{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiauxjj5xxdsxwa3z6ydx7nhkpnj7pxxezlquwamblu5a2hpzbyph4",
    "uri": "at://did:plc:qzjwstutqk2cy7df7jbzd2hx/app.bsky.feed.post/3mhxw4ev5qou2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreietmh55tnjmwboujrne5xyy3vtnhznibknpkvgqho4bgvv7kwvnom"
    },
    "mimeType": "image/jpeg",
    "size": 16546443
  },
  "path": "/article/4150459/panasonic-says-datacenter-batteries-are-selling-out-and-ai-is-to-blame.html",
  "publishedAt": "2026-03-26T11:51:37.000Z",
  "site": "https://www.networkworld.com",
  "tags": [
    "Artificial Intelligence, Data Center, Data Center Management",
    "said in a statement",
    "HBM capacity",
    "2025 Data Center Trends predictions",
    "growing regulatory and utility scrutiny",
    "Five Data Center Predictions for 2026",
    "projected",
    "warned"
  ],
  "textContent": "Datacenter battery backup units are selling out years in advance, and artificial intelligence is driving the shortage — at least according to Panasonic Energy, which disclosed this week that hyperscaler customers have already pre-committed to more than 80% of its planned output through fiscal year 2029.\n\nThe disclosure came as the Japanese company outlined its plans to scale production to meet surging AI-driven demand. The company said it currently holds an 80% share of the distributed power supply market for data centers, citing its own research based on a Synergy Research survey as of December 2025.\n\n“The company has already reached agreements with customers on projects representing more than 80% of its planned sales,” Panasonic Energy said in a statement, adding that the pre-commitments represent “a long-term commitment with hyperscalers to work toward platform designs focused on the next and future generations of the technology.”\n\nThat leaves enterprise operators and colocation customers outside the hyperscaler tier competing for less than a fifth of Panasonic’s planned output — a pattern that has played out across the AI infrastructure stack. HBM capacity from Samsung, SK Hynix, and Micron has similarly been sold out through 2026, with allocations skewed toward the same pool of hyperscale customers.\n\n“This is not a set of isolated shortages,” said Sanchit Vir Gogia, chief analyst at Greyhound Research. “This is a structural tightening of the entire data center stack.”\n\nGogia said hyperscalers are extending their supply reservation strategy deeper into the infrastructure layer than before. “They are not just securing GPUs and memory. They are locking in the infrastructure required to make those systems stable and deployable.”\n\n## AI servers are rewriting the power rulebook\n\nThe root cause, Panasonic noted in the statement, is the electrical behavior of AI workloads. Unlike conventional server applications, AI inference and training draw large amounts of electricity in short bursts to sustain GPU processing, causing peak power levels to spike rapidly and voltages to fluctuate.\n\n“Peak power levels for such servers can rise rapidly, and voltages can often become unstable,” the statement said. “Securing stable, highly reliable power supplies is an absolute necessity for AI datacenters.”\n\nVertiv warned in its 2025 Data Center Trends predictions that AI racks must handle loads that “can fluctuate from a 10% idle to a 150% overload in a flash,” requiring UPS systems and batteries with significantly higher power densities than current infrastructure provides.\n\nPanasonic said the solution gaining traction among hyperscalers is to place a battery backup unit on each server rack rather than rely on centralized UPS infrastructure upstream, absorbing voltage instability at the source. The company said its systems also carry a peak shaving function that stores off-peak electricity and deploys it during demand spikes, reducing peak grid draw at a time when AI-driven consumption faces growing regulatory and utility scrutiny.\n\nSeveral independent research bodies have reached similar conclusions on the severity of the power challenge ahead. Uptime Institute, in its Five Data Center Predictions for 2026, said “developers will not outrun the power shortage,” with research analyst Max Smolaks warning the crisis “is likely to last many years.” The IEA projected global datacenter electricity consumption could exceed 1,000 TWh by 2026, more than double 2022 levels, while Gartner has warned that energy shortages could restrict 40% of AI datacenters by 2027.\n\nGogia said the shift runs deeper than a hardware swap. “This is not backup in the traditional sense. This is active stabilisation,” he said. “Power delivery is no longer passive. It behaves like a dynamic system with control loops, response thresholds, and continuous monitoring requirements.” Most enterprises, he added, are not ready. “Many enterprise datacentres were designed for a different era — lower densities, predictable loads, centralised assumptions. Retrofitting for AI workloads requires redesign, not just upgrades.”\n\n## The window is narrowing\n\nPanasonic said it is moving aggressively to expand supply to meet that demand. The company, as per the statement, plans to convert automotive lithium-ion cell lines to datacenter output from FY2027, establish a new module plant near its existing Mexico facility, and develop supercapacitor-based rack backup units with Panasonic Industry Co., Ltd., targeting load spikes that lithium-ion batteries alone cannot absorb quickly enough.\n\nGogia said the procurement window remains open, but the familiar playbook no longer applies. “The biggest mistake enterprises can make right now is to approach this as a late-stage procurement exercise. By the time procurement becomes the focus, the structural disadvantage is already in place.”\n\n“You do not begin with compute anymore. You begin with power availability, site readiness, and infrastructure compatibility,” he said. “Many facilities will claim readiness. Far fewer are actually engineered for it.”",
  "title": "Panasonic says datacenter batteries are selling out and AI is to blame"
}