Hamburg The workers in the Otto Group’s goods distribution center in Haldensleben will soon have eight very special new colleagues. These are robots that are designed to take on tasks that previously only humans could do: recognizing, gripping and sorting tens of thousands of different products.
What sounds simple at first glance has so far been largely impossible for robots. They are perfect for fixed, repetitive tasks, such as car assembly. But with constantly changing situations in which dynamic decisions are required, the robots have so far been overwhelmed.
Thanks to new forms of artificial intelligence (AI), this is changing fundamentally. “The technology has made immense progress, especially in image recognition,” says Werner Kraus, department head at the Fraunhofer Institute for Manufacturing Engineering and Automation IPA. Robots could therefore take on much more complex tasks in logistics than before.
In order to keep up, the Otto Group has entered into a cooperation with the start-up Covariant from Silicon Valley, which combines artificial intelligence with robotics. “The use of AI enables us to rethink processes that were previously not possible,” says Otto board member Kay Schiebur in an interview with the Handelsblatt.
In the initial phase, around 120 robots are to gain experience in various warehouses. “Together with Covariant, we have already found up to 300 use cases in which we suspect that robots with AI can bring more efficiency and a higher quality of work,” says Schiebur. This is now being tested in practice.
Otto is doing pioneering work in Germany. Similar systems are already in widespread use in the USA. According to estimates, suppliers of self-learning gripping robots there already generated sales of 100 million euros last year. Covariant robots are also already in use in numerous US warehouses.
Experts see a potentially huge and fast-growing market for gripping robots in warehouse automation. The market researcher Interact Intelligence, which specializes in automation, estimates that sales of such robots could increase to 1.34 billion US dollars by 2025. That would be a tenfold increase compared to 2021.
“This is just the beginning, and we will see increased adoption in the coming years,” predicts Irene Zhang, analyst at Interact Analysis. This will continue to grow as the benefits and innovation potential of robotic picking become more apparent.
Zhang estimates that around 150,000 such robots will be installed by 2030. This development is being driven in particular by the rapid advancement of artificial intelligence.
Covariant has raised $222 million in capital
AI also made the launch of Covariant possible. Covariant founder Peter Chen started his career at OpenAI, the company that has just sparked a global hype about artificial intelligence development with its ChatGTP software. That is why he initially had a general AI developed for his new company, which is not limited to special applications.
“Our core product, Covariant Brain, is self-learning software that we offer our customers as a subscription model,” explains Chen. The companies can then either buy or lease the robots that Covariant procures from robot manufacturers such as Knapp or ABB.
The new robots with artificial intelligence are like Chat GPT, they can solve completely new tasks and react independently to new challenges and changes. Peter Chen, Founder and CEO of Covariant
Earlier in April, Covariant raised an additional $75 million from investors led by venture capital firms Radical Ventures and Index Ventures. The total funding thus increased to 222 million US dollars. Among robotics experts, the company’s technology is considered to be particularly advanced.
“Covariant has demonstrated that it is the global leader in AI robotics, where AI is enabling the successful deployment of robots at scale to secure global supply chains,” said Jordan Jacobs, Managing Partner of Radical Venture.
Despite the experiences already made in the USA, the cooperation with the Otto Group is of fundamental importance for Covariant’s further development. The Otto Group “is the perfect partner for us to expand the possible uses of robots and to develop next-generation logistics centers,” emphasizes the Covariant boss.
It is “the first long-term, strategic partnership in e-commerce in which we want to develop many tailor-made applications for logistics centers,” explains Chen. His calculus: “Our system learns from every collaboration, and the data we collect makes our AI better,” says the Covariant boss. These advances would later also benefit other logistics customers.
“You don’t get such an AI in Europe”
In plain language: Ultimately, Otto’s competitors will benefit from the experience that Covariant is gathering in Germany. Otto consciously accepts this disadvantage. It’s about not wasting time. When it comes to AI, logistics are also in a “race,” says Alexander Sukharevsky, AI expert and senior partner at management consultancy McKinsey. It is less of a problem “getting your hands dirty” than being left behind. Ultimately, the integration of AI systems meant that the processes behind them also adapted to the new world. “Companies can only benefit from this,” says the expert.
