Panasonic Holdings announced on Monday, June 8, that it plans to begin mass production of battery cells for data centre applications at its Kansas plant in the financial year 2028, ending March 2029. The company will allocate approximately 350 billion yen – $2.18 billion – of a previously announced 500 billion yen AI infrastructure investment to its Energy unit over fiscal 2026-2028, with the remaining 150 billion yen directed to its Industry segment. Panasonic Energy CEO Kazuo Tadanobu described the unit’s 950 billion yen sales target for data centre energy storage in fiscal 2028 as a “minimum commitment” and said the business would aim to exceed 1 trillion yen. YourNewsClub places this announcement as confirmation that a company best known as a Tesla battery supplier is repositioning its North American manufacturing capacity toward the AI data centre stack.
The underlying power problem Panasonic is addressing is structural. AI servers consume electricity in large, rapid bursts during GPU workloads, creating voltage instability that can cause hardware failures. Standard backup power systems cannot respond quickly enough or sustain output long enough to manage these spikes. Panasonic Energy’s solution is a rack-mounted Battery Backup Unit that installs directly into server racks, providing both blackout protection and peak-shaving capability – storing electricity during off-peak periods and releasing it during demand spikes. Panasonic Energy holds an 80% global market share in distributed power supply systems for data centres, a position built on the company’s control of the full value chain from lithium-ion cell production through system integration. That vertical integration is what a dedicated battery manufacturer cannot replicate at comparable cost or speed.
The Kansas plant that will begin data centre battery production in fiscal 2028 currently produces EV batteries for Tesla. The pivot reflects a wider dynamic: EV battery demand growth has stalled relative to expectations under US tariff policy and the termination of IRA 30D tax credits, while data centre energy storage demand is, in Panasonic’s own language, “growing more than anticipated.” Panasonic Energy is also building a third plant in Mexico with mass production scheduled for fiscal 2028, and plans to expand existing lines in Japan. In FY2029, the company aims to increase Japanese lithium-ion cell production to approximately three times the FY2026 level. Supercapacitors for faster load fluctuations are also in development, with factory output targeted for fiscal 2027. YourNewsClub maps this as the clearest view of how a Japanese industrial manufacturer reroutes EV-scale battery production capacity into the AI infrastructure layer. The Kansas plant is already operational for Tesla battery cells; FY2028 data centre cell production at that facility describes a conversion already underway, not a greenfield commitment.
The energy infrastructure beat at Your News Club expects Panasonic’s next quarterly earnings call to carry an updated bookings figure for data centre storage – a number that will show whether hyperscaler demand has locked in the expanded capacity before a single new cell ships.
Owen Radner, who models digital infrastructure as energy-information transport systems, frames the logic: “What Panasonic is solving is not storage in the traditional sense – it is active power stability. AI servers behave like loads that spike unpredictably, and the data centre must be designed around peak load, not average load. The battery inside the rack is the buffer that makes the rest of the infrastructure function. Whoever controls that buffer at scale controls a non-optional component of the AI compute stack.”
Here is the part nobody likes to say out loud: Panasonic’s forward output is already sold. Data centre battery supply through 2026 is contracted to hyperscalers, leaving enterprise and co-location customers competing for less than 20% of planned supply. The Gartner forecast that energy shortages could restrict 40% of AI data centres by 2027 maps directly onto this dynamic. YourNewsClub places the non-hyperscale access gap at the top of the energy infrastructure watchlist – the Kansas expansion in fiscal 2028 addresses the constraint, but not before it becomes a structural bottleneck.