Why Japan's $67 Billion Bet on Chips Matters | Momentum
Take a look around you. Can you imagine what our world would look like if there were no mobile phones, computers, TVs, cars, trains, or even the internet? These modern comforts, which we take for granted, are all powered by semiconductors. And most of these small but mighty chips are manufactured right here in Asia. While Taiwan leads the world in the production of advanced semiconductors, Japan is gaining momentum, outspending the United States and Germany on chip subsidies, in terms of GDP. Japan is going all-in on semiconductors, making a $67 billion bet to once again become a global chip powerhouse.
But is it all too little, too late? Can Japan buy its way back into the semiconductor race? I’m Haslinda Amin, and this is Momentum. In the 80s and early 90s, Japan ruled the chip world. In 1988, its companies controlled 50% of global chip sales. Then, the dot-com bubble burst.
Japan’s economy began a lengthy period of economic stagnation. And its chip technology began to slip behind its rivals in Taiwan and the US. Japan wants to revive its semiconductor glory days of the 1980s and it’s mobilizing everyone. The government, the private sector, academia, even international partners like TSMC and IBM. That means billions in investments, and billions more in subsidies.
And the stakes are high. Advanced chips will serve as the foundation for a dozen critical technologies, including artificial intelligence, weapons systems, and electric vehicles. Devices like smartphones and gaming consoles are all powered by semiconductors - tiny silicon chips smaller than a fingernail, which often take weeks, even months, to be fabricated. The semiconductor market is projected to be worth $1 trillion by 2030. Currently, a large portion of global production is based in Taiwan and South Korea, making the supply chain vulnerable to geopolitical tensions.
China could be prepared to invade Taiwan by 2027, so clearly there is urgency behind this. Mr. Kanazashi, Japan dominated the chip industry back in the 1980s, 1990s. Then it was overtaken by rivals like South Korea, like Taiwan.
And now, it’s kind of back with a vengeance, betting big on semiconductors. Why is that? During the, you know, time of Covid, and also concerning the situation of the Russian invasion into Ukraine, disruption in the semiconductor supply chain happened, which caused a complicated situation to the public life in Japan. So there are many end products, like cars or other electric devices like microwaves or etc. didn’t come to the people’s lives, so that’s why many public people realized the importance of the semiconductor chip. That was kind of a trigger for us to push the semiconductor industry. Recent disruptions in chip supply rattled the Japanese economy.
Its key manufacturing industries, like cars, game consoles and robots, rely on chips. Without them, production stops. So would you say that chip security is crucial for Japan’s competitiveness, Japan’s growth, Japan’s economy, going forward? Toyota, Honda, Nissan, they are very important industries for the Japanese economy.
But we have to have kind of multiple heads to push economic growth for the future again. Considering the future course of the Japanese car industry, they have to meet the challenge for autonomous driving, for example. In order to realize autonomous driving, they have to have the most advanced semiconductor technology. In order to realize electrification, they have to have accessibility to power chip technology.
So in that sense, to have semiconductor industrial basis, means we can have kind of weaponized semiconductors, for such Japanese user industries to grow more. Japan is not just playing catch-up. It’s also gunning to manufacture advanced semiconductor chips with 2-nanometer process technology, which will be used in products like AI-enabled smartphones, computers, and self-driving cars. This is a massive leap. So massive many experts think Japan will fail. The first pillar of Japan’s strategy was to attract the world leader in semiconductors, TSMC, to their shores.
The joint venture is scheduled to begin mass production later this year, surprising market analysts. The second pillar, and Japan’s big hope, is local company Rapidus. We do have some breaking news when it comes to Rapidus, and what Japan’s Ministry of Economy, Trade and Industry are doing to support the company. The two-year-old company aims to leapfrog generations of innovation and produce bleeding edge 2nm chips in 2027. But you’re also betting big on Rapidus, which is, some people say, a start-up.
Japanese car industries must have accessibility to the most advanced semiconductor technology, 2nm, 1.4nm to realize their most advanced business models, autonomous driving, electrification, etc. So that’s why we decided to launch Project Rapidus. Without advanced semiconductor technologies in Japan, there is no future for user industries, Japanese car industries, Japanese robotics industries. There is no future for them. So that’s why, Rapidus project, we think it is very critical to the future of the Japanese industry.