On the other hand, Sukharevsky warns against blind and thus inefficient technology euphoria. First of all, the decisive question is whether a process would really be improved by AI. The technology is “not the standard answer”.
From Otto’s point of view, Covariant’s “Brain” was the decisive point for the partnership. “We saw many solutions on the market, but we were completely fascinated by their AI capabilities,” says Schiebur, who is responsible for both logistics and AI applications on the retail company’s board of directors. “You don’t get such an AI in Europe,” he says.
Modern, AI-supported robot solutions can also distinguish very similar objects from one another.
(Photo: Covariant)
At the start, the robots in the Otto warehouses are to take over simple tasks from human employees, such as taking items from boxes and placing them on the sorting conveyor belts with the barcode facing up. The aim is more complex processes such as the complete compilation and packaging of a customer order.
Modern, AI-supported robot solutions can also distinguish very similar objects from one another, for example by reading the text on their labels. They locate them in space and calculate in advance how the robotic arm has to reach for them. By overcoming this complexity, “a new level of development” has been reached, says Fraunhofer researcher Kraus. However, there are still problems with the depositing, which is often quite “clumsy”. Objects would typically just be dropped into the shipping carton.
Zalando relies on the combination of people and technology
Other large online retailers also want to make their logistics processes more efficient through automation. “We are already using artificial intelligence and robots to optimize our logistics processes,” says Marcus Daute, Director Logistics Product & Strategy at Zalando. However, he qualifies that it is about the “combination of people and technology in everyday work”.
This is exemplified by the Toru robots from the supplier Magazino, which Zalando uses in its warehouse in Lahr. They take shoe boxes from the shelves and transport them to the shelves. These robots are also controlled by AI, which is not yet so highly developed.
>> Read also: Seven problems of mankind, seven possible solutions – you should pay attention to these technologies
Zalando manager Daute emphasizes that the AI allows them to learn something new with every outsourcing “and to fall back on their wealth of experience in particularly tricky cases”. But they can only react to changes independently to a limited extent and therefore need the support of employees. For example, they can only work without errors if the shoeboxes are positioned very precisely on the shelves.
Amazon invests 400 million euros in new technologies
Amazon, on the other hand, has largely limited the automation of its logistics centers to conventional transport robots and sorting lines. Even simple gripping robots that lift pallets off conveyor belts, for example, are still the exception. While the retailer uses more than 100,000 transport robots worldwide, it only has 30 of these palletizing robots.
The US group has invested more than 400 million euros in new technologies in the European logistics centers.
(Photo: Amazon)
According to Amazon, within three years it has invested more than 400 million euros in new technologies in its European logistics centers. “The exciting thing is that we are just at the beginning and will continue to develop and use cutting-edge technologies in the years to come,” explains Stefano La Rovere, Director Robotics Advanced Technology at Amazon.
The Otto Group also sees potential in the technology that goes far beyond the 120 robots ordered so far. “The advantage for us is that we don’t have to decide on a large investment right at the beginning, but can proceed sequentially,” explains Schiebur. Every single step must prove its success before it can continue.
Covariant boss Chen is already convinced of the successes. “We’re seeing a significant reduction in operational costs in the warehouses where our robots are deployed,” he says. According to business circles, it should typically be between 15 and 30 percent.
>> Read also: Install electronic parts and stack boxes: SMEs overcome skepticism about robots
The fact that the robots are working more and more independently as a result of the further development of AI has a completely different advantage for Otto. Today, the company still employs up to 3,000 people in highly automated warehouses. “But it is becoming increasingly difficult to find employees for many simple tasks, we are facing a very challenging job market,” explains Board Member Schiebur.
At the same time, the company can use the new AI robots to develop more demanding and thus more attractive work profiles for employees. Appropriate training measures have already been started. “In the future,” says Schiebur, “they might then be able to control and monitor an entire fleet of robots.”
More: Tablet replaces the service technician – How AI is revolutionizing the industry