Patrick Schwarzkopf leads the Robotics and Automation Association at VDMA, with about 400 member companies. He says Europe is squeezed in the middle between China, where the humanoid factories that produce at scale are, and a US that is banning imports of humanoid and mobile robots. He came back from Shenzhen describing a market where you fill a shopping basket with fingers, hands, arms and a head and go home to build your own humanoid.

In this interview from The Robot Bureau, Rogier and Lei sit down with Patrick Schwarzkopf, Managing Director of VDMA Robotics + Automation, at the Humanoid Robots Summit Europe in Stuttgart.

Patrick sees no gap in European mechatronics. Europe is strong in the drives, actuators and sensors a humanoid depends on, and by his count one robot can need 20 to 40 drives, yet the factories that produce at scale are in China. Start-ups raise a first round, run short when it is time to scale and end up sold abroad, a valley of death he wants pension funds and insurers to help close, the way France changed its rules so institutional investors put more money into robotics. Germany's SPRIND agency offers up to €7.2 million per team in its Next Frontier Robotics challenge, and Wandercraft is installing 350 robots in Renault factories within the next year or so.

"You cannot win a game if you play in defensive mode."

The US ban on foreign humanoid and mobile robots is a big problem for European makers of mobile robots, who can still export their old certified models but need to certify each new generation within a year or so. Patrick wants Europe to answer with its own supply chain from component to finished humanoid, and with a master plan and a north star, as China had for robotics. Public buyers in care, healthcare and critical infrastructure could give European start-ups their first customers.

Rogier and Lei cover: why Europe stopped arguing about whether robots need legs, why Patrick does not expect a humanoid winter, whether a transatlantic certification bridge is possible, how China makes humanoids popular with its public, and the driverless taxi ride that impressed Patrick most in Shenzhen.

Chapters:
00:00 VDMA and the Humanoid Robotics 2040 study
00:30 What Europe stands to lose
02:25 Playing to win instead of defending
03:32 Signs of acceleration in Europe
05:32 How likely is a humanoid winter
08:09 Safety, world models and public acceptance
11:32 The FCC ban and a transatlantic bridge
14:34 Public procurement and the valley of death
18:56 Europe needs a north star
20:26 Shenzhen's supply chain and Europe's components
24:04 Scale baby scale

Watch the video version on YouTube: https://www.youtube.com/watch?v=7L2tyUvX_bM

Follow VDMA Robotics + Automation Association
Org website: https://vdma.eu/en-GB/robotics-automation
YouTube: https://www.youtube.com/user/TheVDMA

Follow The Robot Bureau
YouTube: https://www.youtube.com/@therobotbureau
Spotify: https://open.spotify.com/show/19Rva9230EwGErD2fg0tQb
Apple Podcasts: https://podcasts.apple.com/podcast/id6812673688
Substack: https://substack.com/@therobotbureau
LinkedIn: https://www.linkedin.com/company/the-robot-bureau
X: https://x.com/therobotbureau
Host Lei Li: https://www.linkedin.com/in/leilibeyondthenow/
Host Rogier Bikker: https://www.linkedin.com/in/rogierbikker/

Francesco Ferro has been building humanoids for more than 20 years, and he says that when he spoke about bipedal humanoid robots, everybody told him it would never work. He was named Robotics CEO of the Year in 2024. PAL Robotics now has around 120 employees, opened in Toulouse in 2020 and in Bari in 2023, and he hopes to open in Germany this year. In this conversation he explains where he sees Europe in the race between the US and China, and what Europe can build on.

In this interview from The Robot Bureau, Rogier and Lei sit down with Francesco Ferro, CEO and co-founder of PAL Robotics, at the Humanoid Robots Summit Europe in Stuttgart.

