Remember Tsinghua Unigroup, a wholly owned business unit of Tsinghua University and itself owner of Yangtze Memory Technologies Co. (YMTC) (Previously mentioned here.) Well, it turns out that a bunch of their top executives just got arrested:
The video shows a picture of six semiconductor executives, all of whom have reportedly been arrested:
Dia Shijing, co-president of Tsinghua Unigroup
Lu Jun, president of Huaxin Investment
Zhao Weiguo, chairman of Tsinghua Unigroup
Ding Wenwu, president of National IC Industry Investment Fund,
Zhang Yadong, president of Tsinghua Unigroup
Qi Lian, another co-president of Tsinghua Unigroup
How a company runs with three presidents I couldn’t tell you. Must be a Chinese thing.
“In the past few days, several senior executives of the organization behind the semiconductor industry in Mainland china have been taken away by the CCP Central Commission for Discipline, Inspection and Investigation.” Given my knowledge of communist nomenclature, I strongly suspect that this is not the sort of organization you want to enfold you in their tender mercies.
“In 2014, the General Office of China’s Ministry of Industry and Information Technology announced the official establishment of the National Integrated Circuit industry investment Fund Company Limited [ICF], also known as the National Big Fund or big fund.” Probably best to think of them like USA’s SMEATECH, but with a whole lot more opportunities for graft.
Together two rounds of government funding added up to 320RMB, or about $47.4 billion, which should have driven additional public/private capital investment of some $240 billion divided up between China’s Ministry of Finance and large central Chinese enterprises, most of which are also owned by the state. Even for the semiconductor industry, that’s a lot of cheddar.
By some estimates, $100 billion of that had already been spent by 2021.
“The two phases of investment cover all aspects of integrated circuits (ICs), including IC manufacturing IC design, packaging and testing semiconductor materials and equipment, and industry ecological construction.”
ICF provides overall direction and management, while Huaxin Investment provides management of the second phase of fund investment.
“Eight years have passed, but high-end Chinese chips haven’t yet been produced, and the management of the state level chip industry has collapsed.” Reading between the lines, this means TSMC is still kicking their ass. If that’s the standard, then it’s a bit unfair because every other semiconductor manufacturer in the world is in the same boat.
On July 28, Xiao Yaqing, head of MIIT, fell from power. “Xiao was the spearhead of the Chinese communist party’s attempt to build a world-class chip industry, and eliminate its dependence on the US.” He supposedly tried to slit his wrists.
“The very next day, Xi Jinping immediately appointed a replacement a longtime aerospace official to take over MIIT.” Yeah, that’s really going to help your semiconductor goals.
“On July 15, Lu Jun, former deputy director of the China Development Bank Development Fund Management Department, was investigated Lu Jun was involved in many investment operations of the Big Fund, of which he was the sole manager. He was also former president of Huaxin.
Yang Zhengfan, another Huaxin executive, was also taken away.
Also arrested: Wang Wenzhong of Hongtai Fund and Gao Songtao, both involved with Huaxin and the Big Fund. And that’s probably not all. Evidently a whole network of semiconductor executives are being rounded up.
Dia Shijing of Tsinghua Unigroup was among those reported arrested, but Tsinghua Unigroup is saying “Nah, everything’s good here! Go about your business, citizens!”
In July 2021, Tsinghua Unigroup announced that it was overwhelmed by 200 billion RMB of debt and filed for bankruptcy because it couldn’t pay its bonds at maturity. Keep in mind that Tsinghua Unigroup, partially owned by Tsinghua University, is itself owner of YMTC, which is (I think) China’s biggest domestic memory chip manufacturer. Tsinghua/YMTC was previously one of China’s biggest semiconductor manufacturing success stories, second only to SMIC, and supposedly “the largest integrated circuit company in China.” They have actual working fabs up and running. And they’re still evidently a money-losing failure.
Tsinghua Unigroup has grown through mergers and acquisitions, buying up over 20 companies. This strategy is not unknown among western companies, as GlobalFoundries and NXP are both the results of a similar strategy. But neither of those companies is on the cutting edge.
