Hello! It’s Becca Szkutak filling in for Brian while he takes a long overdue vacation.
For those that don’t know me, I’ve been contributing to A3 since April. I live in Queens, New York with my husband Kyle and Siberian husky Patrice and have been writing about tech big and small for nearly a decade. When I’m not writing, I’m likely hanging out at my other job, teaching Pilates.
You might think robots and Pilates couldn’t be further from one another, and that would be a very reasonable thought, but the studio is where I find most of my day-to-day interactions with robots. I spend quite a lot of time shepherding our robot vacuum back to its base and rescuing it from choking on the bathroom rugs.
But I digress. While Brian is off touching grass and enjoying what every writer does to relax and get away from work, writing a book proposal, we’ve got robotics news to talk about.
Today we are digging into a massive funding round for Neolix, the Chinese company making delivery robots and beyond. We learn that an Italian startup is building Europe's biggest humanoid training facility. And we meet a startup coming out of stealth that looks to automate the chemical-based product industry.
We also have a conversation with Dr. Jan Zizka, the CEO of Brightpick, our weekly roundup of notable startup raises, industry job postings, and more.
One last thing before we dive in, I’m back next week too, so send me all of your funding rounds, pitches, or grumblings for next week and beyond to rszkutak@automate.org.
Neolix, which builds autonomous delivery vehicles, just raised a massive $400 million Series E funding round led by Boyu Capital with participation from Alibaba’s AEF NextGen Fund, Monolith Management, and Granite Asia Capital, among others.
The company has now raised more than $1.4 billion since its 2018 founding but it seems it has been put to good use. Neolix has deployed more than 30,000 of its robots across 30 cities and 20 countries which means it technically operates the world’s largest autonomous fleet, according to reporting from Bloomberg.
It has aggressive expansion plans too. Neolix’s founder and CEO Enyuan Yu has said the company wants to add another 100,000 robots to the road in 2027.
Neolix currently has two models, the Xolra 3 which is the smaller of the two and designed for last-mile delivery, and the Xolra 6 which can hold up to 2,000 kilograms, or 4,400 pounds for us stateside, designed for urban delivery.
The company is also working on a third model, Xolra 20, designed for scalable long-haul automation, but we don’t have too many details on that quite yet.
What’s really notable here is the scale. Getting the robots to do what they are supposed to do safely and consistently is a challenge enough of itself, and of course there are regulatory hurdles to clear to have these systems out on the streets.
To compare, Serve Robotics, known as one of the industry leading players in the category in the United States had a goal of deploying 2,000 of its robots by the end of 2025. That is not to say Serve Robotics is slow or behind but rather to show the sheer magnitude of what Neolix has been able to do thus far.
Likely some of that momentum has come from more autonomous-vehicle friendly regulations in countries, especially in Asia, and especially outside of the U.S. where regulation is an everchanging patchwork system.
“We’re doing work people don’t want to do, are unwilling to do, or can’t do,” Yu said in a recent Bloomberg interview. Continue reading. >
Robotic training continues to dominate the conversation in the industry right now for good reason. Companies are feeling pressured to implement automation as physical AI continues to take over but systems need to be trained to be effective and safe.
We are seeing a surge of startups focused on virtual and simulated training using things like world AI models, which while an important step in the process, can't replace physical training — at least it won’t any time soon.
Italian humanoid company Generative Bionics announced it is building a network of three physical AI data factories, or training facilities, for humanoids. The first of the three will be 5,000 square meters, based in Italy and operational by the end of Q4 2027, Generative Bionics said. This would be Europe’s largest facility of its kind.
The other two will include one somewhere else in Europe, the company hasn’t disclosed where quite yet, and another designed to serve the global market.
The facility will be powered by AMD Helios rack infrastructure. Generative Bionics will work with partner companies to create test scenarios for humanoids to learn new tasks and skills in a controlled environment.
“We are replicating the world, physically and digitally, to train humanoid robots at scale before they ever enter it,” Daniele Pucci, the CEO of Generative Bionics said in a press release. “The Physical AI Data Factory is how we do it: fleets of GENE.01 learn from real industrial work, we stress-test those skills across thousands of digital-twin variations on AMD compute, and we return validated World Action Models to the field.”
