Edge Executive Interview – Jon Lindén, Ekkono Solutions
In the lead up to Edge Computing World, we’re taking some time to speak to key Executives from the leading companies supporting the show. Today we’re talking to Jon Lindén, CEO and Co-founder of Ekkono Solutions.
Tell us a bit about yourself – what led you to get involved in the edge computing market and Ekkono Solutions?
JL: I’m a serial tech entrepreneur. My latest company prior to Ekkono was in telecom. I traveled the world to meet with telecom operators, and for the last four years they spoke of nothing other than IoT – the opportunity and threat of everything being connected with sensors that constantly stream data. This was in the back of my mind when I did a project for the University of Borås on commercialization of research after leaving that company, and met my co-founder Rikard König who after seven years of research had successfully implemented a full machine learning stack on a connected device. I saw the potential of running the first level of real-time intelligence at the edge – though Edge was still not an expression at the time… This has turned into our Google-size of an opportunity as we’ve designed it as a toolbox to help expedite the use of machine learning for smart features on IoT products in all industries. [For the full story, see https://www.ekkono.ai/the-ekkono-story/.
What is it you & your company are uniquely bringing to the edge market?
JL: Technically we bring a unique solution to market that can scale down to run on very small devices (Cortex-M0 size of MCUs) – AND that actually does machine learning at the edge – learning, not just inference. We do incremental online learning. This means that for every instance of data we can train the model, which enables personalized training that learns normal behavior, use, load and application for every individual device. The benefits of this are extensive – besides automated personalized features, you get instant insights without waiting while collecting data for a year, it can adapt to new conditions like when being moved or replacing a wear part, you don’t have to send sensitive raw data to the cloud, and you don’t have to tell constant conditions to the model.
I don’t think that most people have understood the true potential of IoT yet. It’s NOT about collecting data. Instead, it’s about extending the relationship with your products to after they leave the factory. Now, you can see how every product is used, how it behaves, help customers run them in a sustainable way, avoid fatal wear through condition-based maintenance, upgrade them in the field, and even sell your products as a service. This is all enabled by mass-customization and smart automation. And automation in IoT spells machine learning – Edge Machine Learning!
Tell us more about the company, what’s your advantages compared to others on the market, who is involved, and what are your major milestones so far?
JL: We have a huge technical edge in our unique product, which is the result of seven years of university research that helped us touch the ground running when we started Ekkono in 2016. By then we married academic results and sobriety with successful entrepreneurship since three of us founders have brought companies to Silicon Valley, IPO on NYSE and NASDAQ, and to more than $100M in global sales. As a Swedish company we have the privileged access to globally renowned product OEMs like ABB, Volvo, Alfa Laval, Siemens Energy and Husqvarna – who all are Ekkono customers – within driving distance. This is a great foundation for proving the maturity of our product and as reference for customers elsewhere to take the leap.
How do you see the edge market developing over the next few years?
JL: We ain’t seen nothing yet! With IoT it makes perfect sense to combine edge and cloud. Not only due to available processing capacity at the edge and connectivity, but also to solve the question of integrity and to integrate different components to make the bigger product even better – like the ball bearing helping the motor make smart decisions on how to make the machine run in a better way. This is no longer a question of if, but only when. Everything gets connected, and Edge Machine Learning will become as commoditized as connectivity to make products self-learning, predictive and smart. Our vision is the be the default option for developing smart features onboard connected devices. The when has been delayed by the pandemic, but product OEMs will soon catch up on the innovation backlog. All that’s needed are a few front-runners that lead the way for others to follow, because the business prospect of smart IoT is huge!
What are the main trends you see about integrating edge computing into different verticals?
JL: You need an immediate problem to solve to justify investment decisions. This is today typically related to maintenance. But we’re seeing that the sustainability aspect of running every individual unit as energy-efficient as possible has become a solid reason in itself. I’m so happy about this as it brings even more meaning to what we do. A clear trend is that the use cases become more realistic and focused on real problems. Machine Learning and AI is not seen as “one big brain” anymore. This is great as first features like virtual sensors and individual health indicators enable future grand features like self-optimizing and autonomous machines one step after another. This speaks in favor of our tools-approach where we empower our customers to continuously develop new smart functionality on their products. We don’t necessarily see that the maturity between verticals are that different, but it’s more a question of the maturity of individual companies. This will prove who are winners and losers as you either disrupt yourself or get disrupted by someone else.