“Tsinghua Unigroup has been using short-term loans rolling over to create long-term loans. These made the group’s cumulative liabilities too large and its financing structure unbalanced.” Yeah, I bet. “Get big quick” worked for a few doctcom era mega-success stories, but I don’t think it works in semiconductors.
Zhao Weiguo once boasted he was going to buy TSMC. Also, I’m going to kick Shaq’s butt in the slam dunk contest just as soon as I take time off from dating all these supermodels.
China Development Bank extended Tsinghua Unigroup 100 RMB credit between 2016 and 2020. Still a lot of cheddar.
I’m skipping over a whole lot of blow-by-blow “who owns what” in the corporate structure. Imagine if Spectre, the Gotti Family, and the Bank of England all had shares in Amway.
“Due to debt, Tsinghua Unigroup abandoned its plan to build DRAM memory chip manufacturing plants in Chongqing and Chengdu in southwest China earlier this year.” I bet that left a lot of pissed-off local commissars holding the bag.
“When the chip industry becomes a national strategy, but with no real oversight, it becomes a disaster zone of corruption, and a big cake for those in the circle to get rich for themselves.” True of any industry anywhere, but especially true of China, and especially true of semiconductors, where “fake it until you make it” isn’t an option if you’re actually building fabs.
“China cannot make high-end chips to this day.” True.
“American chip technology is far ahead of the world.” Also true, though with caveats. For semiconductor manufacturing, TSMC is on the cutting edge, with Intel and Samsung within striking distance. For semiconductor leaders, two American companies (Applied Materials and Lam Research) dominate a fair number of technologies, but Tokyo Electron is competitive in many of them, and ASML dominates the stepper market.
Skipping over the bits where China stole US (and other) tech, which should be familiar by now.
Enter the Trump Administration, “blacklisting and embargoing more than 600 Chinese high-tech companies and high-end manufacturing companies, as well as universities and research institutions.” Pissing off your biggest trade partner is generally not a great plan.
Result: Bottlenecks in China’s supply chains.
EDA makes software to design chips, and China has no real substitute.
Semiconductors are still a big item in China most recent Five-Year Plan (and yes, the Chicoms still use Five Year Plans, just like Mama Stalin used to make).
“The outside world has not seen the investment of the Big Fund break any bottleneck. However, the earthquake happening in the industry has directly shown people that there is a deep corruption in the Chinese chip industry.” Why should it be different than any other Chinese industry?
And just who is going to step up to those jobs running China’s increasingly-unlikely-to-succeed semiconductor moonshot, given that the last batch got rounded up by the Chinese Inquisition?
Interesting bit of history: Previous CCP head Jiang Zemin put his own son Jiang Mianheng in charge of developing China’s semiconductor industry, and also managed to make the country even more corrupt than it already was. And here we are.
It’s ironic that just as Washington was passing a giant graft bucket of semiconductor subsidies because China was supposedly kicking our ass, China itself was sacking the very people presiding over China’s own bucket of graft for not catching up to the west. The truth is somewhere between.
China was never going to catch up to western semiconductors because the gap was too large and you need a crazy swarm of free market capitalist entrepreneurs risking private money to eek out important incremental process tweeks to keep Moore’s Law going. China was never going to have that as long as they suffered under Communist rule. And a huge percentage the government money that was sloshed into semiconductors was indeed swallowed up by graft and diversion of funds. But all that money does appear to have helped China close the gap some. Granted, a lot of that was via systematic IP property theft, but it got them into the game.
Twenty nine years ago today, on February 26, 1991, units of the American Second Armored Cavalry Regiment engaged the armor of the Iraqi Republican Guard Tawakalna Division in the Battle of 73 Easting.
