Software

Software Development

Energy management for power storage facilities and AI data centers, Model-as-a-Service for AI, assistive technology and unmanned aircraft systems.

EMS · Energy Management System

Energy Management System Software for BESS and Power Plants

Software for monitoring, controlling and analyzing battery energy storage systems (BESS), solar and wind power plants—from the field device to the cloud.

EMS monitoring dashboard for an energy storage power station on a desktop monitor
01

Equipment Control & Monitoring

  • Real-time monitoring of equipment operating status
  • SCADA-based secondary data acquisition, aggregation and monitoring
  • Cloud-based configuration of control parameters and strategies
  • Local reception, processing and execution of control commands
  • Ultra-low-latency control for real-time power system operations
02

Power Plant Data Analytics

  • Comprehensive analysis for BESS, solar and wind power facilities
  • Revenue data tracking and performance analysis
  • Analytical reports and data export
03

Energy Management & Control

  • Distributed energy management and control
  • Power supply optimization
04

Electrical Circuit Topology Visualization

  • Visualization of the electrical topology across the entire BESS facility or power plant
  • Network configuration, equipment information at key nodes and real-time metering data for each node
05

Revenue & Operational Management

  • Monitoring and analysis of revenue, operational efficiency and KPIs
  • Individual accounts for each electricity distribution destination
  • Segregated management of operational data per destination
06

Cloud Platform / Network Security

Standardized centralized monitoring

  • Collects, aggregates and analyzes computational results from individual BESS stations and presents comprehensive analytical reports
  • Supports safety assessments, fault diagnostics, predictive fire risk detection and security management operations
  • Continuously monitors the operational status of other electrical equipment within each BESS station, detects faults in real time and automatically generates alarms when abnormalities are identified
For AI Data Centers

BESS and Intelligent EMS for AI Data Centers

In AI data centers equipped with high-density GPUs such as NVIDIA B300, GB300 and Rubin, battery energy storage is critical infrastructure—not just emergency backup.

BESS plays an essential role in power quality stabilization, rapid GPU load fluctuation management, uninterrupted power supply support and electricity cost optimization. For large-scale AI data centers from 30 MW to 100 MW, integrating BESS with an advanced EMS significantly enhances power supply reliability, operational stability and economic efficiency.

30–100 MWTarget data center scale
5Core BESS functions
AI data center hall with GPU server racks and a coolant distribution unit (CDU)
01

Fast Power Response

Rapid charging and discharging compensates for sudden fluctuations in GPU power demand. Coordinated PCS and EMS control mitigates rapid load changes and their impact on facility infrastructure and the utility grid.

02

Backup Power & Continuity

Supports power continuity during grid outages, voltage sags and other disturbances. Integrated with UPS systems, BESS helps prevent GPU server shutdowns, computational interruptions and data loss.

03

Peak Shaving & Load Shifting

Charges during low-tariff periods and discharges during high-tariff periods, reducing demand charges and overall electricity costs depending on the tariff structure.

04

GPU Load Smoothing

AI training and inference generate large, rapid swings in power consumption. BESS smooths these fluctuations, stabilizing the demand seen by the grid.

05

Intelligent Energy Management

An integrated EMS coordinates BESS, UPS, power distribution, emergency generators and cooling for optimized distribution and efficiency across the data center.

06

Functions and Roles of BESS

  • Uninterruptible power supply for cooling equipment
  • Power stabilization for cooling systems
  • Response to GPU workload fluctuations
  • Peak power demand reduction (peak shaving)
  • Emergency backup power for cooling systems
  • Power optimization through energy management systems (EMS)

Example: 30 MW AI Data Center

Estimates based on a power demand requiring 30 MW of BESS capacity.

BESS ConfigurationTheoretical Supply Duration at 30 MWPrimary Application
30 MW / 15 MWh30 minRapid load compensation and short-duration backup
30 MW / 30 MWh1 hourPeak shaving and backup power
30 MW / 60 MWh2 hoursElectricity cost optimization and backup power
30 MW / 120 MWh4 hoursExtended backup power and load shifting
Our Product

EMS Software Suite

An EMS and monitoring platform for the safe operation of energy storage facilities.

