In ongoing tariff negotiations between Japan and the US, the Japanese government has proposed purchasing semiconductor products from American manufacturers worth several billion US dollars. This move aims to serve as a bargaining chip for reducing the US trade deficit with Japan.
According to Asahi Shimbun, sources within the Japanese government revealed that during past tariff discussions, Japan suggested procurement plans primarily focused on products from US semiconductor giant Nvidia, involving transactions at the multi-billion-dollar level.
The Japanese government intends to encourage telecommunications companies and IT enterprises operating AI data centers to increase their purchase of AI chips through subsidies.
If realized, this initiative could boost Japan's imports from the US to a scale of hundreds of billions to a trillion Japanese yen (several billion US dollars), potentially offsetting about 10% of the trade deficit. The US trade deficit with Japan stood at US$68.5 billion in 2024.
Beyond chip purchases, Japan has also proposed supporting essential materials like wafers and chemical supplies needed for semiconductor production within the US. Strengthening the chip supply chain jointly between Japan and the US is expected to enhance economic security.
According to Maeil Business Newspaper, the US is reportedly sticking to its position that it is only able to negotiate on the additional differential tariffs applied by country in addition to the 10% reciprocal tariffs imposed across the board and that it is not able to make adjustments on item-specific tariffs such as automobile tariffs.
Japan maintains that tariffs on automobiles, which account for around 30% of exports to the US, must be reduced, and the positions of the two sides remain unchanged.
Minister of State for Economic and Fiscal Policy Ryosei Akazawa is scheduled to leave for the US on May 29 and hold the fourth round of tariff negotiations with US Secretary of the Treasury Scott Bessent and others in Washington DC.
Article edited by Jack Wu
