July 5, 2019

Here is where big corporates can order up a new startup like a pizza

Can a new breed of field lab help companies innovate better?

Maija Palmer

9 min read

“We don’t just match companies with startups, we go into the bedroom with them. We go to their wedding day,” says Arnaud Huvelin, head of industrialisation at IoT Valley, the tech ecosystem and one of Europe's innovation labs, on the outskirts of Toulouse in France. “[The PR team] hate when I use that metaphor, but it is true, we get deeply involved.”

IoT Valley is a combination of incubator, accelerator and corporate innovation lab. Like many company builders, it offers young companies cheap office space and lessons on how to build a business.

Arnaud Huvelin, head of industrialisation at IoT Valley

The unofficial motto IoT Valley, which was founded in 2012, is “no bullshit”. They don’t really do tech events and conferences. There are no foosball or ping pong tables here; no beanbags. Instead, there is a cluttered workshop that smells strongly of industrial solvent and a flowchart outlining how you go take a startup from idea to shipping the first 10,000 units.


IoT Valley, basically, does product and profits. Like a growing number of incubators, such as Founders Factory in London or Beta-i in Lisbon, it has a particular focus on match-making between corporations with problems and startups offering them potential solutions, in this case (the clue is in the name) solutions based on using internet of things, or IoT, technology.

Ordering up a new startup

IoT Valley is one of a number of organisations trying to broker the tricky corporate startup relationship. Corporations are notoriously bad at creating new innovations themselves. A 2014 study by marketing consultants Simon Kucher found that 72% of all new products are a flop.  Corporates have tried a number of approaches to staying fresh and current, from creating their own in-house startup or innovations labs to simply buying up promising new startups in their sector.

Corporate partners at IoT Valley
Corporate partners pay an annual fee to get access to IoT Valley innovation

Innovation labs run by third parties are one idea for trying to get around the deadlock.

Corporations such as SNCF, the French state-owned railway company and the bank BNP Paribas pay an annual subscription of up to €80,000 to be part of IoT Valley. In exchange they get some very tailored consulting on their business problems. In a programme they call “Atelier Catastrophe” the IoT Valley team invite the corporates to talk through their pain points and then work with them to see if any startups were developing a potential solution.

“If no one is solving the problem, then we will create a company to do it,” says Huvelin. There is a stable of would-be entrepreneurs who can be put to work on new companies like this — a little bit like creating a new boy band on X-Factor, back in the day.

We don’t create sexy startups. These are B2B businesses.

IoT Valley will only nurture startups with real customers. If you have a great idea but no clients, forget it. IoT Valley will actively discourage you from making so much as a prototype. Better to put your idea on hold and work on something a real customer is asking for.

The projects can sometimes be deeply unglamorous. Carmila, the property company, for example, told IoT Valley about a problem it was having with facility management. It is hard to keep track of all the maintenance jobs that need to be done in a large office building, every blocked toilet or blown lightbulb. A startup called MerciYanis stepped in to create a solution for this.

“We don’t create sexy startups. These are B2B businesses,” says Huvelin. But they are B2B businesses with real customers and revenues.

In return for helping them get started, IoT Valley will take 4% of any revenue that their fledgling companies create above €500,000, up to a maximum of €180,000 a year.  After the company reaches €4m the revenue share ends — an incentive for companies, which are sometimes concerned about being bled dry by an incubator that doesn’t set a cap.

Most of the company’s revenue, however, comes from the corporate subscriptions, says Thomas Gras, communications manager at IoT Valley.

Mine’s a lobster, please

Another valley, some 1200 kilometres away in Delft in the Netherlands, is experimenting with a similar idea.


RoboValley, founded in 2015, is a centre for robotics, where the University of Delft’s robotics lab collaborates with governments and companies on projects. Much like in IoT Valley, big companies like Rabobank, ABB and Ahold Delhaize partner with the organisation to get access to bright ideas from students, researchers and startups.

RoboValley is a little less single minded about the corporate focus. Some projects here are still very much in the realm of the student and amateur.

Stephan Rutten with the Lobster undersea exploration probe
Stephan Rutten with the Lobster undersea exploration probe

In one room a bunch of undergraduates are designing a deep sea exploration robot called the Lobster. It looks a bit like a torpedo and is filled with oil to keep the plastic structure stable even at great depths. If it is successful the Lobster could significantly decrease the costs of sending probes out to map and catalogue the ocean floor, but Stephan Rutten, one of the mechanical engineering students working on the project, isn’t really thinking about the commercial side yet.

