Solid-state transformers, purpose-built for the grid ahead - from the ex-Tesla team rebuilding the most overlooked box in the energy system.
Somewhere in Scotts Valley, a few cabinets hum where a building-sized iron transformer used to sit. The cabinets do the same job - turn high-voltage grid power into something a data center or a solar farm can actually use - but they do it with silicon, software, and a battery tucked inside. This is Heron Link, and it is the product of a company that has sold roughly 50 gigawatts of orders before its main factory has poured concrete.
Heron Power is two years old. It has raised $178 million. It has 83 people. And it is wagering all of that on a single, unglamorous bet: that the thing slowing down the electric grid is not power itself, but the century-old hardware we use to move it around.
Every wind farm, solar array, battery yard, and data center connects to the grid through a transformer. The design is roughly as old as the lightbulb: an iron core, copper windings, and oil. It is reliable, it is dumb, and right now it is in desperately short supply. Lead times stretch past a year. Projects that could be generating power sit idle, waiting on a box.
The irony writes itself. The world wants to electrify everything - cars, heat, compute - and the gating item is the least exciting component in the entire system. A transformer does not learn. It does not flex. When it fails, you wait, sometimes for the better part of two years.
Heron's read: if you replace the iron core with high-frequency power electronics, the transformer stops being a passive lump and starts being a controllable device - one that can be built on an assembly line instead of wound by hand, and repaired in minutes instead of months.
Drew Baglino spent nearly two decades at Tesla, ultimately as SVP of Powertrain and Energy. He helped build the Powerwall, Powerpack, and Megapack lines, worked on the 4680 cell, and grew Tesla's energy and charging businesses past $10 billion in annual revenue. He knew batteries, power electronics, and - crucially - how to manufacture hard things at volume.
When he left Tesla in 2024, the obvious move was another battery company. He picked the transformer instead. The bet is that the same playbook - take a proven physics, wrap it in software, and scale the manufacturing brutally - applies to grid hardware just as it did to cells. The backers agree: Capricorn led the Series A; Andreessen Horowitz's American Dynamism Fund and Breakthrough Energy Ventures co-led the Series B. Tesla alumni JB Straubel and Zach Kirkhorn are along for the ride.
Four numbers, one story: a young company moving at a pace its order book is forcing on it.
Heron Link converts medium-voltage grid electricity down to 800 volts using high-frequency switching. Each unit handles about 5 megawatts and is assembled from tens of identical modules. When one fails, a technician swaps it in roughly ten minutes - closer to changing a server blade than rewiring a substation. Built-in lithium-ion cells provide about 30 seconds of backup, enough to ride through a blink. Heron says the approach can cut roughly 70% of the traditional gear needed to connect a project to the grid.
An integrated solid-state transformer and battery-backup unit for AI and hyperscale campuses connecting directly to medium-voltage transmission.
A solid-state transformer and inverter combined into one scalable platform for utility-scale solar and grid-scale battery developers.
Because it's electronics, not iron, the box can regulate, monitor, and adapt - a transformer that finally has firmware.
Captioning a cabinet: it looks like a row of gray boxes. So does a data center, and that turned out to matter.
Fundraising stories are easy to inflate. Order books are harder. Heron says customer demand has run past 40-50 gigawatts - which is what pulled the Series B forward. Early customers include Intersect Power and Crusoe, the latter via a letter of intent to supply power-converter technology for large data center campuses in Texas. The demand splits almost evenly across three markets that all happen to be growing at once.
From seed-stage to scale-up in about nine months. The bars get longer; so do the lead times they're trying to kill.
Three buckets, roughly equal. When all three grow at once, the boring box becomes the prize.
Power electronics, purpose-built for the 21st-century grid.
Heron frames its job plainly: help the electricity sector grow faster and more affordably with scalable, reliable, software-integrated infrastructure. That means a transformer you can manufacture on a line, monitor from a dashboard, and repair without a crane. The factory is the strategy. A 40-gigawatt annual capacity would be roughly 10-15% of all transformer production outside China, and comparable to about half of Texas's peak power demand - every year.
The competition is twofold: the legacy giants who still wind iron (Siemens Energy, Hitachi Energy, GE Vernova, Eaton) and the simple inertia of the existing transformer, shortage and all. Heron is betting that a programmable box, made at volume, beats a proven box you can't get.
Return to Scotts Valley, and those few cabinets where a building-sized transformer used to live. If Heron is right, that swap happens millions of times over the next decade - at solar farms in the desert, battery yards on the grid's edge, and the data centers now drawing power like small cities. The bottleneck that kept clean energy and AI compute waiting in line gets a manufacturing answer instead of a queue.
It is, admittedly, a strange thing to get excited about. Transformers are the part of the grid nobody photographs. But the unglamorous parts are usually where the constraints hide - and Heron Power has decided the most valuable thing it can do is make the boring box interesting, fast, and available. The order book suggests it's not the only one who thinks so.
Final frame: a gray cabinet, humming, doing a hundred-year-old job in a way that finally fits the century it's in.
Video: search "Drew Baglino Heron Power" on YouTube for the Latitude Media "Catalyst" podcast episode and recent conference talks on grid power electronics. A dedicated product demo had not been published publicly at the time of writing.