Today’s digest served up a Habr post with a deceptively dull title: “The AutoStore Warehouse Constructor: How a Rubik’s Cube Became Norway’s First Unicorn.” Another “startup success story,” I thought, already about to scroll on — but one detail caught me: the warehouse dreamed up in a Norwegian village has not a single aisle. No racks, no corridors, no people — just a cube of containers and robots driving across its roof. And this August, a company with more warehouse robots of its own than anyone else on the planet suddenly remembered it: Amazon signed a global framework agreement with AutoStore.
How a TV repairman in a shed arrived at an “anti-warehouse,” why the idea was dismissed as madness and then fought over in the courts of three countries — and what air has to do with it, the air Hognaland decided to take out of freight logistics.
It all began in a red shed in the village of Nedre Vats, in western Norway. Jacob Hatteland started a TV repair shop there in 1971, and over time it grew into Scandinavia’s largest distributor of electronic components. By the mid-1990s, the business had expanded so much that the company built a new warehouse — and it filled up within a month. Not with goods, exactly — well, goods too, but that was not the real problem.
Its technical director, Ingvar Hognaland, diagnosed it in a single line, one that Synnøve Matre, who has worked at AutoStore since those days, still recalls:
“What’s the biggest thing in a warehouse? Air. There is too much air.”
A conventional warehouse is aisles for people and forklifts to move through. Every aisle is dead volume that pays rent but stores nothing. Hognaland proposed getting rid of the aisles altogether: pack containers of goods tightly together — stacked one on top of another, in a cubic grid — and retrieve them with something that did not need aisles to move through. First he tried a gantry crane. He quickly realized: wrong answer. They needed robots.
It took six years: the first prototype went into operation at Hatteland Logistics’ own warehouse in 2002; the first outside installation followed in 2005, at Eletec, a Norwegian supplier of security equipment. In a room at the former warehouse in Nedre Vats, there is still a museum of five generations of robots — assembled, improved, and remade again and again. Investors, according to then-CEO Carl Johan Lier, “didn’t really believe in us.” Mostly, Hatteland himself put up the money: “Jacob was willing to take risks because he believed this crazy idea could become great.”
A modern AutoStore looks like this: on the floor stands an enormous aluminum lattice — the “grid” — whose cells hold stacks of plastic bins full of goods. Across the top of the grid, as though across a roof, square robots zip along rail “highways” in two directions: they reach the required cell, lower a telescopic arm, grab the bin, and carry it to a port — a workstation where a person, or now a robotic arm, takes out the item. A person may never enter the storage area during an entire shift: it is physically impassable.
The tightly packed trick delivers storage density up to four times higher than a conventional rack warehouse: no gaps, no aisles, no ventilation for forklifts. Then comes the logistics magic that initially looks like a mistake: if the bin you need is buried beneath fifteen others, the robots — not one but several at once — “dig” it out, temporarily placing the upper bins on the surface of the grid, then stacking them back just as quickly. It is called bin digging, and from outside it looks like a pointless dance of boxes. Inside, it is an optimization problem solved by software in real time, which is why pulling a random item from the middle of the cube can be just as fast as taking one from the top shelf.
Then come the figures that warm an engineer’s heart: ten robots consume as much electricity as one vacuum cleaner. The system has no single point of failure: if one robot goes down, the others keep working, and a replacement can be rolled in without stopping the warehouse. An average project — 25,000 bins, 25 robots, 5 ports — is installed in six months. The robots are assembled in Poland, the software is written in Nedre Vats — and that pairing, “hardware from Poland + brains from a village,” remained unchanged even after the company went public.
Every good story about a brilliant idea needs the moment when someone tries to claim it. This one has an epic patent war with Ocado, the British online grocery giant.
Ocado was founded in 2000 by three former Goldman Sachs bankers, and by the early 2010s they had tried the whole evolution of warehouse automation: conveyors, carts, shuttles. Until one day in Stuttgart, they came across an AutoStore cube. Ocado’s engineers were electrified, bought an installation for their Fetch pet-supplies business, and in 2012 senior executives came to Nedre Vats with a proposal: let’s build an improved version together for grocery retail. And here the sides differ over why the partnership broke down. Ocado’s version is that “the robot arms took up space and did not move across the grid quickly enough” — they needed a different robot. AutoStore’s version is that it all came down to a visit by a Swedish Ocado engineer, who brought and showed the Norwegians a patent application for a robot with a central cavity instead of cantilevered arms. “They stopped trusting us,” the Financial Times sums up tersely.