The furious action lasted twenty-three minutes. The troop stopped when there was nothing left to shoot. Sporadic contact ranged from nuisance machine gun fire to one company-sized counterattack of T-72s and BMP armored personnel carriers. Tanks and Bradleys destroyed enemy vehicles at long range from the dominating position on the ridge. Three Bradleys from first platoon, led by Lieutenant Michael Petschek, encountered and destroyed four T-72s as they moved north to reestablish physical contact with G Troop. Medics treated and evacuated enemy wounded. Crews cross-leveled ammunition. Mortars suppressed enemy infantry further to the east as our fire support officer, Lieutenant Dan Davis, called in devastating artillery strikes on enemy logistical bases. Scouts and a team under the control of First Sergeant Bill Virrill cleared bunkers using grenades and satchel charges, and then led a much-needed resupply convoy through minefields to our rear. A psychological operations team broadcasted surrender appeals forward of the troop and the troop took the first of hundreds of prisoners including the brigade commander. Soldiers segregated, searched, and secured prisoners through the night. Many prisoners cried because they had not expected such humane treatment; their officers had told them that we would execute them. The prisoners were incredulous when our soldiers returned their wallets without taking any of the money that they had looted from Kuwait City. Just after 2200, 1ID conducted a forward passage of lines in Third Squadron’s area of operation to our south.
The morning after the battle, soldiers were exhausted. Many of the approximately fifty T-72s, twenty-five armored personnel carriers, forty trucks and numerous other vehicles that the troop destroyed were still smoldering. Our troop had taken no casualties.
Here’s a video on the battle:
In addition to being an overwhelming victory, and part of the larger overwhelming victory of Desert Storm, the Battle of 73 Easting was important for several other reasons.
For one thing, it was the largest tank battle between American- and Soviet-constructed armor since Israeli M-60 Patton tanks faced off against Egyptian T-62s in Sinai campaign of the Yom Kippur War in 1973. All throughout the 70s and early 1980s, various media outlets talked about how much better Soviet military equipment was than American equipment. (I remember a 60 Minutes episode that talked about Soviet equipment being better “all across the board.”) And Soviet equipment was better—on paper, with thicker armor, higher top speeds, etc. And then 73 Easting happened, and M1A1s wiped the floor with T-72s. A lot of that was American troops being much better trained and led than Iraqi troops. But the Republican Guard was the best the Iraq army had, and on paper the T-72 was a match for the M1A1s. In actual combat, the T-72s started blowing up before they realized the Americans were engaging (and destroying) Iraqi armor at the extreme range of the American computerized fire control systems. Soviet armor still used reticules, where the gunner had to manually calculate distance and windage to put shots on target.
In Vietnam, early computerized combat technology was clunky and unreliable. By the time of Desert Storm, the furious onrush of Moore’s Law had rendered technology smaller, more compact, more reliable, and more user-friendly. By pursuing what Jerry Pournelle called the strategy of technology, the United States was producing weapons that were qualitatively superior to those of its communist foes. That technological gap (especially in the form of SDI) was one of the drivers for the end of the Cold War, and it was on full display in Desert Storm. The Soviet Union itself would dissolve later the same year.
The Battle of 73 Easting has become the single most accurately recorded combat engagement in human history. Army historians and simulation modelers thoroughly interviewed the American participants, and paced the battlefield meter by meter. They came up with a fully interactive, network-capable digital replica of the events at 73 Easting, right down to the last TOW missile and .50-caliber pockmark. Military historians and armchair strategists can now fly over the virtual battlefield in the “stealth vehicle,” the so-called “SIMNET flying carpet,” viewing the 3-D virtual landscape from any angle during any moment of the battle. They can even change the parameters – give the Iraqis infrared targeting scopes, for instance, which they lacked at the time, and which made them sitting ducks for high-tech American M1s charging out of blowing sand. The whole triumphal blitzkrieg can be pondered over repeatedly (gloated over even), in perfect scratch-free digital fidelity. It’s the spirit of Southwest Asia in a digital nutshell. In terms of American military morale, it’s like a ’90s CD remix of some ’60s oldie, rescued from warping vinyl and remade closer to the heart’s desire.
Like Agincourt or Amiens, the Battle of 73 Easting heralded the arrival of a new type of technology to the battlefield, one that every army in the world henceforth need to take into account.