As Francesco sees it, the US is spending to become the brain of all robotics and China, especially with government money, to be the only producer of the robot supply chain. Europe sits in the middle and, in his words, wakes up a little late, like always. His case against chasing a big valuation is about ownership, because the biggest venture capital funds are in the US or China, and once someone enters the shares, governments have no tools to stop a company from being bought by someone else. A company from Abu Dhabi helped PAL with different projects as it grew, and PAL uses components from China, Germany, Italy and Spain depending on price, quality and delivery time.

"Europe is just in the middle. We are now, like always, wake up a little bit late."

Francesco is against raising taxes on imports and wants Europe to work with China, the US, Japan and Korea, while keeping the know-how and not giving away, for example, the cheap production or the batteries in Europe. The Vienna Statement, as he describes it, asks that robots help people rather than substitute them, that Europe thinks long term, and that the more conservative companies in Europe open the door to collaboration. He also says the big companies PAL works with trust open source more and more, and Sony closing its Aibo section and never opening the code is his example of what happens when a company closes its code.

Rogier and Lei cover: how the cobot changed what people expect from robots, why Francesco does not see one dominant monopoly as the European way, what the Kinect gaming camera taught PAL about reusing parts from other industries, what PAL wants to show with industrial proofs of concept in the next year, and why he calls himself crazy for refusing to measure success by valuation.

Chapters:
00:00 Francesco, PAL Robotics and euRobotics
01:23 Robotics CEO of the Year 2024
02:45 From cobots to the US-China race
06:02 The Vienna Statement
07:22 What Europe can compete on
10:02 What PAL Robotics builds
13:53 Sourcing, the FCC ban and sovereignty
17:56 Why PAL is not VC funded
20:42 Ethics and a different idea of success
23:22 PAL's next 12 months
24:22 Open source and the Aibo lesson

Watch the video version on YouTube: https://www.youtube.com/watch?v=d0OZehrG5BM

Follow euRobotics
Org website: https://eu-robotics.net/
YouTube: https://www.youtube.com/user/euRobotics

Follow PAL Robotics
Company website: https://pal-robotics.com/
YouTube: https://www.youtube.com/PALRobotics

Follow The Robot Bureau
YouTube: https://www.youtube.com/@therobotbureau
Spotify: https://open.spotify.com/show/19Rva9230EwGErD2fg0tQb
Apple Podcasts: https://podcasts.apple.com/podcast/id6812673688
Substack: https://substack.com/@therobotbureau
LinkedIn: https://www.linkedin.com/company/the-robot-bureau
X: https://x.com/therobotbureau
Host Lei Li: https://www.linkedin.com/in/leilibeyondthenow/
Host Rogier Bikker: https://www.linkedin.com/in/rogierbikker/

Swisslog automates warehouses where robot cells already depalletise and pick, and its AI lead Jan Andreas Daske expects the first humanoid prototypes to be working in its sites within two years.

Swisslog is a KUKA company, and KUKA's majority owner is China's Midea Group. Jan sees humanoids unloading trucks and carrying orders from picking station to packing station 24 hours a day, and the problem he wants solved first is a plain one, what happens when a humanoid falls over on the way to the picking station or its battery runs low.

In this interview from The Robot Bureau, Rogier and Lei sit down with Jan Andreas Daske, Senior Director Global AI & Industry Intelligence Lead at Swisslog, at the Humanoid Robots Summit Europe in Stuttgart in September.

In the high-throughput core of a Swisslog site the work is already automated, with robot cells building mixed pallets and AI picking arms taking single items out of boxes, so Jan looks for humanoids at the handovers where a person still steps in, at the truck dock and in manual pick zones. Brownfield warehouses were built for people, which in his view may make a humanoid the best way in, and it still needs a software layer that coordinates it with the classical automation already on the floor. Swisslog sells that layer as its own SynQ software or as SAP EWM, on top of KUKA AMP, and by Jan's account the same integration know-how sometimes gets more than 20% extra performance out of an AutoStore system. He expects AI to break up today's monolithic systems such as warehouse management systems, with AI agents for inventory management, consolidation and reservation strategy as the next step, and over the next three to five years a lot of new automation that is more fragmented and better orchestrated by AI-driven software.