The drive to produce cutting edge chips is pulling in companies from across Japan. Success is a matter of national security. One company that is critical to its success is Preferred Networks, a Japanese tech start-up which is already producing its own chips, optimized for AI.
Some say that Preferred Networks could be Japan’s answer to Nvidia. What do you think? At this point in time, Nvidia has a very large market and is dominating that market, so we are still a small company compared to them, but our goal is to have a firm grip on the next generation of semiconductors. Founded in 2014, this tech unicorn creates hardware for AI models used in robotics, drug discovery, autonomous driving and e-commerce. In 2024, it secured $463 million of private investment to develop their own semiconductors. The reason why we decided to create such a processor is that AI consumes a tremendous amount of energy.
Sustainability will be gone. The energy AI requires will grow exponentially. We believe that suppressing this is essential for the development of AI, so we are developing a unique processor.
If we can create chips and computers with lower power consumption, many people will be able to take advantage of this new AI. More people will have access to this technology. This will make it possible to create jobs that are convenient for many people and increase the number of opportunities for AI to be used. Preferred Networks is confident of bringing its AI chip design and supercomputer to market in a few years’ time. But for the Japanese tech unicorn to go further, it needs an ecosystem of partners, like advanced semiconductor manufacturers in Japan. Mr. Nishikawa, how do you think Japan’s semiconductor industry
would look like in five years? In terms of our company, the foundation of our company, the emergence of generative AI and the current evolution of semiconductor technology have come at the same time, which is a rare opportunity for us. We would like to win this battle by creating something competitive, and we would like to be the dominant player. I would say the next 2 to 3 years is going to be the critical milestone because Rapidus will start to operate by then. We need to have the whole ecosystem ready and all the players ready to work as a whole so that we can commercialize a new semiconductor that is actually useful and maybe it will take a couple more years to fine-tune and integrate the latest technology to chase the rivals. But I would say that the next 2 to 3 years will be critical and key to the Japanese semiconductor industry because unless we are able to achieve some commercial grade product by then, I don't think we will be truly competitive.
This semiconductor renaissance is happening far away from Japan’s economic center - Tokyo. Instead, it is being built in places like Hokkaido, and Kyoto. Areas more famous for their scenic beauty and ancient history. Beyond the temples and palaces of Kyoto, are the foundries of semiconductor companies, like Screen Holdings. Screen Holdings designs and builds machines that clean chips, a process critical to the manufacturing of semiconductors. I wonder what the private sector thinks of the Japanese government’s ambitious plan to revive the semiconductor industry.
How much has the ecosystem improved, given the funding from the government? With this funding, we are thinking of taking on a number of new initiatives. There are manufacturers of manufacturing equipment in Japan, as well as many manufacturers of materials. We have heard that the 2nm challenge will be followed by a new packaging challenge, and I have also heard that various changes will be made to circuit design. So we see this is a huge opportunity for our business going forward. Still, I have my doubts.
After all, Japanese semiconductor manufacturers have not been chasing the leading edge for a long time. Overseas competition. Technological changes. Not enough skilled workers. These are just some of the challenges Japanese semiconductor businesses face.
We know that most of your clients come from China, the demand comes from China, 40, 50 percent in fact. And the demand is from the auto industry in particular. With Japan looking at a two-nano chip, really high-tech, how are you positioning to cater for that? We are supplying legacy node cleaning technology to our customers in China, and also cutting-edge cleaning technology to our cutting edge customers. I hope you will understand that we are providing different equipment for different customers. So you’re prepared and in a position to cater to the demands in Japan, when it goes to two-nano. Yes, right.
And in terms of demand for the future, where do you see it coming from? This is a difficult question to answer. What we are aware of is that in the future, AI chips will be produced in a way that each hyperscaler appropriately manufactures a different type of chip. Most of the chips we use now are supplied by a big company, Nvidia, but I think we will see a shift to suppliers making less volume. In that sense, Rapidus sees the supply chain as being slightly different and we expect that demand for small-quantity, high-variety, cutting-edge chips will emerge. The US announced billions of dollars in subsidies to grow its semiconductor market.
