Factories and warehouses have been two of the largest markets for robots in recent years, with machines that take mundane, though limited, processes to accelerate work and free humans to perform other more complex tasks. Now, a startup in Poland that is expanding the reach of what these robots can do is announcing financing, a sign not only of how robotic technology has evolved but of the growing demand for more automation, specifically in the world of logistics and compliance.

Nomagic, which has developed a way for a robotic arm to identify an item from a messy selection, pick it up and then package it in a box, announces today that it has raised $ 8.6 million in funds, one of the largest seed rounds in history of Polish startup. Co-directed by Khosla Ventures and Hoxton Ventures, the round also included the participation of DN Capital, Capnamic Ventures and Manta Ray, all previous Nomagic sponsors.

There are a number of robotic arms on the market today that can be programmed to pick up and deposit items from point A to point B. But we are only beginning to see a new wave of companies focused on bringing them to compliance environments due to limitations of those arms: they can only work when the items are already “sorted” in a predictable way, such as in an assembly line, which means that the fulfillment of, for example, online orders is usually made by humans.

Nomagic has incorporated a new degree of computer vision, machine learning and other technologies based on artificial intelligence to raise the capabilities of these robotic arms. Robots powered by their technology can successfully select elements from a group of “unstructured” objects, that is, not an assembly line, but potentially another box, before picking it up and placing it elsewhere.

Kacper Nowicki, the former CEO of Googler de Nomagic who co-founded the company with Marek Cygan (an academic) and Tristan d’Orgeval (formerly of the Climate Corporation), noted that while working on the problem of objects has not Structures and Industrial Robots: in the US In the US, there are some live implementations that are taking shape, with one, Covariant, which recently came out of stealth mode; It has been primarily a “missing piece” in terms of the innovation that has been made to make logistics and compliance more efficient.

“If each product were a tablet or a phone, it could automate a regular robotic arm to pick up and pack,” Nowicki said. “But if you have something else, say something in plastic, or a great diversity of products, then that’s where the problems come in.”

Nowicki was a long-time Googler who moved from Silicon Valley to Poland to build the company’s first engineering team in the country. In his years at Google, Nowicki worked in areas such as Google Cloud and search, but he also saw the developments of AI in progress at Google’s DeepMind subsidiary and decided that he wanted to address a new problem for his next challenge.

His interest underlines what has been a kind of bifurcation in artificial intelligence in recent years. While some of the first implementations of AI principles were applied to robots, these days a lot of robotic hardware seems outdated and even outdated, while much of the AI ​​approach has shifted to software and software. “non-physical” systems intended to replicate and improve human thinking. Even the word “robot” is now so likely to be seen in the phrase “automation of robotic processes”, which in fact has nothing to do with physical robots, but with software.

“Many applications of AI are not as attractive,” Nowicki said (in fact, although Nowicki did not explain it, DeepMind in particular has faced a lot of controversy over his own work in areas such as medical care). “But I was struck by improvements in existing robotic systems by applying machine learning and computer vision so they can operate in unstructured environments. There has really been very little automation in physical systems, and I think it is a place where we will still see many changes. “

Interestingly, although the company is focusing on hardware, it is not actually building hardware per se, but is working on software that can run on the most popular robotic arms on the market today to make them “smarter.”

“We believe that most of the intellectual property in AI is in the software stack, not in the hardware,” said Orgeval. “We see it as a mechatronics problem, but even there, we believe this is primarily a software problem.”

Having Khosla as a sponsor is notable since a large part of VC’s prolific investment has been in North America so far. Nowicki said he had a connection with the company during his time in the Bay Area, where before Google, Vinod Khosla backed his start (which went bankrupt in one of the dotcom crises).

While there is an opportunity for Nomagic to take its idea globally, for now, Khosla is interested because of a closer opportunity at home, where Nomagic is already working with third-party logistics and compliance providers, as well as retailers such as Cdiscount, a French Amazon-style online nut soup market.

“The Nomagic team has made significant progress since its founding in 2017,” says Sven Strohband, Managing Director of Khosla Ventures, in a statement. “There is a great opportunity in the European market for robotics and warehouse automation, and NoMagic is well-positioned to capture part of that market share.”

Leave a Reply