They are not the only company pursuing this strategy. German robotics startup NEURA Robotics announced a fleet of robotic gyms earlier this year designed for both their company, and beyond to test and train their humanoids in real world scenarios to collect data.
It will be interesting to see what happens once these facilities come fully online. Will we see a robot-training-facility-as-a-service industry bloom? If so, I hope we will find a better acronym than RTFAAS. Continue reading. >
Brightpick CEO Dr. Jan Zizka discusses the tension between engineering quality and startup speed, and why robotics companies need to balance both. Watch on YouTube >
Jagdeep Singh, (Rhoda AI)believes the industry needs a much larger and more diverse source of training data to build robots that can handle the unpredictability of real work. At Rhoda AI, that starts with internet video.
Andy Zhang (ABB) - What's an industrial giant like ABB to do amid a humanoid hype bubble? Lots. While general purpose nabs headlines, industrial arms and cobots dominate deployment.
Karen Panetta (Tufts University) - From helping create the digital twin to solar-powered race cars, Karen Panetta has lived far more life than can fit into a single podcast. We still give it the old college try.
Automated contributor Rebecca Szkutak (hey that's me!) rounds up the most consequential recent funding rounds in robotics, automation, and physical AI.
120ft, €576,000, Angel round
Dream Labs, $90 million, Seed round
Eka Robotics, $145 million, Series A
Origin, $20 million, Undisclosed
ZAPs Surgical Systems [pictured], $36 million, Undisclosed
We often talk about building dexterous robotic hands that can mimic a human’s for tasks that require precision or a light touch. But sometimes, you just need a good, strong grip to get the job done. There is no shortage of companies building robots that can pick up boxes and move them but that doesn’t always work when boxes are open or are less rigid.
OnRobot announced a new solution to pick up those kinds of boxes called the 2FGP20 Fork, which offers an electric gripper that operates almost like a forklift in sliding under boxes and packages before using an adjustable top clamp to help keep them in place.
Chemical-based product manufacturing is rigid and forces brands and companies to order large batches of materials that often go unused and wasted. Formulate Robotics CEO Osmaan Shah said that it’s a “huge challenge to find manufacturers that will do sub 10,000 unit orders”. His company designed a platform that uses AI and robots for small industrial batching from just an AI prompt.
“We've talked about dark factories, automated factories I feel like for 30, 40, years,” Shah said. “The naysayer would say, well, ‘we've been talking about this for 30, 40, years, why hasn't it happened?’ Technically, the software and the robotics or hardware existed 30 years ago or 20 years ago at least to do this. Why did it not happen then? Why can it happen now? And I would say, it is squarely in what AI is unlocking from a software standpoint.”
Tuesday Labs, which was founded by former Boston Dynamics staff robotics engineer Ros Shinkle, announced its first robot this week and opened up pre-orders. Voxel is a bi-pedal robot with two arms that the company describes as a gentleman in training or a droid butler. For a cool $20,000 the system can help pick up your home and help you find and retrieve items.
If you’ve been reading my weekly startup funding roundups — which you should be by the way — you’d know that the construction robotics sector is not only well populated with founders but also very hot with investors. It’s good to see university researchers diving in here too, experimenting with what works without the constraints of having to make returns for said investors.
Liam has a fun look at two University of Stutgartt projects wanting to do just that. One is focused on building prefabrication timber modules with the goal of improving the manufacturing of this key building material. The other one is particularly interesting to me. It’s looking at building lightweight structures with a swarm of small robots. The concept of hive-minded systems is becoming more common in the defense space, so I’m curious to see how it could play out in this sector too.
The Association for Advancing Automation (A3) is North America’s largest automation trade association representing more than 1,500 organizations involved in robotics, artificial intelligence, machine vision & imaging, motion control & motors, and related automation technologies.
.png?upscale=true&width=1120&upscale=true&name=2FGP20%20Fork%20Photo%20-%20Media%20Kit%20%20(4).png)