"Challenge me in two years, please."

Rogier raises a question he hears a lot in European logistics, that robotics companies coming mostly from China expect things to work as they do there and then find warehouses built 20 years ago. Jan's answer is that on a brownfield site a humanoid may be the best way in, as long as software controls and steers it and it meets CE certification, safety controls, collaborative safety rules and the new Cyber Resilience Act. By his account customers ask about capacity, technology, time and cost, and Swisslog's Chinese ownership normally does not come up. Jan describes Swisslog as the master solution integrator inside the KUKA environment.

Rogier and Lei cover: whether Jan has seen Midea's humanoids in its washing machine factory, why Swisslog is moving from loving steel to loving the process, what a humanoid can do that a fixed picking arm cannot, why he thinks two years may even be too slow, and the AI picking arm that fetches the pasta and a tin of tomatoes when you tell it you want spaghetti today.

Chapters:
00:00 Swisslog, KUKA and Midea's factory humanoids
01:41 Where humanoids fit in a Swisslog warehouse
03:42 KUKA AMP and the three software layers
05:50 Brownfield, greenfield and new EU rules
09:08 Midea ownership in customer talks
10:07 From loving steel to loving the process
12:02 Getting 20% more out of an AutoStore
13:35 AI agents and the end of the monolithic WMS
14:50 Humanoids across the whole workflow
16:32 A humanoid prototype within two years
18:07 Why two years may be too slow

Watch the video version on YouTube: https://www.youtube.com/watch?v=0mG-KyVDeJI

Follow Swisslog
Company website: https://www.swisslog.com
YouTube: https://www.youtube.com/@SwisslogLogisticsAutomation

Follow The Robot Bureau
YouTube: https://www.youtube.com/@therobotbureau
Spotify: https://open.spotify.com/show/19Rva9230EwGErD2fg0tQb
Apple Podcasts: https://podcasts.apple.com/podcast/id6812673688
Substack: https://substack.com/@therobotbureau
LinkedIn: https://www.linkedin.com/company/the-robot-bureau
X: https://x.com/therobotbureau
Host Lei Li: https://www.linkedin.com/in/leilibeyondthenow/
Host Rogier Bikker: https://www.linkedin.com/in/rogierbikker/

Maarten Witteveen, co-founder and CTO of MANUS, took his company from VR to entertainment to robotics, where it now makes data gloves that record how human hands work. He was in Los Angeles in 2023 when the actor and writer strikes sent entertainment off a cliff, and MANUS had to reinvent how it does business. In his five-year plan, everybody can wear the gloves for their whole work, and MANUS becomes an end-to-end data collection provider. This conversation explains why MANUS wants robots to learn from skilled people doing their real jobs, and what Maarten thinks that will do to those jobs.

In this interview from The Robot Bureau, Rogier and Lei sit down with Maarten Witteveen, co-founder and CTO of MANUS, at the Humanoid Robots Summit Europe in Stuttgart in September.

Each MANUS glove carries an electromagnetic sensor on every fingertip that measures its position against a source on the glove, and from those five points the software calculates every joint angle and maps the motion onto a robot hand with different proportions. Fine-tuning a robot today usually means teleoperation, by his account, with an expensive robot in the loop, maintenance and a trained pilot, and Maarten would rather record a mechanic with ten years on the job doing the actual work. Video from the data farms of China and India helps with pre-training and carries no kinematics or sense of touch, so MANUS is hiring engineers to build tactile sensors that do not exist yet. Its buyers range from humanoid makers to AI labs such as Google DeepMind, and in his example a carmaker like BMW would train a general purpose humanoid on its own assembly line and keep that data in house.

"We need to re-envision what it means to have a job."

Maarten says MANUS ranks its product against competitors on data quality and comfort. He has tested the top five Chinese gloves he sees and says none hit MANUS's build quality or the comfort needed for eight-hour shifts, though some have added pressure sensing. He says MANUS did get EIC funding from Europe, and he wants European investors to be more gutsy, because in his view Hugging Face is one more company Europe could not keep.

