Can battery monitoring keep high-voltage AI server backup systems reliable? Two ICs combine cell monitoring with daisy-chain communication.
Nuvoton Technology has announced mass production of two battery monitoring ICs designed for battery backup units (BBUs) used in AI servers and other high-voltage energy storage systems. The KA49703A and KA49713A support battery architectures operating at 400V and 800V, with shipments scheduled to begin in September 2026.
The devices address the growing complexity of battery systems used in AI infrastructure. As server power requirements increase, backup systems need to monitor larger numbers of lithium-ion cells while maintaining reliable communication between monitoring devices. A daisy-chain architecture can reduce wiring and component requirements as battery packs scale.
Each IC can monitor up to 16 cells and connect with as many as 55 monitoring devices through the daisy-chain interface. This allows multiple ICs to be linked in series without requiring a dedicated microcontroller and isolation device for every monitoring IC, helping simplify the overall battery management architecture.
The devices support both daisy-chain and SPI communication and provide cell anomaly detection for battery safety. They offer voltage measurement accuracy of ±2.9 mV and are housed in a 7 mm × 7 mm QFP-48 package, helping fit monitoring circuitry into space-constrained AI-server power systems.
The KA49713A adds a transportation and storage mode that reduces shutdown current to below 0.1 μA. This is intended to limit battery discharge while large battery packs are transported or stored, potentially extending storage periods and reducing the need for recharging before deployment. The KA49703A has a shutdown current of 5.0 μA.
The monitoring ICs are designed for BBUs in AI servers as well as large-scale energy storage systems. Their architecture also builds on daisy-chain communication technology previously used in automotive battery applications and adapts it for higher-voltage industrial and infrastructure power systems.
With AI servers placing greater demands on power delivery and backup systems, monitoring individual cells accurately becomes important for maintaining battery reliability. The devices combine cell-level measurement, fault detection, and scalable communication in a compact package for these higher-voltage architectures.
Click here for the official announcement.
