Co-Founder & CEO / Fourier
Siva Yellamraju
He built the camera that follows you around the room. Then he left to build the box that makes hydrogen anywhere.
The fourth time is the one that took over him.
He built the camera that follows you around the room. Then he left to build the box that makes hydrogen anywhere.
Hydrogen has a transportation problem. It is light on energy-per-mass and miserable on energy-per-volume, which means trucking it across a state can cost more than making it. Siva Yellamraju's answer was to stop moving it at all.
Fourier, the Palo Alto company he co-founded in 2022, builds a hydrogen generator no larger than two server racks standing side by side. You wheel it onto a skid, connect water and power, and it produces hydrogen on-site and on-demand. Inside sit roughly twenty small electrolyzers - Fourier calls them blades - sharing one water pump and a set of power supplies originally manufactured by the billion for data centers. Fourier reprograms and retrofits those supplies to run electrolysis instead. The hardware is deliberately unremarkable. The intelligence is in the software that watches every blade and tunes it in real time.
That framing - hydrogen as a data problem rather than a chemistry problem - is the thing that makes Yellamraju different from the refinery builders. "Push the overall efficiency problem and production problem into a data-optimization problem," he says, and you can hear two decades of camera and video engineering in the sentence. The model is Tesla, not BASF: an array of small intelligent cells, managed like a battery pack, scaling like cloud compute.
Hydrogen is a great molecule in terms of mass to energy ratio, but it's actually a pretty sucky molecule in terms of volume to energy ratio.
The economics are the point. Industrial buyers - pharmaceutical plants, chemical and petrochemical facilities, metal sintering shops, and increasingly data centers - pay somewhere between $13 and $20 a kilogram for hydrogen that arrives on a truck. Fourier aims to deliver at $6 to $7 a kilogram before any government incentive. "With our margin, they're still saving half the price of hydrogen," Yellamraju has said. The business can run either way: sell the hardware up front, or sell hydrogen-as-a-service with a payback around eighteen months.
Fourier is not Yellamraju's first company. It is his fourth, and the first three all ended in acquisitions by names you know. In 2007, during the financial crisis, he co-founded ViVu, browser-based video conferencing years before Zoom made the idea obvious. Polycom bought it, and he stayed on as a director of engineering. Then came Baarzo, applying deep convolutional networks to understanding video, audio, and images. Google acquired it, and he went on to run YouTube engineering. Next was Akruta, a smart speaker with a 360-degree camera - the technology folded into Apple, where its ideas surfaced as the iPad's Center Stage feature.
At Apple he worked on the cameras directly: Portrait mode, Cinematic mode, the features that decide what your phone keeps in focus. He has a degree in computer science from IIT Guwahati and an MBA from Stanford. By any normal measure, he had already won. Then he had young children, started thinking about climate and energy, and gave himself permission to poke at the space as a side project - maybe ten or fifteen percent of his time.
The more I got into it kind of took over me. There's no other reason. It just grew on me.
There is no origin-story heroics here, which is what makes it believable. He did not have a manifesto. He had a curiosity that refused to stay at ten percent. "Create a better, cheaper, faster energy that just doesn't constrain people" is how he describes the aim - abundance through distribution, not activism through scarcity.
The throughline across all four companies is harder to spot than the logos. ViVu, Baarzo, Akruta, Fourier - video conferencing, neural networks, cameras, electrolyzers. They look like four different careers. They are really one habit: take a problem that everyone treats as fixed hardware, and find the part of it that is secretly software. Compress video before bandwidth is cheap. Teach a network to see. Make a camera decide where to look. Make an electrolyzer tune itself. The molecule changed. The move did not.
An array of small PEM stacks, each one fully automated and tuned by software - the same logic that runs a battery pack or a server farm.
The cost gap Fourier is chasing, drawn to scale:
Yellamraju went in assuming electrolyzer stacks were commodities, interchangeable parts you bolt together. He learned the opposite - the mechanical design and construction of the stacks is where a surprising amount of innovation lives. The bigger jolt came from customers. He expected price to win every conversation. Instead, the thing buyers cared about most was safety, and many said they would pay more than they do today for a system that was meaningfully safer. For a founder coming from consumer cameras, learning to run teams of mechanical and manufacturing engineers instead of pure software talent was its own steep curve.
Jean-Baptiste Joseph Fourier was the first to correctly propose that gases in the atmosphere warm the planet - the greenhouse effect, two centuries early. Yellamraju calls him "the father of the greenhouse effect," and the name does double duty. Fourier transforms run through his entire engineering life, from video to cameras to signal processing. A transform takes a thing in one form and reveals it in another. "What we are doing is transforms," he says - water and electricity in, clean fuel out. It is the rare company name that is both a tribute and a thesis.
Where he is careful is the hype. He is openly skeptical of hydrogen passenger cars - "that just doesn't make any sense to me" - while bullish on heavy trucking, maritime, and aviation, where energy density actually decides the outcome. Batteries and hydrogen, in his telling, are not rivals but complements: lithium-ion for the short and frequent, hydrogen for the long and heavy. It is a refreshingly un-tribal position in a field full of true believers.
For a long stretch Fourier kept its head down. The website was a single thesis statement, the company technically only eighteen months old, the work largely in stealth. The proof points came quietly. An early lab pilot ran with an Indian pharmaceutical company. Letters of intent followed for US commercial pilots - one at an Ohio petrochemical plant, another at an aerospace parts manufacturer in Fremont. The plan Yellamraju laid out was deliberately staged: commercial pilots first, then production, then a wider commercial rollout, with the early geographic focus on the US, India, and Indonesia, and Europe deliberately left for later.
By 2026 the boxes were leaving the lab for good. Fourier announced a partnership with Payal Industrial Park to deploy a hydrogen-based energy storage system at its site in Dahej, Gujarat - exactly the kind of decentralized, industrial deployment the company was built for. It is a long way from Portrait mode, and a fitting place for a company named after the scientist who first warned the planet was warming.
The capital backs the ambition. The $18.5 million Series A, announced in April 2025, was led by General Catalyst and Paramark Ventures, with Airbus Ventures, Borusan Ventures, GSBackers, MCJ Collective, and Positive Ventures joining. Energy companies are capital-intensive and unforgiving on timing, something Yellamraju says plainly. But the flywheel he keeps describing - more deployments produce more data, more data builds better stacks and better algorithms, which win more deployments - is the kind of compounding loop a software founder knows how to love. If it turns, hydrogen stops being a thing you ship and becomes a thing you switch on.