Rogier and Lei cover: why the move from entertainment to robots felt like the same motion capture work on Linux, why MANUS will leave data farms to others, whether AgiBot could build its own gloves, who should own the data a worker's hands produce, and how being mistaken for Manus AI sent a million views to the MANUS website overnight.

Chapters:
00:00 From VR to entertainment to humanoids
02:27 Tactile data and how the glove works
05:19 Who buys the gloves
06:19 Data farms and costly robot pilots
08:35 From glove maker to data provider
12:44 Ahead of cheaper Chinese gloves
15:28 Robot Games gold and AgiBot's own hands
17:02 Workers selling their own hand data
19:35 Who owns the data and the jobs
21:49 Mistaken for Manus AI
23:20 Why Europe can't keep its companies

Watch the video version on YouTube: https://www.youtube.com/watch?v=FQjOhyZ27mY

Follow MANUS
Company website: https://www.manus-meta.com
YouTube: https://www.youtube.com/@ManusMeta

Follow The Robot Bureau
YouTube: https://www.youtube.com/@therobotbureau
Spotify: https://open.spotify.com/show/19Rva9230EwGErD2fg0tQb
Apple Podcasts: https://podcasts.apple.com/podcast/id6812673688
Substack: https://substack.com/@therobotbureau
LinkedIn: https://www.linkedin.com/company/the-robot-bureau
X: https://x.com/therobotbureau
Host Lei Li: https://www.linkedin.com/in/leilibeyondthenow/
Host Rogier Bikker: https://www.linkedin.com/in/rogierbikker/

Teaching a robot an assembly task used to take weeks of programming, and Martin Willers says vision-language-action models have cut that to between two and seven hours, with seven minutes in sight by next year. His company, 120ft Factory, puts that robotic assembly into a 12 metre shipping container that can stand next to Europe's small manufacturers, and one of its first pilot jobs in Stockholm is a christening gift for the Swedish design brand Klong, which asked 120ft to engrave the baby's name on each one.

In this interview from The Robot Bureau, Rogier and Lei sit down with Martin Willers, co-founder and CEO of 120ft Factory, at the Humanoid Robots Summit Europe in Stuttgart in September.

Martin takes his model from the PC. In his telling, the PC put an expensive graphics card from NVIDIA into a cheap standardized metal box so more people could access the technology, and he credits a government grant that made PCs less taxed for the tech scene that runs from Spotify to Lovable. When he visits a manufacturing site he looks at how many standard operating procedures it has, and for him 12 metres and 12 SOPs clicked, so a shipping container could hold those assembly steps and be built to demand, where a factory typically takes two to three years to plan. He says they do everything in simulation first, where a robot learns from a million variations of a task, and getting a new customer's assets from real to sim quickly is a bottleneck when a new customer may arrive daily. He expected the OpenUSD format to set off a Cambrian explosion, but says it is not quite there yet. He hopes the model can scale to something like Amazon Web Services, and 120ft is building a permit for 50 lines next to each other on about two hectares, where trucks bring components to one end and pick up finished goods at the other.

"We have this blind spot where how we make things is not that attainable to us anymore."

Europe's family-owned manufacturers, many of them in Germany and Sweden, struggle to find people for repetitive work, and few can afford to buy and program a robot of their own. Martin says that through his work with Kickstarter he helped many small hardware companies get off the ground, usually with first production runs in China that typically run to thousands of units, paid for with the first money from a customer. A container that can make a couple of hundred units close to home would get them to market sooner, and by his count the million or so small European firms stuck in the old system could become several million within a few years. He also wants every product to leave with a digital product passport and its own serial number, so it can be taken back and remanufactured.

