Arthur T Knackerbracket has processed the following story:
Japanese Prime Minister Shigeru Ishiba revealed the $65 billion plan this week. Reuters saw an early draft of the proposal, which is scheduled to be submitted during the country's next parliament session, and notes its support of domestic chipmaker Rapidus among others.
The homegrown semiconductor manufacturer was founded in 2022 with support from several major Japanese tech companies including Sony and Kioxia, and entered into a strategic partnership with IBM in December 2022. The outfit expects to start mass production of advanced chips built on a 2nm process by 2027. It is an ambitious goal, but one that could be helped along with a significant infusion of cash.
[...] As Tom's Hardware highlights, it took many years for established players like TSMC to get to where they are today. The publication questions whether or not Japan has enough workers with the skills necessary to achieve their goals. As we have seen both domestically and abroad, finding workers with the smarts to get the job done can be a real challenge. Even China, with its heavy investments and accusations of IP theft, hasn't been able to compete toe to toe with leading chipmakers.
It remains to be seen whether or not Japan's investment will pay off, but it is hard not to think that increased competition will benefit the masses via lower prices and a more robust supply chain.
(Score: 2, Offtopic) by Username on Monday November 18, @11:07AM (2 children)
Almost all major camera optics or sensors come from Japan. Sony, Nikon, Sigma, Olympus, Pentax, Fujitsu, etc, all Japanese. I think Canon and Samsung are the only ones to have their own fabs, but still get other electronic parts from the former. Leica and Zeiss are probably the only other well known lens manufactures.
(Score: 2) by Unixnut on Monday November 18, @11:31AM (1 child)
Considering how much Japan dominated the electronics industry in the 80s and 90s (especially with semiconductor electronics) I am quite frankly surprised to hear they did not have their own domestic chip production until 2022.
I always thought their electronics industry used primarily home grown silicon, at least until Taiwan and China took over the high and low end respectively.
(Score: 4, Interesting) by VLM on Monday November 18, @02:43PM
Coincidentally in that dominant era they had about 50% of chip production worldwide.
They had special government financing and permission to dump chips to take over markets. Other governments were pissed off but not much could be done.
They got "trapped" in the conversion in the 90s away from vertical integration. Before 1990 or so, pretty much every company designed AND made their own chips and the industry split around then. The companies that split, survived. The companies that did not split (Pretty much Japan and Intel) either did not survive or had some kind of special monopoly thing like Intel. General economic decline in Japan also dried up the government funding that had supercharged industry. The old days where the designers and the fabs worked at the same corporation are pretty much over now.
Japan kind of stagnated for 30 years. Not just in chips, either.
(Score: 1, Flamebait) by pkrasimirov on Monday November 18, @01:40PM (2 children)
Japan is too close to China for comfort. If Pooh goes with his plan to conquer Taiwan in 2027 then Japan will be in the hot waters too. I wonder what's the actual USA/EU plan, because so far it looks like nothing.
(Score: 1) by khallow on Monday November 18, @02:41PM
The default plan of inaction is that Japan remilitarizes, complete with nuclear weapons. South Korea will be in the real hot waters, sandwiched between China and Japan.
(Score: 3, Interesting) by ikanreed on Monday November 18, @10:16PM
The 2027 thing is essentially entirely the Pentagon's invention, because they believe China's economy will collapse and they will no longer out-produce the US in heavy ship manufacturing anymore around 2028. So they think China has to exploit their short term competitive to seize Taiwan before the US can catch back up in naval capacity.
Essentially no one in China's government thinks that 2027 is some kind of goal for conquering Taiwan. And by and large Chinese doctrine surrounding Taiwan is that if they go too far in being US friendly and claiming independence, a cuba-style blockade is all they'll need to "resolve" the situation.
Both sides are a bit of wishcasting(Taiwan will absolutely not politically buckle under an embargo, and the US has no actual strategy to improve our shipbuilding capacity), but I think your version's understanding is pretty dumb.
(Score: 2) by VLM on Monday November 18, @02:49PM
It always comes down to too little pay and too much interference from HR acting like Soviet political officers.
Even now, for example, 50% to 75% of EE grads go into programming because there's no EE jobs or they can make more money doing CS despite having an EE degree instead of a CS degree.
Its almost as bad as K-12 education degrees where most of the grads get jobs as waitresses and baristas after graduation and only a minority get a job in the field.
Offer an EE perhaps 80% to 90% of what a CS grad gets to do EE work and you'll have more applicants than you know what to do with, because that would be about a 100% pay raise...
Similar issues in technician field. You can make $18/hr as a electrical tech doing automation wiring but if everything else pays more and the supply is enormously larger than the demand ...
(Score: 1) by khallow on Monday November 18, @02:51PM (1 child)
(Score: 2) by VLM on Monday November 18, @03:48PM
MITI tried to run it like a NASA research project with predictable results. The money all got burned and some cool research papers got written and that's about it.
I'm surprised that in an industry where "everything old is new again" and nothing ever really dies in FOSS, nobody is pushing "Concurrent Prolog on a JVM" or similar retro-new-wave stuff. AFAIK this is all that's "active" out there today:
https://www.ueda.info.waseda.ac.jp/software.html [waseda.ac.jp]
Note that Japan's chip success was more like their SEC said here's some seed money, a license to print more money, and a promise you don't have to follow any international trade regulations, then got out of the way, resulting in massive growth. This is quite a contrast to how MITI ran their project.
They could go old school and hand out bags of money and promises not to enforce regulations (environmental, etc) and it would work pretty well.
(Score: 2) by quietus on Monday November 18, @04:17PM
The quoted articles fail to mention that ASML and IMEC are also involved: IBM is there [rapidus.inc] because it managed to demonstrate a 2nm chip in 2021 (Rapidus engineers working alongside with IBM engineers in Albany, New York), ASML has opened an office across the street with 40-50 staff and works together with IMEC on developing production technology beyond 2nm, which is also the stated aim of another cooperator, LSTC (Leading-Edge Semiconductor Center [www.lstc.jp]).
Interestingly, the Japanese Ministry's project proposal [meti.go.jp] also mentions photonics -- which was mentioned in another article on this site, this week: a project funded in part by a NATO investment fund (in this context, note the use of the word allies in the Japanese project). So, the way forward looks to be photonics in combination with next-gen CFET transistors [tomshardware.com].