Pentagon, With an Eye on China, Pushes for Help From American Tech

For the latest technology other than TSMC, Samsung is also a viable alternative. But for components of US origin, I think Intel is the only choice. Intel could have contracted out its production like TSMC when it had the chance. Instead it wasted billions chasing after ARM and Qualcomm and has nothing to show for, meanwhile, it lost its leadership in semiconductor processing. Lucky for Intel, it still the leader in supplying chips to data centers. Using cash flow from its core business, it can afford the $20bn cost of new fabs.

Chinese savings finance $4 billion fabs and China's market, growing 6% annually, makes it the no-brainer location. Staffing them with smart, tech-literate workers is easier, too: Chinese high schoolers graduate three years ahead of ours in STEM subjects. Taken together, those advantages will be difficult to surmount, even for the Pentagon, and suggest that we have work to do if we want to remain competitive.

TSMC is the reason why Huawei couldn't be pushed around, because we can't push TSMC around.

We need to have a good plan to make sure the US has access to the chips necessary to support our military during war. This may mean subsidizing plants. like TSMC's, in the US (despite the unattractive economics of subsidies).

This is a very tricky issue. On one hand, an extended semiconductor supply chain increases security risks. On the other hand, the companies ask for too much. TSMC and other Taiwanese chip makers historically have often made more money after taxes, even in profitable years, because of accumulated R&D credits. U.S. school districts often get shorted because of tax holidays to tech companies. You're essentially talking about giving massive subsidies to very profitable companies. At some point it does seem like extortion. Maybe the solution is to make the purchasing rules more strict.

The problem of US dependence on foreign manufactures turned first up on displays on which I worked in the US Flat panel displays were invented in the US. What I learned working on military displays The "kill ratio" of fighter plane pilots depends critically on the quality of the displays that feed information to the pilot. As the fighter turns the display alternatively goes from deep shadow to bright sunlight - and instantaneously has to adjust its brightness to feed the pilot crucial information. You may be annoyed that you can not read the screen of an Iphone in very bright sunlight. At 20 000 ft, not able to see your fighter display might kill you. For about a decade the Pentagon kept two display manufacturers in the US alive. It then gave up. There is no US manufactures of displays left

I have co-founded and led several "fabless" semiconductor startups, including the one that enabled a camera in every pocket. TSMC is a great company, but I wish TSMC had a state of the art trusted fab in the US. Perhaps Apple and other cash-rich high tech giants can partner with TSMC and the US government to make this happen. Increased support for training US students in micro and nano electronics is also needed. As the fabs left the US,so shrunk university activity in this area in our country as well as ready access to state of the art technology. There are also other elements of the ecosystem that need strengthening. We can rebuild it if we have the will. Investment really isn't that expensive in the big picture and will probably pay for itself many times over. How much has been invested (using that term loosely) in US farmers in the past 18 months?

The cheapest production is single source/just in time. This can also be one of the most fragile, sensitive to disruption. In the Old Days (say, 20 years and more ago) multiple sourcing was fairly common with geographically diverse suppliers and buffering of parts. Now, this is pretty rare due to cost pressures. Between Taiwan (TSMC) next to China and South Korea with an unstable North Korea, and years long lead times for fab tooling never mind skilled expertise needed, there is a very real need for military and critical civilian supply chain management. Some invention needed (the dimmer people will think just money is sufficient. Good luck with that, did not work for the Spanish or Romans, won't work for Americans either).

I and many others in engineering saw this situation coming 25 years ago and tried to warn about it, but our political ‘leaders’ refused to listen, due to some combination of ignorance about technology and manufacturing, and greed for donations from investors looking for a quick profit from offshoring. The US government must require that all electronic circuitry, including integrated circuits (‘chips’) used in any security related application, especially in defense and vital infrastructure, be manufactured in secure facilities in the US, and should finance as necessary the construction of the required production facilities and their necessary supply chains. The US defense budget is over $700 billion annually, a small fraction of which, say 5%, spent as direct investment or set aside as underwriting, would be money well spent to ensure that our military hardware won’t be rendered useless by a lack of secure electronics. Also, the requirement for US manufacture would ensure a market for US plants large enough to support a self- sustaining supply chain. The US government must act quickly to fix this problem or it could become unsolvable sooner than anyone expects. The risk of inaction is catastrophic.

Intel used to be the number one chip company. Then their first Republican CEO Paul Otellini, a Bush supporter, took over. He brilliantly told Steve Jobs that Intel was not interested making the microprocessor for the iPhone. Intel then went from being the number one maker of microprocessors to being number two. He was replaced by their second Republican CEO Brian Krzanich, a notorious Trump supporter. Under his leadership Intel went from being the number one semiconductor company to being number two. Apparently Republicans turn everything into number two.

I don't see the problem here, as the US is obligated to defend Taiwan against all efforts to undermine its sovereign nation status. If we were so concerned about the supply of these advanced silicon, we would be increasing our military clout in the Asia Pacific and doubling down on China tariffs. It's the only way forward.

The US still leads in many fields but it needs to realize and make peace with the fact the center of innovation and research are moving to East Asia. You can throw money at the issue but at some point you are going to run out of talented researchers and engineers. For decades the US can count on apolitical talents from East Asia and India but now a lot of them are staying home because of the opportunities there. The problem is structural. The US wants talents but is increasingly hostile to East Asians seeing them as competitors or bring less diversity.

The more you outsource , the more you become dependent on other Nations. At some point, all countries put national security over economic gain by importing from low wage countries That happened one hundred years ago in agriculture. There is no free trade, for reason that food supply is crucial Manufacturing is next.

The Republicans believe that any kind of industrial policy is a form of socialism. China isn't restricted by such a ideology. Republicans have no problem with farm, oil and other subsidies, but the recipients tend to be in red states, so that isn't seen as 'socialism'. This rigid ideology, coupled with a hatred of scientific expertise has cut into our technological edge. It has become a national security issue.

This is no surprise. I worked in the semiconductor manufacturing industry for decades, and the outsourcing of technology has been draining U.S. capacity for over 25 years. U.S. companies don't want to invest in advanced technology as the equipment has become prohibitively expensive, and unlike Taiwan, China and Korea, the U.S. doesn't subsidize the major U.S. manufacturers. This article doesn't mention that the U.S. no longer has significant silicon substrate manufacturing, and virtually all chip packaging is overseas. This is a big problem that anyone who worked in the industry could see coming, but the focus on short term profits at expense of longterm viability was more important to shareholders, and hence executives.

Given the money spent on weapons development, particularly aircraft, $15 billion seems like a drop in the bucket.