Then comes the classic sequence: Ocado builds its “hives” — huge cubes swarming with 2,400 robots with cavity grips; AutoStore responds by painting its robots black, calling them the Black Line, and in October 2020 files patent-infringement suits in the UK and the US. “We will not tolerate the continued infringement of our rights,” Lier declares. The courts begin producing mixed results: in December 2021, a preliminary US ITC ruling goes Ocado’s way — the British company gains more than a billion pounds in value in a single day, while AutoStore drops; in March 2023, the British patent court rules definitively for Ocado. The war continues to this day, and both sides keep churning out patents: hundreds in the Norwegians’ names covering “motor placement in wheels” and “bin-moving algorithms.” The author of some of them, Hognaland, had already been retired for eight years by then.
The juiciest detail of this war is the scale of the enemy hives: two cubes measuring 240×80 meters, eight-meter tunnels, 180 ports, 45,000 product lines; an average order of 50 items is picked in five minutes — 5.5 seconds per item. Ocado’s own infrastructure, meanwhile, has caught fire twice: in 2019, a robot in Andover caught fire and the sprinklers had been mistakenly switched off — the entire hall burned down; in 2021, a collision involving three robots caused £10 million in damage at Erith. The hive is an impressive spectacle, but temperamental.
While the courts were sorting out who got to the cell first, money did its work. In 2017, the EQT fund bought AutoStore from the Hatteland family — the Financial Times puts that deal at roughly €500 million. Two years later, the company was resold for 15 billion Norwegian kroner — around $1.7 billion — and that was when AutoStore officially became Norway’s first unicorn. In April 2021, SoftBank acquired 40% for $2.8 billion. In October that same year came the IPO on the Oslo Stock Exchange, valuing it at $12.4 billion: Norway’s largest listing in two decades.
Then came a fall to $8.3 billion amid the patent war and legal uncertainty — followed by a slow but steady recovery. By 2026, AutoStore had nearly two thousand systems in more than 60 countries, over a thousand employees, and in 2022 it launched Pio — a smaller cube for small businesses, paid for as “robot as a service.” But the main thing happened in August.
On August 13, 2026, AutoStore Holdings announced a strategic framework agreement with Amazon.com Services LLC for the global supply of its solutions. The official press release is dry, as such things should be: the agreement “sets the terms for future purchases of systems but contains no purchase commitments at this time.” Put simply: Amazon has not bought anything yet — but it has come to the table and agreed on the terms under which it will buy.
And this is where the plot gets interesting. Amazon is the company that built the world’s largest army of warehouse robots: its own Kiva family, hundreds of thousands of machines in its fulfillment centers. Amazon set the standard for “the goods come to the person, not the person to the goods.” And yet, back in 2022, at the height of the patent war, the Financial Times wrote the following about Ocado’s approach, built on Hognaland’s idea: “it is even more advanced than Amazon’s.” In other words, the planet’s most advanced warehouse empire, the owner of Kiva, eventually came to negotiate with the heirs of a shed-born idea from Nedre Vats — even without guaranteed volumes. For comparison, elsewhere in an American warehouse, the cube is already live in production: CVS in Hainesport, New Jersey, using AutoStore alongside Tompkins sorting robots, processes more than two million items a week.
This story has a moral, and it is bitter in moderation. Hognaland retired in 2015, gives no interviews, and still lives in the same village. His name appears on most of the patents — but the patents themselves belong to a global corporation registered in Bermuda (“a Norwegian tax resident,” as the IPO prospectus puts it). Professor Nari Lee of Helsinki, commenting on this for the FT, said what everyone is already thinking: “We have this romantic conception that the inventor in a Norwegian village has an inalienable moral right to his patents. But that is not how the law works — and it has not worked that way for a long time.” When the journalist asked whether that was fair or sad, she answered briefly: “I think it is very sad.”
Though, to be completely honest, the idea itself was not absolutely new: containers in ports had always been stacked and retrieved by cranes. Kevin Kelly, co-founder of Wired, put it this way in an essay: “Ideas want to be shared. Sharing is their nature.” Everything brilliant is simple — and that is exactly why it is so hard to protect: once you prove there is another way to do something, the whole world begins doing it differently. Former AutoStore CEO Carl Johan Lier said it best: “The most brilliant ideas are always simple. People look at them and say, ‘It looks so easy.’”
It does — yes. Coming up with it, refining it through five generations of robots, taking giants to court, and waiting for Amazon itself to come to the negotiating table — not so much.