Semiconductor MIC, Explained: China’s Semiconductor Crown Jewel
Race, founded by the intensely driven richard chang smic, ruthlessly hired from its taiwanese rivals out competed its chinese ones and quickly established itself as the top dog in china’s semiconductor space. In this video let’s, look at smic, china’s, fiercest semiconductor, foundry player and the industries crown jewel. But first i want to take the time to ask you to subscribe to the agentometry newsletter. Getting tired of me saying the same thing over and over again. Well, imagine me having to say it so many times over and over again, i want to call out a recent post. I just finished about lin biao, the ccp’s ill fated number two to help better show his character and the unusual nature of his illness. I added a collection of anecdotes to the story to help better illustrate his unfortunate plight. It really goes to show just how a troubled individual he was. You can find the link to the newsletter in the video description below or you can just go to. Asianometry.Com subscribe and i’ll try to make it worth your while you can expect a new newsletter every four days at 1 am taiwan time. Much thanks. Smic was founded in january 2000 by richard chang. Chang is an interesting fellow born in nanjing. His family fled to taiwan, along with a million other ethnic chinese migrants. The wai shun ren, as they were called, grew up in taiwan, but held deep roots to the chinese mainland. I did a video about them earlier many of the waist and run disoriented and feeling unwelcome in taiwan emigrated to other locales.
Like the united states. Chang was one such person after graduating from the prestigious national taiwan university. He got a master’s degree from the state university of new york in buffalo. Then he and his wife went to work for texas instruments in dallas tsmc, founder morse tang. No relation also worked at ti for over 25 years at ti richard chang oversaw foundry work in east asia for 20 years. He did well there and took early retirement in 1997.. Returning to taiwan, he started a foundry of his own worldwide semiconductor manufacturing corporation and eventually sold it to tsmc in 2000.. Richard left and soon then itched for another opportunity and china came calling with cheap land, water and electricity. He crossed the strait and started smic with 1.6 billion of investment from goldman sachs and a few other private equity firms. Chang and smic, soon found themselves in fierce competition with grace semiconductor grace was founded around the same time as smic by two political heavyweights. The first was winston wong son of powerful taiwanese industrialist wang yong ching huang yong qing had founded one of taiwan’s biggest chemical empires, formosa plastics corp, his daughter runs htc. The second heavyweight was zhang. Miyan, hang son of chinese paramount leader zhang zaman. Naturally, having such connections would put grace at the very forefront of the chinese semiconductor foundry industry smic knew that it would have to compete extremely hard in order to distinguish itself from the competition they did as it turns out being a rich guy’s son does not help Much in competitively running a semiconductor foundry for richard chang and his team, this was their second startup and they knew how to get things done.
In contrast, winston wong had little experience in building and running leading edge, independent foundries, not that he didn’t have any. I think he worked it before in in memory, but it wasn’t the same as logic grace could not keep up. In the same time, grace took to get its first foundry to mass production. Smic had two in shanghai a bunch more on the hopper and was already prepping for their new york stock exchange. Ipo grace settled into a number two position, and the chinese government soon shifted their favor to smic in terms of subsidies, cheap loans and more by 2004. Less than five years after its founding smic had ascended to become the fourth biggest player in the independent foundry industry, only tsmc umc and singapore national champion chartered semi were ahead of it and it was rapidly eating up market share two big things. You need when running a startup. Foundry is a fab to make the chips and a lot of talented people to run it otherwise, i’m, not sure you might have a feasible business. You might have a hong qing smic’s business model at the time is not all that different from an nfl or nba sports team. A foundry needs an expensive multi billion dollar fab to do their work. Likewise, a sports team needs quote unquote an expensive multi billion dollar arena. Neither their organizations can raise or risk the billions of dollars to self finance. One so get someone else to do it.
Unable to tap bank financing. Smic thus went to chinese municipalities around the country and got those local governments to finance a fab’s construction. Smic would then move into the fab on a contract basis to operate it. The local government would get some fee. I guess and the opportunity to build a quote. Semiconductor cluster kind of like how nfl teams like to say that, on game day, local businesses benefit from all the increased traffic around their billion dollar stadium. You should google those claims to see how it turned out for those cities as smic ascended up the chain and the leading edge process got more and more expensive. These capital expenditures became onerous even for the richest local governments. Smic began performing a type of regulatory arbitrage shopping, their services to cities around the nation. Looking for the best deal, for example, in 2005 smic was running two factories in shanghai, then the next year they engaged with the cities of chengdu and wuhan to start foundries. There too, two more joined ventures at the cost of 1.5 to 3 billion each it allowed smic to grow extremely fast, while keeping their actual invested capital low. In a previous video, i likened advanced semiconductor foundry work to baking and, like baking, accumulated unwritten experience is critical for the life of me. I can’t make a decent cake, even if the steps in the recipe are right in front of me. This is a broad generalization, but foundry work is in many aspects, an art that cannot be easily boiled down into steps.
Not everything is or can be, written down so it’s critical to have with you the best chefs, both grace and smic aggressively recruited overseas town to fill their ranks, including taiwanese from tsmc and umc being a startup. These companies lured such employees with the promise of stock option riches at the time salaries on the chinese mainland were far below that which can be found in the u.s or taiwan. A senior manager in china, for instance, would get paid a quarter or a third of what that same person could get in the united states. Grace handed out stock options, but smic gave them out like candy for certain key tsmc employees, smic offered 80 000 shares and stock option equivalents. Such amount of shares at the 1750 us ipo price in 2004 would be worth 1.4 million dollars, not to say that those people actually got their hands on all of that 1.4 million. I have worked at startups myself and it is very common for that. 80. 000 number to quickly get diluted, as the company goes to funding rounds and the like, beyond the potential riches people gravitated to smic for two other reasons. First, many of the engineers were ethnic, han chinese living and educated overseas and wanted to give back to their country of smics founding 1000 engineers. Nearly 40 percent of them were chinese citizens returning to the motherland. Second, many of them felt that they had hit the glass ceiling in an american or taiwanese company.
