A 板块｜“北海道谷”把希望押在Rapidus身上
HOKKAIDO HAS long been known for dramatic volcanic peaks, creamy dairy products and powdery snow. These days, Japan's northernmost major island is being touted by enthusiasts as the country's next high-tech hub. “Hokkaido is the new Taiwan,” writes James Riney of Coral Capital, a Tokyo-based venture-capital firm. Japanese officials speak, somewhat breathlessly, of “Hokkaido Valley”.
The vision hinges on Rapidus, a bold if risky chipmaking effort. Launched in 2022, the company aims to bring Japan back to the frontier of advanced semiconductor manufacturing. Supported by the Japanese government and a consortium of big Japanese companies, Rapidus broke ground on its first fab in September 2023, in a field near the international airport in Chitose, a suburb south of Sapporo, Hokkaido's capital. Pilot production of 2-nanometre chips, the cutting edge for chipmaking technology, is under way. The ambitious effort is seen as a model for the sort of new-age industrial policy favoured by Japan's prime minister, Takaichi Sanae.
B 板块｜人才难补，商业成功更没有保证
Yet despite all the enthusiasm, the Hokkaido Valley plan faces several big obstacles. First is cultivating the third necessary resource for chip manufacturing: human talent. Hokkaido hopes to sell itself as a place where skilled engineers might want to live. True, it is bitterly cold in winter. But the island is home to world-famous ski spots. And local culinary delights, culled from the icy surrounding seas, include tasty scallops and sea urchins as well as enormous crabs.
Most important is the success of Rapidus itself. While the firm has made impressive technological advances, its commercial success is far from guaranteed. The true test will come when it begins mass production, which it aims to do next year. TSMC produces standardised chips at large scale; Rapidus, by contrast, plans to offer rapidly made, small batches of bespoke AI chips. Given the dearth of firms able to make cutting-edge semiconductors, and the growing demand for them, the strategy is plausible. But even in the best case, Rapidus will need to rely heavily on public subsidies. The government has already promised more than $15bn in support for the project. It is a big bet on a single firm. The Hokkaido Valley concept depends on it paying off.
