Humanoid robots shelf-stack in a Danish store: A taste of the future?

At a grocery store in Lyngby, 12 kilometers north of Copenhagen, humanoid robots are being trained to pick up products and stack them on store shelves.

The hope is that the robots could soon be clocking in for the night shift, while human workers sleep.

Since June, their robots have been in training at this Føtex grocery store after closing time, allowing the bots to operate in a controlled, safe environment.

"We're starting out with replenishing all the different kinds of products on the shelves and we'll move on to trimming or date checking or maybe taking away cardboard boxes," explains Hive Robots co-founder and CEO, Charlotte Sand. 

"Then, while we're already working the aisles, we can also start scanning for out-of-stock products or misplaced items. There are kind of no limits to the very, very end goal."

For now, the robots – made by Chinese company Unitree – are slower than their human colleagues, but Sand says with the help of artificial intelligence, that could soon change.

And unlike human workers, the robots could in principle work round the clock.

"With the speed of AI's development, we will also be able to carry out more and to do our tasks better," she says.

"Eventually we'll work all night, all day. So, maybe it doesn't have to be that fast, it can just continue working."

Changing environment 

Supermarket environments constantly change, so training the robot presents a technical challenge.

Pasta packets can fall, glass can break, and shelves can be arranged slightly differently from one day to another.

That's why, rather than writing code, Hive developers are training an AI algorithm to handle an array of products, based on real-world data collected in this grocery store.

"We're controlling the robot to collect data and restocking cans, restocking shelves," explains robot pilot Roger Sala, who controls the humanoid robot using virtual reality glasses and two handheld controllers.

"Then we can use this data to train our AI models to basically perform the same thing but autonomously."

It's challenging – the robots have limited tactile sensing and there's some control delay.

Sala says they're avoiding fragile products like glass jars for now, and pasta packets can be particularly fiddly.

"If we want the robot to do a good job, we need to train it to do the same we do," explains Sala. "And there are a lot of small things that we as humans don't think about that basically we need to teach the robot."

Labor replacements

Operating grocery stores including Føtex and Netto, Salling Group is Denmark's largest retail group.

According to its 2025 report, it operates over 2,000 stores with almost 70,000 workers in Denmark, Germany, Poland, Latvia, Estonia and Lithuania. Føtex operations director Simon Kudsk estimates 40-50% of workers' time is taken up restocking shelves.

Denmark, a Scandinavian country of six million, in August reported an unemployment rate of 3.1%, below the EU average of 6.1%. A study published last year by the International Monetary Fund said about 20% of Denmark's labor market is at risk of job displacement by AI, similar to other European countries.

Kudsk says the goal is not to replace human workers but to support them, so they can perform more "value-adding activities" such as customer service.

"Our vision or the dream right now is that we eventually could have one or two humanoids in every Føtex store that could support our colleagues to make it more efficient and create more time for value-adding activities focused on customers," he says.

And Sand says retailers like Føtex struggle to attract staff to complete such jobs.

"For them, their staff can focus on customers, while the robot will do the hard work," she adds.

Timescales

Kudsk admits it could take up to five years before every Føtex store has a robot worker.

Danish legislation also currently does not support such humanoid robots working independently among human shoppers, so a law change will be required, plus new safety rules.

But Kudsk isn't deterred – after shelf-stacking, he believes best-before date checking – another time-consuming task – could be automated.

"Replenishing and filling up shelves is really the main target at the moment, but we see a huge potential in best-before date checks," he says.

"That's a huge task for our colleagues today. It's really hard to do and you need to be really detailed and we use a lot of hours in that matter every year."

At Hive Robots, they have closer goals – hoping by the start of 2027, their robots could begin autonomously restocking grocery store shelves at night.

The company plans to offer its humanoid robot retail technology as a service, with stores paying by the hour, rather than purchasing the robots outright.

Sand says the first milestone is to be as "good as your worst employee".

"We will work on from there, increase speed, increase preciseness and include more skills," she says, adding they have also attracted interest from other European retail chains.

"By New Year's, we just hope to be able to restock some of the products like the cans, the pastas that everybody is buying – because the worst nightmare for any retailer is empty shelves."