Japan’s trade minister, Ryosei Akazawa, said Friday in Washington that artificial intelligence and semiconductors will carry “very significant weight” in the next round of projects selected under the country’s $550 billion investment agreement with the United States. The statement comes more than a year after the pact was signed and signals how strongly Japan now wants to shape the future of advanced technology supply chains, rather than simply financing traditional energy infrastructure.
Akazawa spoke after two formal rounds of investment commitments had already been announced. The first tranche directed $36 billion into American oil, natural gas, and mineral projects. The second, worth $73 billion, went to a nuclear development project spanning Tennessee and Alabama, as well as natural gas power plants in Pennsylvania and Texas. Together, these commitments amount to $109 billion of the total $550 billion envelope. Notably, none of that initial spending was targeted at artificial intelligence or semiconductors. Friday’s remarks signal that the third tranche will look very different.
How the Japan-US investment pact works
The investment framework was the centerpiece of last year’s tariff agreement between Washington and Tokyo. Under that agreement, the United States set tariffs on Japanese goods at a 15% ceiling and lowered the rate for automobiles. Both sides confirmed on Friday that no additional tariffs would be imposed beyond what had already been agreed. The broader structure gives Japan a distinctive role: it is not simply a list of corporate projects but a negotiated mechanism through which Japan chooses and directs billions of dollars into sectors considered strategically vital to the relationship.
Japan’s ability to publicly name future tranches is itself a noteworthy innovation. The country negotiated a financial commitment, a governance process, and a timeline for deciding what to buy in the United States. This means Tokyo can make deliberate choices about which industries receive its capital, then announce those choices in a structured public way. Akazawa’s latest statement is effectively a forward-looking signal to both governments, industry executives, and investors that AI and semiconductor-related projects will be central to the next phase of cooperation.
The timing is important. The first two tranches were dominated by energy security and critical minerals, reflecting the immediate economic pressures caused by the war in Ukraine and by disruptions in global commodity markets. Those remain important priorities, but the technology landscape has shifted sharply since the pact was first negotiated. The explosive growth of generative artificial intelligence, rising demand for advanced data centers, and export controls around cutting-edge chips have made AI hardware one of the most contested trade battlegrounds in the world. Japan is now moving its bilateral investment vehicle into that fight.
What Japan might target in the next round
Japan has strong reasons to focus on AI and chips. Its own semiconductor industry retains leadership in many upstream materials and manufacturing equipment, while the United States produces many of the world’s most advanced chip designs and AI processors. Coordinating investment in those areas could strengthen supply chains that connect American design capability with Japanese manufacturing expertise. It also gives Japan a voice in American industrial policy decisions related to chip production and data-center development.
AI infrastructure is increasingly expensive, and the capital requirements for cutting-edge facilities are enormous. Japanese institutional investors and corporations could participate in the construction of new American data centers, AI computing complexes, or advanced packaging facilities. Semiconductor production has also become a major focus of government policy in the United States, with programs like the CHIPS and Science Act already supporting domestic fabrication plants. A Japanese allocation under the $550 billion pact could complement those efforts, possibly encouraging joint ventures between Japanese and American firms.
The mention of semiconductors also carries geopolitical undertones. Chips are central to national security, and both Washington and Tokyo have been working to reduce their dependence on a small number of manufacturing locations. By directing investment toward AI and chip projects, Japan can help build alternative sources of supply in friendly countries. That helps allies hedge against the disruption of existing trade routes and gives Tokyo a more direct role in shaping the development of the semiconductor ecosystem.
Europe’s parallel framework takes a different shape
Europe signed its own trade and investment framework with the United States roughly one month after Japan, in August 2025, under the same 15% tariff ceiling. Yet the comparison between the two arrangements reveals a crucial structural difference. The joint statement accompanying the European deal says European companies “are expected to invest an additional $600 billion across strategic sectors in the United States through 2028.” That figure, though large, is a prediction about what private firms will do rather than a funded commitment controlled by European officials.
In Japan’s model, government representatives can stand before the press and announce which projects will be funded and which sectors will get priority. No equivalent machinery exists in the European framework. There is no European official who selects specific American projects and no equivalent to the Japanese tranche announcement process. The difference means that Japan has a more active, coordinated role in directing capital, while Europe has agreed to a statement of anticipated private-sector behavior.
The European document also includes language on technology purchases. It says the European Union “intends to purchase at least $40 billion worth of US AI chips for its computing centres,” alongside $750 billion of American energy through 2028. The AI chip purchase commitment is particularly striking given that the United States is home to the three most prominent advanced chip designers used in AI systems: Nvidia, AMD, and Qualcomm. These chips are the essential input for the AI supercomputers and data centers now being built around the world.
Europe’s AI gigafactory plans
European governments have moved in parallel to develop their own AI infrastructure. Earlier this year, Brussels opened bidding for up to seven AI gigafactories with an estimated total cost of around €30 billion. Of that amount, roughly €10 billion would come from public sources, with only about €1 billion of the European Union’s share actually committed to date. The funding gap illustrates the gap between ambitious public targets and the reality of national budgets, and it makes the promised $40 billion purchase of American chips politically and financially important.
Each planned gigafactory is meant to house at least 100,000 advanced AI chips. If all seven sites are built, they would require a total of at least 700,000 high-end processors, almost certainly sourced from American suppliers. Those chips do not come cheaply; advanced AI processors can cost tens of thousands of dollars per unit, and system-level designs are even more expensive when memory, networking, and cooling are included. The $40 billion purchase commitment may therefore align quite closely with the scale of equipment needed to fill these facilities, although significant additional funding will still be required.
European governments approved the trade deal despite concerns that the investment and purchase commitments are not fully binding. The pact passed through the European Parliament, but the technology debate did not end with that vote. Many European lawmakers and industry executives have questioned whether the bloc should be committing to buy American chip technology while simultaneously trying to build its own semiconductor competencies. Others argue that no viable alternative exists at present, since Europe lacks large-scale production of cutting-edge AI processors.
Two different strategies for two allies
The comparison between Japan and Europe is not about which arrangement is better; it is about different philosophies of economic statecraft. Japan negotiated a mechanism that allows a single ministry to direct investment according to strategic priorities. The results may be slower because they require official selection processes, but they provide transparency and predictability. Each announced tranche is a tangible expression of the partnership, and it gives Japanese officials a way to explain to domestic audiences what they are getting from the agreement.
Europe, by contrast, created a framework that relies largely on market actors. Its $600 billion figure is a forecast, not a treasury account. The enforceable parts are more visible in the side commitments, such as the purchase of American energy and chips. That does not necessarily make the European approach less effective; private companies often make better investment decisions than governments. But it does make the political and strategic dimension very different. Japan can decide where its money goes through a formal government process, while Europe has essentially agreed on spending categories and allowed individual companies to make their own choices.
The next few months will show whether Japan’s third tranche can provide meaningful momentum for AI and semiconductor projects. If the tranche is large enough, it could support new data centers, advanced packaging facilities, or semiconductor equipment agreements that bind the two economies even more closely. It would also demonstrate that the $550 billion pact is not simply a political gesture but a flexible tool for responding to a fast-moving technological landscape. With global demand for AI chips still rising sharply and trade barriers increasingly common, the practical decisions made in the third tranche could matter just as much as the underlying tariff agreement.