Chemistry gains are banked once and shared with every competitor. Control gains accumulate —and they belong to whoever owns the architecture.
The Question That Started the Company
When I began the engineering work that became Premergy’s patent portfolio, the electric vehicle industry had settled on a shared assumption: the path to better vehicles ran through better cells. More energy per kilogram, faster charging, longer cycle life. Every roadmap I saw was a chemistry roadmap.
I found myself asking a different question.
Given the cells that already existed—cells that were, by any historical standard, remarkable—how much performance was being left on the table by the way those cells were being used?
The answer turned out to be a great deal.
That answer is the foundation of everything Premergy has built since.
Chemistry Improves. Control Compounds.
There is a structural difference between the two layers that I do not think the industry has fully internalized.
Chemistry improvement is linear, capital-intensive, and slow. Commercial lithium-ion cells today reach roughly 275 Wh/kg, against a theoretical ceiling that practical engineering cannot approach closely. Industry projections put annual energy-density improvement at around 4%, and that rate has been moderating as electrode design runs into physical limits. Each increment requires new cell lines, new qualification programs, and years of validation. And when it arrives, it arrives for everyone at once—which means it is rarely a durable competitive advantage for anyone.
Control improvement behaves in the opposite way. An improvement to switching logic applies to every system running that logic, immediately, without new hardware. The next improvement builds on the previous one. Over-the-air deployment means the installed base gets better after it ships. And because control operates above the cell, the same improvement lands in an EV, a drone, a data center, and a grid installation without redesign.
Chemistry gains are banked once and shared with every competitor. Control gains accumulate, and they belong to whoever owns the architecture.
What We Built, and Why It Is Shaped That Way
The platform has three parts, and each exists because of a specific observation about how batteries fail to deliver their potential.
Multi-Chemistry Battery System
The multi-chemistry battery system exists because a single chemistry cannot be optimal across a variable duty cycle. We use two banks of distinct chemistries—one for steady-state power, one for surge—and switch between them in real time. The cells are ordinary. The organization is not.
Adaptive Energy Regeneration
The adaptive energy regeneration patents exist because conventional regenerative systems are designed for high-speed deceleration and do almost nothing in the low-speed, stop-and-go conditions where a great deal of real-world operation actually happens.
Our system captures energy at speeds as low as 5 mph and routes the recovered charge to whichever bank can absorb it most efficiently at that instant.
Battery Control System
The Battery Control System exists because the first two only work if something makes the right decision fast enough. It operates in three tiers:
- Supervisory Controller for system-level decisions.
- Drive Mode Controller for operating profile optimization.
- Switching Controller for real-time bank transitions.
One design constraint governed all of it: the system had to integrate into existing software architectures with minimal rewriting.
I have watched better technology than mine fail because adoption required a partner to rebuild their stack. Elegance that cannot be integrated is not engineering. It is a demonstration.
Why We Tested the Way We Did
I insisted on independent validation because I knew the claim would not be believed otherwise.
A greater than 20% efficiency improvement from control architecture alone, with no change to the cells, sounds like an overstatement. It is exactly the kind of claim that gets a company quietly dismissed by the engineers whose opinion matters most.
So we took it to Clemson University’s International Center for Automotive Research and ran more than 100 hours on a dynamometer, then more than 700 miles of on-track validation at Atlanta Motorsports Park. The efficiency result held.
The finding I care about most, though, receives less attention: across the entire test program, the thermal management system never activated.
Not once.
That was not luck. It is the direct consequence of never asking the energy-dense bank to absorb the current transients that generate heat. Thermal stress is the primary driver of degradation, and a system that does not generate the stress does not need to spend energy managing it.
Where I Think the Industry Goes Next
Three developments seem to me clearly directional.
First, Control Moves Into Software Permanently
Software-defined vehicles are the leading edge of a pattern that will reach data centers and grid installations. Once the control layer is software, it can be licensed, updated, and improved continuously—which changes both the engineering and the business model beneath it.
Second, Scale Makes Control Decisive
In a single pack, control is worth a few percent. In an array of thousands of modules under one point of dispatch, coordination overhead grows faster than capacity, and control becomes the limiting factor on what the installation can actually deliver against its nameplate rating.
Third, the Control Layer Becomes the Natural Home for Machine Learning
Our architecture was designed from the start to accept an AI overlay:
- Predictive load balancing that pre-positions energy ahead of demand.
- Self-learning switching algorithms that refine timing against measured outcomes rather than programmed parameters.
- Fleet-wide optimization across installations.
That is not a future product concept. It is the logical extension of a platform built to accommodate it.
The Layer Above Chemistry
I want to be precise about what I am claiming and what I am not.
I am not claiming chemistry does not matter.
Better cells are genuinely better, and the people building them are doing difficult and valuable work. Premergy does not compete with them—every improvement they deliver makes our systems better too, because we sit above their product rather than against it.
What I am claiming is that the marginal return on effort has shifted.
The next significant gains in battery-powered systems will come from the layer above the cell, because that is where the inefficiency now lives and where improvement compounds rather than resets.
We spent a decade patenting that layer while the industry’s attention was elsewhere. I did not know in advance that the market would arrive here in 2026. But the engineering reasoning was sound then, and the market has since supplied the confirmation.
