Why would a 1 MW backup system emphasize only 60 seconds? Because power and duration describe a duty, not the sufficiency of an entire facility design.
The cited company says the disclosed system made its first public appearance at YOTTA 2026. In a five-module configuration, the company states that each cabinet delivers 1 MW rated discharge power and at least 60 seconds of backup at a 1 MW load. Every number and function below remains a company claim, not independent verification.
Disclosure and boundary
1 MW and 60 seconds describe the disclosed five-module configuration
The cited company describes a hybrid of lithium-ion capacitors (LIC) and lithium iron phosphate (LFP) batteries, direct connection to an 800 V DC busbar (±400 V), and a portfolio that also includes rack-level battery backup units (BBU). The disclosure does not establish an industry standard, prove universal compatibility or show that 60 seconds is sufficient for a data-center site.
Duty layers
Separate transient response, seconds-scale ride-through and minutes-scale backup
A layered design assigns each timescale a job. Transient support addresses rapid load movement; seconds-scale ride-through may bridge a short transfer or stabilize a bus; longer backup supports the recovery sequence defined by the facility. Response time, power curve, duration, transfer logic, controls and recovery target must be specified together.
Architecture
LIC, LFP, BBU and conventional UPS/VRLA occupy different positions
LIC may be evaluated for frequent pulse cycling, LFP for stored backup energy, and BBU for protection closer to compute loads. Conventional UPS/VRLA can still have a role where the architecture, high-rate discharge curve, runtime, maintenance method, temperature and complete system validation support it. No chemistry should be assumed to cover every AI load.
AGM applicability
AGM suitability begins with the battery and system duty
For a short-duration high-rate UPS battery, separator discussions may include thickness, basis weight, acid absorption, electrical resistance, compression response, dimensions and batch consistency. These inputs must be tied to plate design, electrolyte, assembly, charging, thermal conditions and the complete-battery test plan.
Evidence boundary
AGM separator data cannot prove system-level backup performance
Separator measurements can support material comparison and specification matching. They cannot alone prove capacity, 1 MW output, 60-second runtime, life, safety, compatibility or a customer result. Complete-battery and system tests remain necessary, and Viking does not claim use in or compatibility with the disclosed system.
FAQ
Four questions buyers ask about layered backup
Why layer backup power? Different timescales perform different jobs. What does 1 MW / 60 seconds mean? Only company-stated capability for the disclosed five-module configuration. Does UPS/VRLA still have a place? Potentially, when architecture and validation support it. Can AGM data prove backup performance? No—separator data cannot replace complete-battery and system validation.
Target power
State continuous power, pulse power and the time-based load curve.
Backup duration
Separate transient, seconds-scale and minutes-scale requirements.
Architecture
Confirm AC/DC topology, bus voltage, UPS/BBU position, transfer logic and controls.
Environment and maintenance
Provide temperature range, space, cooling, service access and replacement method.
Battery structure
Provide plate-group design, compression, electrolyte and complete-battery validation scope.
