Few industries generate as much data as factories, from sensors and pumps to motors and compressors. Today, Oden Technologies announced it has raised a $28.5 million Series B round, led by Nordstjernan Growth, to help manufacturers use AI to turn all that data into useful recommendations that literally empower machine operators up to date.
The result, said Willem Sundblad, CEO of Oden Fortune, could finally deliver on the long-awaited promise of so-called Industry 4.0 – or the full integration of digital technologies into manufacturing – and, in the future, directly link data insights to autonomous manufacturing processes powered by AI. But right now, the biggest challenges facing the industry are massive labor shortages and high employee turnover, driven by harsh working conditions and strong competition from companies like Amazon. This means manufacturers often need to quickly recruit new, less experienced machine operators, and real-time, industry-specific insights gleaned from billions of complex data points can help them succeed.
“There’s so much data that people can’t really analyze it,” said Sundblad, who co-founded the now New York-based company in 2014 with Peter Brand, now its chief commercial officer. “This could be temperatures, pressures, flow meters, different chemical additives, weights, even the essential engine revolutions per minute and screw speeds. »
Manufacturing data turns into predictive recommendations
Experienced workers take years to learn their machines, gaining what Sundblad calls “fingertip feel,” but new operators often don’t have that luxury. By gleaning AI-driven insights from manufacturing data, they can get predictive recommendations such as how to reduce material or energy costs or increase speed while maintaining quality and managing the constraints defined by the company. For some customers, the results have been significant, with factory line speeds increasing by more than 40 percent, Sundblad said.
Besides increased productivity, Oden’s AI-based recommendations also made work less stressful for machine operators, he added. “There’s an example in the paper industry, where when you’re making things, you can’t know whether what you’re making is good or bad while you’re making it,” he explained. “You take a sample in the lab and find out five hours later if it was good or bad.”
For the operator, if they find out the run was bad, “they’re exposing themselves to a world of pain because it’s very expensive,” Sundblad said. Oden’s real-time AI models, which use historical data combined with supervised machine learning, predict product quality as it is manufactured, “so when the operator makes a change, it sees him instantly, which changes his stress level. »
The future is autonomous systems with humans involved
In the future, Sundblad says, Oden’s AI-based recommendations will help power autonomous systems with humans in the loop: its AI-based insights platform could serve as an intermediary between operators and machines. “I know people who are taking a holistic approach to autonomous manufacturing by building new factories from scratch with autonomy in mind,” he said. “We’re looking at the assets that are already there, layering that layer of intelligence on top of that.”
But right now, he says, it’s all about helping machine operators “so that these jobs are better, that they are more efficient, that they are better paid and that they are safer”. In the future, as autonomous systems become more common across industries, “eventually, these people will perform higher value-added tasks.”
New investors in Oden’s new funding round include Flat Capital, the investment firm of Klarna founder Sebastian Siemiatkowski; Oden client INX International Ink; and Recurring Capital Partners, with participation from almost all existing investors, including Atomico and EQT Businesses. The company’s latest valuation was not disclosed.
Oden, whose clients also include Southwire, Sonoco, Viakable, GreifTeknor Apex and Lake Cable, compete in the manufacturing data analytics space with companies such as Braincube and TwinThread.
