What if Taiwan Gets Taken Over

With the epic disaster our "most popular president ever" unleashed in Afghanistan in the last week, I feel certain that our rivals in the world (enemies if you prefer) are licking their chops and thinking this is an ideal time to do things they've never thought they could get away with.  After watching how the US handled China's institution of harsh shutdowns in Hong Kong, I'm sure that the leadership in Taiwan is planning how they could respond to a Chinese invasion since they can't depend on the US.

Ask yourself if you can envision that our current crop of "perfumed princes of the pentagon" would go to war with China.  About all I can envision is another "strongly worded press statement."  

Peter Grant at Bayou Renaissance Man made reference to this in his morning post today:

If I were a resident of Taiwan right now, I'd be packing my bags and looking for an exit as fast as I could.  (If I were a US manufacturer reliant on computer chips from Taiwan, which makes about half the world's supply, I'd be begging, borrowing and stealing any and all supplies I could get out of there before the invasion.).

Emphasis on the part in parentheses.  The heart of the issue is the Taiwan Semiconductor Manufacturing Company or TSMC.  I think everyone is aware that there's a chip shortage that's affecting the delivery of many products, from pickup trucks to phones.  While there's still a fairly large group of the semiconductor fabrication plants ("fabs") in the US, they tend to be older plants, not capable of the leading edge of performance found in the most modern processors (Central Processing Units or CPUs).  TSMC is the monster in the field. 

Around thirty five years ago, when I was working in defense electronics, many of the chips were made in Asia for American companies.  In the years since then, the situation expanded and concentrated more fabs in Asia with fewer in the US.  This was quite deliberate on the American companies' part.  Coincidentally, Electronic Design did a feature article about the shortage of parts and the need to build more fabs in the states that showed up in today's email.  

How We Got Here

In short, the relentless outsourcing and lack of investment in our own semiconductor sourcing moved the U.S. into last place at the most advanced nodes of this ever-growing industry. At present, American companies are at zero production capacity on a technology that’s coming more and more swiftly to define the world, both inside and outside of tech development.

The governments of Taiwan, Korea, and China, on the other hand, enacted swift, articulated industrial policies and incentives to semiconductor foundries. They traded short-term profit for the potential of long-term dominance. They did this at a time when capital investment subsidies were manageable and translated into ongoing competitive advantage.

The U.S. did not do the same. To the contrary, American companies took advantage of outsourcing deals that maximized near-term profit at the expense of long-term strategy.

It's important to note that first sentence; our problem is in the "most advanced nodes of this ever-growing industry."  It's not like we have no fabs in the US, it's just that the most advanced processors can't be made here without substantial investment.  The article notes that of the three types of semiconductors—digital, analog, and power—the U.S. holds only a slim lead on analog and faces a tense race with Germany for dominance in power microchips.  FWIW, the difference between analog and power chips is the power chips are bigger.  It's in the digital chips where we've fallen far behind.  

We did a useful overview of the industry with emphasis on fab issues back in August of '16.  While the "glamor" side of the semiconductor industry is smallest transistors, 14 nanometers at that time, the actual work is done by older, less fine geometries.  From that 2016 post:

The graph on the left is showing that 43% of worldwide semiconductor production is in the five largest geometries: from 65nm up to the largest sizes used.  Further, the graph on the right shows that 85% of new designs are 65nm and larger.  Or-Bach takes this as evidence that the industry is bifurcating; very few designs move to the finest geometries while most designs are being designed into older, bigger geometries.  This is leading foundries to invest and develop enhancements to the older processes, keeping older facilities in production longer.

The other aspect that goes along with most parts being larger is that they're cheaper parts. Investing in the new machines, new processes and learning how to make those components winnows out the field. This graphic shows that while there were 22 companies capable of building parts with 130 nm features, as features got smaller, the number of companies that could or would invest in the plant went down, so that at the finest features (densest number of transistors in the highest performance processors) only four companies were still in the game.  I'll bet one is TMSC and all four of them are somewhere in Asia.

I don't know if China has one of those plants, but I find it easy to think part of their motivation for going after Taiwan might be for the income TMSC brings in.  If the Chinese Communist Party owns the production of all the densest, most powerful processors, do you think they'll sell them to American companies?  Do you think the supply issues that Ford and other carmakers have are going to get better or worse?  

With one move into Taiwan, they could cripple us by cutting off processor shipments from TMSC, or raise the prices prohibitively.  It's looking more like a low risk move for the CCP.