“We are engineers, not business people. We’re not in a hurry to make a lot of money,” he shrugs.

On the other hand, some of the tenants at RoboValley are not researchers for hire, they are more like the typical young businesses at an incubator.  IMSystems, for example, is designing a high-precision gearbox without gears and is in the middle of pilot projects with a number of big producers of factory machinery. If all goes well, in a few years the company will have outgrown the cheap office space and be out on its own.

Somnox, a co-sleeping robot whose regular breathing pattern helps restless sleepers get a better night, also started as a robotics project here.

"Real life if buggy" says Jaimy Siebel
"Real life if buggy" says Jaimy Siebel when the haptic glove fails to work.

But there is a halfway point between these two: RoboHouse, a field lab that RoboValley opened earlier this year, where companies get researchers to build and test out robotics projects for them at relatively low risk and low cost. 

“We aim to be a be a bridge between academic research and industry,” says Anouschka Verleijen, managing director of RoboValley.

RoboHouse has developed some formulas to help corporations think about robotics and automation projects. The first step is usually to break down the project into a series of actions that need to be automated.

For example, say you wanted to automate chopping the stalk off a lettuce. You’d need to break that down into a series of actions: grab the lettuce, manipulate it round to the right orientation, find where the stalk is, make the cut.

RoboHouse has created a set of coloured cards pinned onto a whiteboard that help companies visualise this, and which show them all the different robotic options. Do you want a metal gripping claw to hold the lettuce or some kind of suction cup? It takes just 45 minutes to plan out the basics of a robotics project in this way, says Jaimy Siebel, manager of RoboHouse.

Once a company has defined what is needs, RoboHouse can pair them up with a team of researchers or students from Delft University to build it.

RoboHouse system for visualising a robotics project
RoboHouse helps corporate customers visualise automation projects with these coloured cards.

One past project involved Delft University students building a cheese-flipping device for Lely, the dairy automation company. The famous yellow, wax-covered gouda cheeses of the Netherlands need to be flipped and recoated every two days during the six weeks to a year they take to mature. Big dairies have already automated this for large volumes of cheese, but Lely wanted to explore a solution that could be used by smaller farms. It worked better than expected.

“They were so excited, there was a van waiting to drive the machine away as soon as the students’ final work had been presented,” says Siebel.

Cheese-flipping robot built for Lely
Cheese-flipping robot designed for Lely

Ahold Delhaize, the food retailer, meanwhile, is building a miniature supermarket in RoboHouse, to explore the use of robots in stores, and Heineken has installed a small production line in one of the rooms where it is testing a robot that can lift up fallen bottles as they trundle past on a conveyor belt.

“If Heineken did this themselves, it would be expensive. But here they can spend two years trying it out, and even if it doesn’t work it is still useful to have found that out,” says Siebel.

RoboValley is set up as a not-for-profit foundation and receives a mix of funding from industry, government grants, EU Horizon 2020 funding, and contributions from various educational partners. A project like Heineken’s might cost a couple of hundred thousand euros, but still far less than it would cost to do the same in an industrial facility.

Heineken robot that helps pick up fallen bottles
Party of one: an assembly line robot tested by Heineken

Things often go wrong, laughs Siebel. One project featuring a small drone that can locate parcels in a warehouse is on hold because the tiny machine has struggled to read the large barcodes on the sides of boxes. A virtual glove that can help you “feel” virtual objects is out of action because of a software problem. That is the point.

“Real life is buggy,” shrugs Siebel. Ironing out things like this is what the field lab is designed to do. It is the part that companies — which often have overly ambitious expectations of what robotics can do — struggle with initially.

“People often see videos, for example of impressive robots from a company like Boston Robotics and they don’t realise they have filmed thousands of takes before it works like that,” he says.

“That is why it is so important to see it in real life, to get your hands dirty,” says Siebel. “A field lab is your chance to do that.”

Other corporate-backed accelerators to check out:

Beta-i in Lisbon

Founders Factory in London

The Bakery in London]

Combient Foundry in Helsinki/Stockholm

?As you can see, this is far from being a complete list. Please let us know about other innovation labs matching up corporates and startups that we should feature.