Where Industry Meets Innovation

  • Contact Us
  • sign in Sign In
  • Sign in with certificate
mit campus


Search News

  • View All
  • ILP News
  • MIT Research News
  • MIT Sloan Management Review
  • Technology Review
  • Startup Exchange

August 2, 2018

Arundo: Pushing data analytics to the edge

After many years leading the global technology practice for asset-heavy industries, MIT alumnus Tor Jakob Ramsøy launched Arundo two and a half years ago to provide large-scale analytics and increased asset utilization.

Steve Calechman

The declining costs of data acquisition, data storage, and cloud computing have transformed the economics of numerous industries, from consumer products to retail to financial services. However, fields such as energy, maritime, chemical, and manufacturing haven’t experienced the benefits at the same rate. These sectors face unique challenges with their capital-intensive physical assets and related information technology investments. MIT alumnus Tor Jakob Ramsøy (S.M. ‘95) understood these dynamics, and launched Arundo two and a half years ago to provide large-scale analytics. The company’s technology enables machine learning for performance insights and automates processes for equipment monitoring, safety, logistics, and scheduling, helping its clients to minimize unplanned downtime and increase profitability, because as Ramsøy says, “It’s all about increased asset utilization.”

Tor Jakob Ramsøy, CEO of Arundo Analytics

Reviving data science
Arundo was founded in 2015; the inspiration came a couple of years earlier. Ramsøy had been a senior partner at McKinsey & Company, leading the global technology practice for asset-heavy industries. In that role, he saw a recurring theme. Big companies with old assets had lots of data, but they weren’t storing or using it in any systematic or accessible way. Even if they tried to use the latest machine learning tools, he says, “A lot of data science was dying in PowerPoint.”

More than merely rescuing information, Ramsøy says that he wanted to apply consumer business algorithms, such as how Netflix makes recommendations based on past viewing, to industrial companies. The intent, he says, was to be dynamic and to anticipate and predict issues in real time, such as equipment or operation failures and points of improvement. Three offices were then opened: Silicon Valley for technology innovation; Oslo, by the North Sea, for shipping and renewables; and Houston for oil and gas. A Boston office was added in January 2018 to take advantage of MIT’s talents and the ILP’s partnerships with large industrial companies that could benefit from Arundo’s capabilities.

While oil and gas was the initial focus – the field basically invented big data 30-40 years ago, Ramsøy says – Arundo’s industry targets share common ground. They have advanced equipment with sensors, making data harvesting and sharing that much easier. There’s also the need to minimize downtime without a compromise in safety. For example, with an oil rig in the North Sea, the average utilization is 84 percent, but the planned utilization is 95. As Ramsøy says, if a company produced 50,000 barrels a day at $45 per barrel, that 11 percent gap could mean hundreds of millions of dollars in lost revenue. With 50 oil rigs in play, it’s a multi-billion dollar opportunity. “The magnitude of the business case is enormous,” he says.

Going to the edge
Arundo’s initial customer was Statoil. For the Norwegian multinational energy company, Arundo built algorithms and developed the ability to not only capture data, but also compare it to performance history and be able to offer predictive maintenance, either with a specific recommendation or even by implementing something like a controlled stop, Ramsøy says.

That’s one part, the ability to pinpoint the cause of downtime, Ramsøy says, and fix it. What sets the company apart is being able to easily introduce machine learning at scale into daily operations, and it’s due to leveraging the cloud, which he says is accessible, safe and inexpensive. Because of the abundance of sensors, Arundo can tap into heavy industrial equipment, stream data, build and train machine learning models, and then publish those models into a business process, with “one click,” he says.

Additionally, Arundo provides “edge” analytics, which involves a few things. The company can sample and intelligently stream data from rugged or remote industrial sites, which may not always have internet connectivity, into the cloud. It can provide edge compute capabilities to compress data with local calculations and stream just the results back to the cloud. And, most uniquely, it can push trained machine learning models down to the edge, enabling them to interact with local operators and decision processes and sync with a cloud-based model management framework when connectivity is available. “So then you have the wisdom of the crowd. All machine learning models are learning from each other,” he says.

All this wouldn’t mean much without speed. For one client, a manufacturing company with 35 plants around the world, Arundo was able to pull streaming machine data from a plant in China in just one day, and deploy real-time cloud analytics in just a couple of weeks, once the right hardware was in place. Ramsøy says that at minimum every customer is guaranteed delivered business value in less than 90 days.

Being ready for changes
Ramsøy says that he’s happy with the state of his company, but he’s not satisfied, as untapped opportunities remain. One is MIT. Arundo recently joined the STEX25, and while the relationship is still taking shape, Ramsøy knows that high expectations and optimism aren’t unrealistic in being part of the ecosystem. What’s often said is true, he adds. MIT offers expertise and talent with its faculty, researchers and students, who can help improve Arundo’s technology, along with being the current and future generations of employees. “We see MIT as a great meeting place between industrial companies, research, and start-ups like ourselves,” he says. “We hope to take an active and leading role in this with our IoT products.”

All that will certainly assist in the other great potential that Ramsøy sees. No industry wants to merely buy a pump anymore. It wants to buy pumping. This industrial shift from product to service economy hinges on readily having usable data. For both a company and its equipment manufacturers, data analytics would let them know exactly what they’re delivering, and, rather than trying to sell an input, businesses could guarantee an outcome. That ability hasn’t existed and it’s one area that Ramsøy says gives Arundo an advantage. “Think of us as the Android of the industrial internet,” he says. “That is what really makes me excited.”

About MIT Startup Exchange, STEX25, and MIT’s Industrial Liaison Program (ILP)
MIT Startup Exchange actively promotes collaboration and partnerships between MIT-connected startups and industry. Qualified startups are those founded and/or led by MIT faculty, staff, or alumni, or are based on MIT-licensed technology. Industry participants are principally members of MIT’s Industrial Liaison Program (ILP).

MIT Startup Exchange maintains a propriety database of over 1,500 MIT-connected startups with roots across MIT departments, labs and centers; it hosts a robust schedule of startup workshops and showcases, and facilitates networking and introductions between startups and corporate executives.

STEX25 is a startup accelerator within MIT Startup Exchange, featuring 25 “industry ready” startups that have proven to be exceptional with early use cases, clients, demos, or partnerships, and are poised for significant growth. STEX25 startups receive promotion, travel, and advisory support, and are prioritized for meetings with ILP’s 230 member companies.

MIT Startup Exchange and ILP are integrated programs of MIT Corporate Relations.