China is doing the same. Can Japan, with its much smaller economy, really hope to compete? With everyone getting mobilized to fulfill this semiconductor dream, can Japan be one of the top players in semiconductor? Don’t ask me, it’s very difficult to say. I hope so. But it’s very difficult, I think. We are challenging the target. In Taiwan, TSMC already has the most advanced technology, and in the US, IBM has the most advanced technology, so I think Rapidus is now in a position to take on challenges.
And we will do our utmost to make it a success. As for whether it will work or not, I would like to answer by saying that we will make the utmost effort. Transformation in the way that Japan’s private and public sector think, talk and dream about its chip ambition is the name of the game, and it’s nowhere more evident than in the formerly quiet town of Kikuyo, in Kyushu island. JASM, a joint venture between Taiwan chip giant TSMC, Sony Semiconductor Solutions and Denso, completed their first fab this year and has another on the way. It’s aiming to produce advanced chips for smartphones and AI applications.
What role did Sony play in getting TSMC to Japan? Ever since TSMC decided to expand into Japan and it was decided that construction would be built here next to our factory, we've supported a lot of areas where they weren't familiar with in Japan. When constructing a factory, we provide advice on how Japan's methods are different from those in Taiwan, and when it comes to actually starting up a factory, we send our Sony engineers to TSMC to support the start-up. Sony hopes its partnership with JASM will benefit its development of image sensors, for smartphones and self-driving cars. What can be expected of EVs in five years, with your image sensing technology? I think that probably in five years, about 35% to 40% of the cars in the world will be replaced by EVs. At that point, the improvement of logic wafers is also necessary. For example, to create an image sensor with higher precision, the scale of the circuit will also increase, so logic wafers will need to evolve to match this.
TSMC is currently developing a roadmap from 22nm technology to the next 12nm technology and beyond. In addition, we are also loading the TSMC scene on our Sony roadmap. So we can match our Sony roadmap with the TSMC roadmap. Furthermore, by further miniaturizing TSMC's logic, there will be the benefit of lowering device power consumption. On the other hand, TSMC's expansion is an advantage for Japan, and with TSMC coming to Kumamoto, many suppliers are gathering in Kyushu. We believe that being near a TSMC supplier is very important from the perspective of stabilizing the supply chain.
While the presence of JASM has boosted the economy in Kumamoto, questions have emerged abouta the long-term viability of this push to turn Kyushu into Japan’s own Silicon Valley. You talked about the challenges, what’s the biggest challenge you have to overcome? The current challenge for the residents of Kikuyo town is to welcome people from other countries to live harmoniously in a multicultural environment in Kikuyo town. In order to cope with this challenge, we have set up a Foreign Consultation Desk on the first floor of the Kikuyo Town Hall. I think the residents are a little confused because of the sudden changes.
Kikuyo’s 44,000-strong population is set to grow with JASM’s two plants bringing in an influx of foreign staff. In 2024, 750 Taiwanese workers and their families moved into town and nearby areas like Kumamoto city. Semiconductor manufacturing requires lots of land and water, which Kikuyo has.
But careful management is required to balance the needs of the town’s residents with the demands of JASM’s operations. I should say there are hardware and software approaches that we can implement. The hardware approach is securing the land and responding to various requests as they arise.
The software approach is securing human resources. There are new faculties being created at various schools, such as Kumamoto University, to train human resources in semiconductor, so we are collaborating with such institutions. The most important thing is that Kikuyo Town fully accepts JASM and people related to the company. What's important to that end is that we ensure that tax revenue, increased in line with the arrival of JASM, gets properly distributed to all of our townspeople. The key to the next wave of cutting edge technologies is undoubtedly semiconductors.
And Asia is at the forefront of fabricating the world’s tiniest chips. With the ambitious strategy to lure foreign investment and develop local capabilities, and deep pockets to fund research as well as foundries, Japan is set. JASM is planning to begin mass production soon, while Rapidus is scheduled to open a pilot fab next year. Once considered unlikely, Japan’s push to reclaim its past glory in semiconductors seems to be gathering momentum.
On my journey across Asia, in search of the brightest sparks, many burning questions have been answered. But now, one question remains. Which cutting edge technology will next take the world by storm? This is Momentum.
2024-11-30 11:25