Big companies have often established ranks of who gets promoted, and there is often more qualified people than open spots. We should also mention the possibility of racism against asians and chinese in the united states. The bamboo ceiling to join smic would be to step into a silicon valley like environment and get in on the ground floor of something exciting to build something for the future and glory of china hard to say no to that by 2003. Tsmc realized that grace, and especially smic, were growing faster than it should have. For example, smic had managed to ramp up on a 0.18 micrometer process in just 12 months. This is without a prior track record of success. In doing such things before, this was implausible and implied the loss of private, extremely valuable trade secrets. Thus, in that same year, as smic prepared to go ipo on the nyse tsmc launched a pioneering intellectual property theft, lawsuit against his mainland rival. This lawsuit was filed in california and, interestingly enough, tsmc asked for a jury trial. It would not be resolved for many years, but when it did, the jury found that smic had indeed infringed on tsmc intellectual property. The lawsuit went back and forth for several years, but when the dust finally settled six years later, tsmc won, they received the 10 stake in smic. Richard chang resigned in 2009. After attempting a counter suit against tsmc that failed. At the same time, tsmc took to battling smic directly on its home turf in 2004.
Tsmc won a public debate in taiwan for permission to enter the mainland market in 2007, tsmc’s first fully owned foundry began operations in shanghai. This appears to be a page out of their playbook. In my video about singapore’s chartered semiconductors, i mentioned tsmc founding a semiconductor foundry joint venture with the goal of battling chartered on their home market. It helped to weaken the overall market enough to decelerate smic’s once torrid growth. Throughout this entire saga, smic lost money. The company had been running on subsidies, loans and the money it received from investors, but the company could not consistently turn a profit on its own. There are a few reasons for this. First, there were the upstarts smic was the largest chinese foundry, but it was not alone. Other chinese cities saw what was happening in wuhan beijing and chengdu and wanted in on the action they began, creating small smics of their own saturating, the market and bringing down profits for everyone else. One upstart story takes place in the city of ningbo, the place of tsmc founder morris chang’s birth and also my ancestral homeland ningbo zongwei was founded in 2002 by a team of mid level. Tsmc managers they bought used equipment from tsmc and invested 150 million of taxpayer money to build a leading fab capable of outputting 40 000 wafers a month. The fab reached the 10 000 wafers a month benchmark behind schedule. The company could not reach the necessary scale, as tsmc’s equipment was quite old and required, expensive servicing 100 to 150 dollars per hour, maintenance, eventually the company folded and it sold their factory to electric car maker byd.
Another reason for the struggling losses has to do with the product being sold in order to make sure that some of its early fabs got online as fast as possible. Smic devoted them to creating commodity ram. The fabs indeed got up and running really quickly, but making a commodity product meant low margins and intense competition it’s the same issue that charter dealt with. You need to extend the value chain and get to the leading edge. That means a lot of investment into r d, so too many startups too much supply sloshing around, not enough differentiated product and, of course, some of the most economically challenging years with the global financial crisis. Where can you get the money to invest in the r d? For better product, if you are not making money from your products, is the chinese government just going to have to constantly pump billions of dollars into your company? I guess they can technically. But how long can you put money into a company that doesn’t turn a profit? Singapore is quite rich: singapore, owned charter for 22 years and for the majority of those years chartered churned out losses, despite becoming the third largest independent foundry in the industry. Just because you have a lot of money, doesn’t mean you like rolling it up in a cigar and smoking. It that’s what’s happening with smic and chartered so long as it did not own and create the technology behind the process themselves.
They were trapped over time. Local city governments like wuhan and chengdu, began to feel uncomfortable about the seemingly unending operating losses. After richard’s resignation, chengdu and wuhan terminated the partnerships, chengdu sold its fab to texas instruments and wuhan decided to go it alone with a company called xinchin semiconductor smic, hired a new ceo in 2010 and sought to dial down on growth in order to achieve profitability. If richard chang was uber’s, travis kalaknik the vicious driven founder, determined to take over the world, then his successor, david wong, was more like dara koshrashahi, never gon na get that right, trying to turn the company around and make a profit david wong did not last long. He resigned after losing a vote of confidence. A year later, tu yincho led the company has ceo until 2017, whereupon he helped stabilize the ship and turn out a very small profit, he retired in 2017, and was succeeded by co. Ceos dao, haijun and niang mong song diang is famous in taiwan for jumping ship from tsmc to samsung and helping them reach the leading edge process. Note it seemed that he had been doing the same with smic too leading a team of 2000 r d engineers to ready the seven nanometer process node. He realized just how important it is to research, create and thus own the process, technology itself, not to take shortcuts and partner with others. In sharing r d costs per the recent reports, it seemed like that they were getting close to high volume production with 7 nanometer.
This is, of course, great news for the chinese semiconductor industry and represents a return to form for smic. After many long stumbles, smic remains the premier foundry on the china mainland, their recent billion dollar ipo on the hong kong and shanghai markets. After a 2019 nysc d listing makes it clear, the chinese national government is starting to invest real financial resources in actually developing an indigenous semiconductor space from the ground up smic will likely be its centerpiece for making it happen.