Rogier and Lei cover: why Christmas editions and Instagram launches are the way into small brands, why robots in a closed box are free to dance, whether customers will ever get a live stream from inside the container, why being the Dell of the PC is more attractive to Martin than the 3D printing comparison, and the weird robot babies he once expected simulation to produce.

Chapters:
00:00 From weeks of programming to hours
00:40 Robot assembly in a shipping container
01:58 Klong and the engraved christening gift
02:34 Sweden's Home PC and the 12 metre box
04:13 Simulation first and the new bottleneck
06:16 Europe's missing factory workers
07:15 A karate dojo for robots
08:42 A magic box and circular products
09:53 A million small firms and the 3D printing hype
11:42 Fifty lines on two hectares
13:24 Reindustrialising the West

Watch the video version on YouTube: https://www.youtube.com/watch?v=hqb_OqnYahA

Follow 120ft Factory
Company website: https://120ftfactory.com
Martin Willers on LinkedIn: https://www.linkedin.com/in/willers/

Follow The Robot Bureau
YouTube: https://www.youtube.com/@therobotbureau
Spotify: https://open.spotify.com/show/19Rva9230EwGErD2fg0tQb
Apple Podcasts: https://podcasts.apple.com/podcast/id6812673688
Substack: https://substack.com/@therobotbureau
LinkedIn: https://www.linkedin.com/company/the-robot-bureau
X: https://x.com/therobotbureau
Host Lei Li: https://www.linkedin.com/in/leilibeyondthenow/
Host Rogier Bikker: https://www.linkedin.com/in/rogierbikker/

Xsens has been building motion sensors for 26 years, and by Dennis Kloppenburg's account the largest corporations now put people in motion capture suits, force gloves and cameras for whole shifts of ordinary work and record everything they can to train robots. He tells factory owners to think hard before anyone films their floor, because the way they make things is the data their competitors would most like to have. The first time he walked into a robot factory in China, he watched robots run, climb stairs and cross rubble at a scale he had never seen.

In this interview from The Robot Bureau, Rogier and Lei sit down with Dennis Kloppenburg, Innovation, Growth & Business Manager at Xsens, at the Humanoid Robots Summit Europe in Stuttgart in September.

Each Xsens suit carries inertial measurement units, small modules with an accelerometer, a gyroscope and a magnetometer, and the company's sensor fusion and biomechanical models turn their readings into a moving human body, with magnetic immunity for factory floors full of metal. A camera loses track when the light changes or a box blocks its view, and it misses the elbow and shoulder rotation a worker uses to put something down without scratching it, so Dennis sees the suit as the ground truth that tells a video model which of its footage is right. The first buyers were humanoid makers who needed a way to steer their robots, data farms came next, and he expects integrators working for logistics companies and manufacturers to follow, because the last five percent of each task has to be recorded on site. Chinese factories can mass-produce good electronics cheaply, he says, so Xsens puts its money into the algorithms and licenses its MVN software, which game developers already use to stream a suit straight into Unreal Engine.

"We need to digitize ourselves into the computer."

Dennis counts three or four companies at the front of physical AI data collection, all of them university spin-offs that most Dutch people have never heard of, and he says people around the world use Xsens technology without knowing where it comes from. Most of the investment still goes to the US and China, by his account, while Europe gives Xsens a large customer base, and the company is Dutch at its core with an American owner. His advice to plant owners, many of whom he says are short of people for the night shift, is to keep control of their recordings and think about vendor lock-in before an American or Chinese company films how they work.

Rogier and Lei cover: whether cheaper synthetic data can stand in for motion capture, when gloves and a head camera are enough and when a worker needs the full suit, how Xsens keeps up with demand while component prices rise, whether Xsens is turning into a data company, and the suits worn by Messi and Rico Verhoeven that Xsens still keeps in its office.

