Batteries x Software
Powering Japan’s Energy Future
Powering
At POWERD LAB, engineers turn ideas into solutions with exceptional speed. From digital twins and VR design to edge-deployed AI, we harness software to accelerate hardware innovation.
Expanding the Possibilities of Energy Storage with Software
Scalable Cloud,
Optimized Cost
We leverage auto-scaling and multi-cloud architectures to flexibly meet demand, embedding security and cost-efficiency at every level.
Our Code.
Total Control.
We develop and operate proprietary software that centrally manages the entire power flow—from batteries and inverters to EV charging stations.
RUL Prediction
with AI/ML
We deploy Remaining Useful Life (RUL) models powered by AI and machine learning to predict battery and device lifespan, enabling real-time performance analysis.
POWERD LAB’s Proprietary Development System
Digital Twin & MR Design
Test Virtually. Perfect in Reality.
Test Virtually.
At POWERD LAB, we use digital twins for real-time simulations and
VR/MR tools to review and optimize large-scale designs before they are built.
The CTO’s Vision for POWERD LAB
What we value most is maximizing the potential of energy through the power of software. We stand at the forefront of this mission, shaping the future of energy.
— Deepak Raut,
Where Physical and Digital Converge

In spring 2025, PowerX launched a new development hub called POWERD LAB. As CTO, I was deeply involved from the concept stage through to its realization.
We also leverage this setup for automated calibration, intelligent alerting, asset management, predictive maintenance, and inventory planning. As a result, our systems achieve lifecycles exceeding 15 years while significantly reducing total cost of ownership. In recent years, I’ve also been pushing forward with the application of AI and ML. Our BESS platforms generate massive amounts of data.
Over the past three years, we’ve analyzed extensive operational data to engineer features for our machine learning models. We follow an “80:20” data strategy—only about 20% of this data is sent to the cloud for high-level analytics and business insights, while 80% remains on-site at the edge, running fast inference models for battery management, auxiliary control, and power processing.
This is where adversarial learning comes in—the models at the edge and in the cloud compete and learn from each other, leading to faster convergence and higher optimization precision. Beneath it all, embedded real-time software leverages these insights to power real-world operations. All of this is managed in-house at POWERD LAB, where we’ve recreated a “mini power grid” for hands-on experimentation. We’re drawing on three years of historical data, which lets us build and test highly realistic scenarios.
Driving Japan’s BESS Innovation from Tokyo

