The Tiny Part Behind the AI Boom: Why Murata Is Betting ¥1 Trillion on Capacitors
Ask most people what powers artificial intelligence and they will name a chip maker. Few would mention a component smaller than a grain of rice. Yet that is exactly where Kyoto-based Murata Manufacturing has decided to place one of the largest bets in its history.
Murata says it will invest roughly ¥1 trillion over five years, from fiscal 2026 to fiscal 2030, to expand production of multilayer ceramic capacitors, better known as MLCCs. About ¥350 billion is earmarked for buildings and around ¥650 billion for production equipment. The company’s stated goal is ambitious: MLCC sales in fiscal 2030 should be at least double those of fiscal 2025.
What an MLCC actually does
An MLCC is a stack of extremely thin ceramic and metal layers that stores a small electrical charge and releases it in an instant. On a circuit board it acts like a shock absorber, smoothing out voltage so that sensitive processors receive steady power. Smartphones contain hundreds of them, cars thousands. The accelerator boards inside AI servers are especially hungry, because their processors draw huge and rapidly changing amounts of current.
That is why a sleepy-sounding component has turned into a growth story. Murata reported September-quarter operating profit of ¥147.5 billion on revenue of ¥588.5 billion, and it raised its full-year operating profit outlook to ¥520 billion. Its president, Norio Nakajima, has said that the company’s earlier capacity expansions were not enough to fill every order.
Where the money goes
The plan spreads investment across Japan and Southeast Asia rather than concentrating it in one place. The largest named project is a new MLCC production building at Okayama Murata Manufacturing in Setouchi, budgeted at around ¥150 billion. Construction is due to start in April 2027, with completion at the end of 2028, and the site aims to employ 1,500 to 2,000 people by 2030. Murata also plans a new electronics plant in Yasugi, Shimane Prefecture, a new building in Thailand and upgrades to an unused section of an existing factory in the Philippines.
Spreading production across Japan, ASEAN and China is a deliberate choice. A company that supplies components to almost every electronics maker on earth cannot afford a single point of failure, whether from earthquakes, trade frictions or power shortages.
Why investors hesitated
The market’s first reaction was not applause. Murata’s shares fell sharply on the day of the announcement, even though the quarterly numbers beat expectations. The worry is less about demand today than about timing tomorrow. Much of the new capacity arrives in 2028 and later, and the heavy equipment spending will add depreciation costs just as some analysts expect data-centre investment to cool. On the same day, Wall Street sold off technology stocks over fears that AI companies are burning cash too quickly.
Murata’s answer is that AI will not stay inside data centres. The company explicitly links its plan to edge devices, mobility and robotics, markets where every new sensor, motor controller and on-device AI chip needs its own ring of capacitors.
A quiet lesson about industrial power
The story also says something about Japan’s place in the technology race. Japan no longer leads in smartphones or leading-edge logic chips, but it remains dominant in many of the materials and passive components that those products cannot do without. Murata’s bet is a reminder that the AI economy is built not only on headline processors but on thousands of unglamorous parts made with extraordinary precision.
If demand holds, the investment will look prescient. If it fades, Murata will have spent heavily on capacity it cannot fill. Either way, the next time someone talks about the AI boom, it is worth remembering that it also runs on ceramic layers a few microns thick.