Chapters:
00:00 26 years of motion sensors
01:39 Motion capture versus synthetic data
03:37 What a suit sees that cameras miss
06:07 Collect as much as you can
08:58 From humanoid makers to data farms
11:24 Where the money and customers are
14:06 Sensor fusion as the moat against China
15:27 One scalable data collection system
16:33 Who owns the data
19:47 Messi's suit and China's robot factories
21:41 Dutch at its core

Watch the video version on YouTube: https://www.youtube.com/watch?v=-IGXbcUPO94

Follow Xsens
Company website: https://www.xsens.com
YouTube: https://www.youtube.com/@XsensMotionCapture

Follow The Robot Bureau
YouTube: https://www.youtube.com/@therobotbureau
Spotify: https://open.spotify.com/show/19Rva9230EwGErD2fg0tQb
Apple Podcasts: https://podcasts.apple.com/podcast/id6812673688
Substack: https://substack.com/@therobotbureau
LinkedIn: https://www.linkedin.com/company/the-robot-bureau
X: https://x.com/therobotbureau
Host Lei Li: https://www.linkedin.com/in/leilibeyondthenow/
Host Rogier Bikker: https://www.linkedin.com/in/rogierbikker/

By Werner Kraus's count, Fraunhofer IPA ran 66 tests on a Unitree G1 and found a humanoid that is hard to knock over and could work in a cleanroom, with a shoulder joint that overheats when it holds a box and a Bluetooth link his team used to hack it. One robot hand his team tested lasted 100 hours where production needs 10,000, and the day before this interview his team unboxed an AgiBot G2 in the morning and had it putting a bottle into a bin by noon.

In this interview from The Robot Bureau, Rogier and Lei sit down with Werner Kraus, head of the Automation and Robotics research division at Fraunhofer IPA, at the Humanoid Robots Summit Europe in Stuttgart.

Werner's case for deploying now starts with what he saw in China, where by the count in his talk only a handful of the 150 to 200 Chinese robotics companies own the component stack, the way one factory can make dishwashers for Bosch and Siemens, and that is how a new Chinese humanoid can appear within a few months. He visited 14 Chinese companies with a delegation of 50 people from Baden-Württemberg, saw German machine tools cutting gears there and came home convinced Europe can build its own drives and actuators. Until European robots are ready he wants factories to learn on the Chinese machines already on the market and switch platforms if the data risks prove unsolvable. He wants robots that last 10,000 or 15,000 hours rather than 5,000, since 5,000 hours is used up after one year of three-shift production.

"We can do it, but we have to do it. Otherwise others will take the market."

Werner doubts European companies can win the mass market, and he sees their chance in high-value robots for places like critical infrastructure, where buyers will want a European machine. Shortly before the interview he was at the European Parliament with 15 robots built in Europe. What Europe lacks, in his view, is a moment like COVID, when Fraunhofer went from booking a web conference a day ahead to using Microsoft Teams within two weeks, and a short list of the first ten jobs where only a humanoid will do.

Rogier and Lei cover: which humanoids Fraunhofer IPA has tested and why some results stay confidential, why neutral testing matters, with an anecdote about KUKA, Stäubli and a shared gearbox, why testing a robot's AI model is harder than testing its body, why Werner expects humanoids on factory floors within two years, and the three Universal Robots arms his team put into three-shift production in 2011, which all broke within nine months.

Chapters:
00:00 Benchmarking humanoids for the factory floor
01:32 What the Unitree G1 tests showed
02:23 Cleanroom tests and a Bluetooth hack
05:07 Use Chinese robots now, switch later
07:29 The handful of firms behind China's humanoids
10:17 Europe needs its COVID moment
12:55 Neura Robotics and neutral testing
16:24 The Humanoid Capabilities Navigator
20:24 An AgiBot G2 skilled in four hours
24:44 Humanoids in factories within two years
29:08 Ten thousand hours and the 2011 lesson

Watch the video version on YouTube: https://www.youtube.com/watch?v=tyn7cRG8TV8

Follow Fraunhofer IPA
Company website: https://www.ipa.fraunhofer.de/en.html
YouTube: https://www.youtube.com/user/FraunhoferIPA

