kim sought to reform the chebour add more transparency and restore the country's competitiveness he also had to abide by the debt and economic requirements set forth by the IMF even before his inauguration Kim brought together the heads of the five biggest treble then in June 1998 Kim unveiled a series of quote-unquote big deals to swap the group's business operations to trim their excess capacity there were multiple big deals involving various companies across Industries but the most prominent big deal would have LG's semiconductor business be given to Hyundai I guess the big deals make some sense except for the part that the chebro didn't want to do it for instance one of the deals was to have Hyundai take over the struggling Samsung motors with the idea of strengthening Hyundai as an automobile company but Hyundai did not see the point of taking Samsung's trash and Samsung did not want the loss of face with the LG Hyundai Electronics merger LG in particular did not want to see control of their semiconductors business which they saw as a crown jewel it might damage their close relationships with Hitachi and it would also mean Fab closures with thousands of workers losing their jobs public pledges were made yet when it came to actions nothing happened for months LG had apparently tried to pull out but the government pushed them back in again president Kim had personally told LG chairman kubonmu that the government would not allow the deal to falter threatening to have State Banks call in LG's loans finally in January 1999 LG announced that it would sell a controlling stake in LG semicon to Hyundai Electronics and then on April 1999 Hyundai bought 60 of LG semicon for 2.2 billion dollars created a new company called heinix semiconductors the third largest Memory Maker in the world heinix immediately became egg on the face for the Kim Administration president Kim received heavy criticism for arranging this shotgun marriage industry magazines fell Hyundai paid too much for a third place player with aging technology others thought that the deal was being forced through by the IMF after a difficult birth the newborn struggled to get to its feet heinix came into this world with a staggering 10 billion dollars of debt just in time for the dram cycle to take another terrible turn downwards if you don't recall the.com bubble burst in 2000 the two or three years thereafter were some of the worst ever for the memory Market memory chip prices plunged 80 percent the more money the company lost the less they were able to invest into their Fabs in 2002 Samsung had a capital expenditure budget eight times larger than hynix's but less investment meant worse products leading to more losses scientifically speaking heinix ate a turd sandwich all the way down the company posted a 1.9 billion dollar loss

in 2000 the year after that another three billion dollar loss without heinek's every company in Korea raised their 2001 profits by an average of 25 percent year over year heinex alone brought that average down to a negative 3.3 percent profit gain it is kind of like the opposite effect of Bernard Arno walking into a room and turning everyone in that room into a billionaire on average heinek's seemed locked in a death spiral with no way out and the Raging dumpster fire became an ongoing embarrassment for the Kim Administration in May 2001 heinek sold nearly a billion dollars in bonds to foreign investors and the Korean government backed those bonds through their State Banks a few months later in October 2001 another very complicated Financial restructuring deal to inject another 4.2 billion dollars in Aid through the use of debt swaps interest payment suspensions and Loan write-offs the Americans and Europeans were Furious over these later Micron and Infineon would file anti-subsidy protests with their relevant governments arguing that the Korean government was unfairly propping up heinix this brouhaha lasted for several years the company pulled out all the stops to get itself back on track heinix cut their Workforce from twenty two thousand to ten thousand a third of their Executives was fired heinek shut down their Fab in Eugene Oregon they gave all their staff a five-month unpaid holiday and they sold or spun off their non-core businesses and land their mobile phone unit curitel their Telecom system unit Hyundai networks Customer Service Unit Hyundai DigiTech their 47 stake in display maker image quest and their automotive electronics division Hyundai Autonet critically the sale of their valuable TFT LCD and OLED Technologies to China's Boe technology these sales funded with loans from hynex's own creditors gave the Chinese a future national champion hi Nick sought to sell itself to someone anyone really talks with the American Memory Maker Micron dragged on for months and Micron eventually did offer 2.8 billion dollars and mostly equity for hynix's Korean semiconductor operations combining the industry's number two and three largest memory makers feels a bit anti-competitive but creditors approved it simply because it brought money to the table however the employees protested and threatened to strike a big labor movement gathered in Incheon the center of hynix's operations and Korean nationalists for their part decried the loss of Korean technology at a fire sale price so in the end hynix's board voted no they probably felt that heinek's might be able to survive independently 128 megabit dram prices were on an uptrend climbing from a low of one dollar in November 2001 to four dollars by February 2002. but the refusal had consequences taken aback after so many months of negotiation Micron walked away and heinek's CEO resigned in protest High Nexus 100 plus creditors then converted their debt into equity and took over the company firing the entire board of directors corporate fantasy for every activist investor out there finally at the end of 2002 the Government tried one more time leading a massive 4.9 billion dollar restructuring that rolled over some debt and relieved others this combined with more asset sales new products and a continued Market recovery thanks to the smartphone boom finally worked after having lost another four billion dollars or so in 2002 and 2003 the company finally turned a substantial profit in 2004.

in 2012 high Knicks was finally sold off to SK group and renamed to SK hynix it remains Korea's second largest semiconductor company after Samsung in 2003 the EU decided that the Korean government did indeed subsidize heinek's through the use of loans this caused the EU to put duties on hynix's imported memories but in 2005 the World Trade Organization overturned the case the interplay between the Korean government and its semiconductor companies is a very interesting one Administration set the initial direction towards semiconductors though it seems like they might not have known its significance back then and they did disperse some money like that through the four megabit projects but after 1983 it was basically the treble in the driver's seat with the government along for the ride it was probably necessary such massive ketchup could not have happened otherwise but when things ran aground in 1997 the government struggled to handle the crisis and stabilize its massive treble in some ways since the dangers of this approach have only intensified all right everyone that's it for tonight thanks for watching subscribe to the Channel Sign up for the newsletter and I'll see you guys next time

2023-08-04 22:41

Show Video

Other news