MENU
Store Energy Locally — Discharge on Demand.
During simultaneous boot sequences or traffic surges, supercapacitors deliver instantaneous peak power, shielding PSUs and UPSs from severe transient stress.
Mitigate transient impact, maintain high conversion efficiency, and minimize overall power and cooling footprint.
Shipment Record (cells)
100,000,000+
units
Kick-start your technical evaluation and comparison process.
.png)
Simultaneous AI server boots, clustered training jobs, and post-sag recovery events generate massive inrush currents and voltage excursions. In modern data centers, the real challenge isn’t the steady state—it’s the transient envelope.
Overbuilding upstream assets to survive short-duration peaks reduces PSU efficiency, consumes UPS/breaker headroom, and forces cooling overshoot.
Panasonic Supercapacitors can be deployed at the Rack (48 V bus), Zone, or Distribution level. They store energy locally and release it at high speed, presenting a leveled load to upstream sources.
Result: PSUs remain in a high-efficiency region; UPS and distribution systems avoid nuisance trips; and thermal spikes are reduced—enabling more flexible cooling strategies and layout options.
“Stay running. Stay cool. Stop over-provisioning.”
A store-and-use power architecture makes this achievable.
Phenomenon:
Nightly batch jobs, auto-scaling, or orchestrated startups cause cascaded inrush events, briefly pushing PSUs, UPS units, and breakers toward their nameplate limits.
Impact:
PSUs drift outside their optimal efficiency band; input-current swing (ΔI) spikes. PDU/breaker margins tighten, increasing nuisance-trip risk.
Result:
Upstream systems must be sized for worst-case surges—wasting capacity, copper, and thermal budget, driving up TCO and consuming valuable rack U.
Phenomenon:
Streaming, inference, and storage workloads often run out of phase yet occasionally align, producing unexpected power peaks.
Impact:
Power and cooling plans become overly conservative; contracted redundancy inflates.
Result:
Resource allocation becomes rigid, limiting new rack additions and slowing tenant onboarding.
Phenomenon:
Limited ride-through and high inrush on restoration cause resets or mis-operations.
Impact:
In Medical and Industrial systems, uninterrupted and fail-safe operation is mandatory-instability is unacceptable.
Result:
Battery-only redundancy cannot fully absorb fast transients; maintenance load and lifecycle complexity increase.
By absorbing inrush currents and fast load spikes, Supercapacitors keep PSUs operating within their high-efficiency region (~50–80%). This reduces conversion losses, lowers thermal stress, and minimizes the need for derating.
Supercapacitors smooth input-current fluctuations (ΔI), keeping breakers below nuisance-trip thresholds. This allows required safety margins to be maintained without upstream over-provisioning.
Millisecond-class current delivery. Handles AI-server inrush and spikes locally, smoothing input current.
.png)
No chemical reaction for storage; supports repeated peak absorption with slow aging—far longer intervals than batteries.
Minimal characteristic drift across rack hot spots and seasonal variation.
Series/parallel configuration at rack, zone, or system level. Works upstream or downstream of PSU/UPS.
Track voltage, temperature, and ESR for early warning and planned maintenance.
Placement (Rack 48 V bus):
Supercapacitor module tied directly to each rack’s 48 V bus; discharge thresholds synchronized with the boot sequence.
Observed Effects:
ΔI reduced by ~60–70%; nuisance-trip rate down ~80–90%; PSU load points return toward nominal.
Operations:
More flexible boot windows; avoids upstream over-design.
Placement (Zone supercapacitor bank):
Place a bank on the zone bus; charge off-peak, discharge at composite peaks.
Observed Effects:
Peak reduction 5–10%; ΔI cut 50–65%; PUE improved 0.02–0.05.
Operations:
Minimizes re-wiring when onboarding tenants; shortens deployment lead times.
Placement (Below distribution panel):
Provide multi-second ride-through plus inrush absorption on restoration; fail-safe preference set to trip-first.
Observed Effects:
Several seconds of ride-through achieved; restarts drastically reduced (often zero in defined windows); inspection time reduced 30–40%.
Operations:
More stable scheduling; documentation aligned with regulatory and quality requirements.
(Example only. Contact us for detailed product options.
New products are also under development.)
ー A Proposal for Peak Power Countermeasures for AI Data Centers Using Supercapacitors ー
Kick-start your technical evaluation and comparison process.
Contact
Technical Inquiries & Engineering Consultation
From spec checks to pre-deployment studies and design support, Panasonic Industry engineers are here to help.
Typical requests:
Detailed specifications and technical data
Design suitability checks
Customization feasibility
Safety and durability information
Get in touch — we’ll connect you with the right technical expert.
Area/Country