BWS
Battery Warning System
  • Real-time risk assessment
  • One-click system self-diagnostics
  • Automatic fault alerts
  • Advanced algorithm-based support
EDR
Energy Storage Power Station Event Data Recording System
  • Multi-source data recording
  • Real-time fault event recording
  • Fault location identification and visualization
  • Highly reliable and secure data traceability
EMS
Energy Monitoring and Management System
  • Real-time monitoring of equipment operating data
  • Power measurement and energy management strategy control
  • Data analysis and visualization
  • System-wide equipment topology visualization
  • Substation account management
HFU
High-Speed Frequency Control Unit
  • High-speed frequency detection
  • Rapid execution of frequency regulation strategies
  • High-speed control command activation
  • Fast-response coordinated power control across multiple units
OS
Operator Workstation
  • Dynamic process flow diagram visualization
  • Alarm monitoring and management
  • System status and operational parameter monitoring
  • Equipment monitoring and control
MP
Monitoring Platform — Standardized Centralized Monitoring
  • Collects and consolidates operational data and computational results from multiple BESS stations, providing comprehensive analytical reports
  • Supports safety assessments, fault diagnostics, predictive detection of fire risks and security management operations
  • Continuously monitors the operational status of other electrical equipment, detects faults and abnormalities in real time and automatically generates alerts
MaaS · Model as a Service

Hybrid Architecture for Proprietary and Open-Weight AI Models

Our MaaS platform transforms access to multiple AI models into a unified, enterprise-grade inference service.

Proprietary models

GPT, Claude, Gemini and more — via official APIs

Because the weights of these models are not publicly available, we access them through the official APIs of each vendor and integrate them into a single, unified API gateway.

  • Unified API interface for multiple AI models
  • Centralized authentication and access control
  • Token usage tracking and cost management
  • Model routing, monitoring and failover
  • Standardized model invocation

No need to build and maintain separate integrations with each AI vendor.

Open-weight models

DeepSeek, Qwen, Llama, GLM, MiniMax and more — on our own GPUs

Subject to their licenses, open-weight models will be deployed on our own GPU clusters once our data center is operational, exposed through standardized APIs.

  • Granular control over latency, bandwidth and SLAs
  • Optimized inference economics through KV caching, batching and elastic scaling
  • Tenant-level isolation of resources and data
  • Data security and sovereignty within our controlled environment
  • Seamless model switching by changing parameters, not applications

Balanced for enterprises

For enterprises operating multiple content platforms, our hybrid architecture balances:

  • AI model quality and performance
  • Data security and operational control
  • Service stability and reliability
  • Infrastructure efficiency and cost optimization

Expanding to multimodal and video generation

We plan to support advanced multimodal and video generation models, including future models such as MiniMax H3. Proprietary models such as Seedance will be integrated through official APIs or other authorized channels—covering text, image processing, multimodal reasoning and AI video generation.

Commercial-grade inference

Public availability of model weights does not make inference free. Production operation requires high-performance GPUs, storage, power and cooling, network bandwidth, monitoring, redundancy and model lifecycle management. We deliver commercial-grade AI inference designed for stability, reliability, scalability and cost efficiency—a scalable foundation for next-generation AI applications and digital content platforms.

Assistive Technology

AI-Powered Accessibility Software

We specialize in AI-powered accessibility solutions for people who are blind, visually impaired or have reading difficulties.

Our technology combines computer vision, optical character recognition (OCR), artificial intelligence, speech synthesis and voice commands to convert visual information into spoken information.

AI assistive smart glasses with built-in cameras

Main functions

TechnologyFunction
OCRReads printed and digital text aloud
Computer VisionRecognizes objects, products and visual information
Facial RecognitionIdentifies previously registered faces
Text-to-SpeechConverts recognized text into natural-sounding speech
Voice RecognitionEnables hands-free operation
Smart ReadingLocates requested text, numbers, dates and sections
AI AssistantAnswers questions and summarizes text
Digital MagnificationEnlarges text and adjusts its appearance for easier reading

Software architecture

  • Wearable camera / handheld device
  • Visual information capture
  • OCR / text recognition
  • Computer vision / AI
  • Voice command processing
  • Context & content analysis

Integration with AI data centers

Our accessibility platform can use centralized GPU infrastructure for advanced services such as real-time scene descriptions, multilingual assistance, video understanding and AI-powered navigation.

UAS · Unmanned Aircraft Systems

UAS and Counter-UAS

A UAS (Unmanned Aircraft System) is the complete system used to operate an aircraft without an onboard pilot—typically an unmanned aircraft (UAV or drone), a ground control station, communication links, navigation equipment, sensors and supporting software.

C-UAS (Counter-Unmanned Aircraft System) refers to technology designed to detect, identify, track and potentially neutralize unauthorized or hostile drones. We offer counter-drone capabilities as part of our proximity-data intelligence solutions.

Unmanned aircraft system (UAS)

RF Detection

Detects radio-frequency signals associated with drones.

Drone Tracking

Monitors detected drone activity and threats.

Threat Mitigation

Provides RF-based countermeasures against unauthorized drones.

Real-time Monitoring

Supports tactical situational awareness.

Mobile Deployment

Supports portable, stationary and airborne configurations.

System Integration

Integrates with broader intelligence and security operations.

Interested in our software?

Request detailed materials or talk to our team about your project.