Follow The Robot Bureau
YouTube: https://www.youtube.com/@therobotbureau
Spotify: https://open.spotify.com/show/19Rva9230EwGErD2fg0tQb
Apple Podcasts: https://podcasts.apple.com/podcast/id6812673688
Substack: https://substack.com/@therobotbureau
LinkedIn: https://www.linkedin.com/company/the-robot-bureau
X: https://x.com/therobotbureau
Host Lei Li: https://www.linkedin.com/in/leilibeyondthenow/
Host Rogier Bikker: https://www.linkedin.com/in/rogierbikker/

Staff costs in restaurants have climbed from 30% to 40% of revenue since before Covid, and two robots now say they can help. One folds about 1,600 table-ready napkins a day for Din Tai Fung, the other makes ice cream in 55 steps inside a Dairy Queen store in China. Maurice Boumen, Director of Rooms at a three-Key Amsterdam hotel, says 'in!' to one and 'out!' to the other.

Dyna Robotics, a Californian start-up founded in 2024, put a two-arm robot into a pilot with Din Tai Fung, which asked for at least 1,500 napkins a day, one every 43 seconds. In August Dyna announced that its robots had passed the chain's return on investment test. Sharpa, known for its dexterous Wave hand, runs a humanoid in a Dairy Queen store that was not rebuilt for it. This episode explains where robots in hospitality already make business sense and where they do not.

In this episode from The Robot Bureau, Rogier and Lei start a monthly series on one industry at a time, beginning with hospitality, with Maurice Boumen of Hotel De L'Europe Amsterdam joining by video call. It was recorded at NIO House Amsterdam (thank you NIO for hosting us!)

By Dyna's count its latest robot folds about 95 napkins an hour, roughly 1,600 over an 18-hour day, where the same hours take three people on eight-hour shifts and about 130,000 euros a year per restaurant at the Dutch minimum wage, before other employer costs. The Dairy Queen humanoid costs an estimated 100,000 to 125,000 euros according to Rogier, and it takes five to six minutes per cup where a person takes two to three.

"Absolutely in." Maurice on the napkin robot, and then a firm "Out" on the ice cream robot.

Rogier and Lei cover: why staff costs reached 40% of restaurant revenue, why Dyna started with napkins, why the Dairy Queen robot is not a vending machine, what a humanoid costs against a human team, how Maurice answered the three either-or questions in the lightning round, and how long it takes Lei to fold a napkin by hand.

Chapters:
00:00 In or out, the trailer
00:47 Welcome and the idea behind the series
02:04 Why hospitality, with staff costs at 40% of revenue
03:28 Meet Maurice Boumen of Hotel De L'Europe
03:44 What luxury hospitality means today
06:11 30,000 vacancies in the Netherlands
07:12 Where hotels look for technology
08:05 Story 1, Dyna Robotics and Din Tai Fung
08:30 The 36 second napkin test
12:04 1,600 napkins a day
14:34 The 130,000 euro staffing sum
18:17 Maurice's verdict on story 1
19:45 Story 2, Sharpa and Dairy Queen
21:47 The 55 steps behind one ice cream
22:35 Why this is not a vending machine
24:36 What the robot costs against a human team
27:43 A robot that fails every hour
28:25 Maurice's verdict on story 2
29:36 Lightning round

Watch the video version on YouTube: https://youtu.be/LS7ttnspJ-c

Follow The Robot Bureau
YouTube: https://www.youtube.com/@therobotbureau
Spotify: https://open.spotify.com/show/19Rva9230EwGErD2fg0tQb
Apple Podcasts: https://podcasts.apple.com/podcast/id6812673688
Substack: https://substack.com/@therobotbureau
LinkedIn: https://www.linkedin.com/company/the-robot-bureau
X: https://x.com/therobotbureau
Host Lei Li: https://www.linkedin.com/in/leilibeyondthenow/
Host Rogier Bikker: https://www.linkedin.com/in/rogierbikker/