He stopped designing batteries to solve the problem that haunted every lab he worked in: knowing, fast, whether a cell is any good.
Walk into any battery test lab and you will find rows of cells charging and discharging, quietly generating thousands of channels of data - voltage, current, temperature, capacity, the slow creep of degradation. For most of the industry's history, the only way to know if a cell was good was to wait. Tal Sholklapper got tired of waiting.
Today he is CEO and co-founder of Voltaiq, the company credited with coining the phrase "battery intelligence" and turning it into a category. The platform is the de facto standard in battery analytics - the layer that sits between raw test equipment and the engineers who need answers, and it is trusted inside the labs and manufacturing lines of Toyota, Mercedes-Benz, Meta, and Amazon.
That is the present. The origin is quieter, and stranger: it begins with a physicist-turned-materials-scientist who kept hitting the same wall in every job he held, and finally decided the wall was the business.
Before he ran a company, Sholklapper ran experiments. He was one of the people designing new batteries and fuel cells, first as a researcher at Lawrence Berkeley National Laboratory, then as co-founder of Point Source Power, a low-cost fuel-cell startup spun out of that lab work. After fuel cells came batteries: he served as lead engineer on a DOE ARPA-E funded project at the CUNY Energy Institute in New York, chasing an ultra-low-cost, grid-scale cell.
Across all of it, one annoyance refused to go away. As he tells it, the hardest part of building a battery was not the chemistry or the manufacturing - it was the lag between making a cell and learning whether it worked.
It all started about 13 years ago. I was one of the folks designing new batteries and fuel cells, and one of the key challenges I kept on seeing is it takes a long time to see if a battery is good or bad.
At the CUNY Energy Institute he was running two parallel ARPA-E energy storage projects alongside his eventual co-founder, Eli Leland. The two were physically driving out to the test lab to check on cells - a routine so tedious that they wrote a small piece of software to collect and analyze the data from a browser instead. That little convenience tool is the seed of everything Voltaiq became.
Most engineers, Sholklapper argues, reach for the wrong mental model. The automotive world tends to think in mechanical and electrical terms - parts that behave predictably and can be specced on a sheet. Batteries do not play along.
These are electrochemical living and breathing organisms that also have those mechanical properties you have to deal with and electrical properties you have to deal with.
That framing is the whole pitch. If a cell expands, contracts, ages, and misbehaves like something alive, then you cannot manage it with a one-time inspection. You have to watch it continuously, read its signals over time, and catch trouble before it ships. Voltaiq's software is built to do exactly that - to make the invisible electrochemical life of a battery legible to the humans responsible for it.
The alternative, in his telling, is a losing game. Quality defects slip through and only surface later, in the field, where they are expensive and dangerous to chase down.
A lot of these quality defects aren't being identified - we're sort of doing whack-a-mole trying to identify these in the field.
Sholklapper has become a candid voice on a tense subject: whether Western and domestic battery makers can catch Chinese cell suppliers on cost and efficiency. His read is not comfortable. A timeline he once measured in roughly five years has, by his own account, compressed.
It feels like it accelerated. Even from last year, we've continued to see efficiency improvements and lower costs.
His prescription is not nostalgia for old manufacturing - it is more software, more automation, more analytics, applied faster. New domestic plants, he notes, are not yet running at the efficiency or the price points the best Chinese suppliers hit today, and closing that gap is a training-and-data problem as much as a hardware one.
You need to really train domestic staff, and build in more automation, build in more analytics to allow people to do this faster.
It is a convenient argument for a man who sells analytics. It is also, awkwardly for skeptics, the same argument he was making back when he was the engineer driving out to the lab to read cells by hand.
The intellectual pedigree is dense. Sholklapper holds a BS in Physics and Applied Mathematics plus an MS and a PhD in Materials Science and Engineering, all from UC Berkeley. The PhD is the detail people remember: he is said to have finished it in two and a half years, among the fastest engineering doctorates the program has seen.
It is a useful thing to know about how he operates. Speed is not an accident of his career - it is a temperament. He compresses. The same instinct that turned a four-or-five-year academic slog into thirty months shows up in the company he built: stop driving to the lab, put the data in the browser, see the answer now.
It takes a long time to see if a battery is good or bad.
These are electrochemical living and breathing organisms.
We're sort of doing whack-a-mole trying to identify these in the field.
It feels like it accelerated. Even from last year, lower costs.
Build in more automation, build in more analytics, do this faster.
Stop driving to the lab. Read the battery from your browser.