Ottawa supply chain software provider Kinaxis Inc. is beefing up its machine learning and artificial intelligence capabilities by buying Toronto startup Rubikloud Technologies Inc. for $60-million in cash.
It’s the second acquisition this year by one of Canada’s most valuable software companies and comes “at a time where we have never been more relevant in terms of what we do,” Kinaxis chief executive John Sicard said. The company in February bought Indian supply chain consultancy and partner Prana Consulting.
Kinaxis’s stock has surged by more than 80 per cent this year as manufacturing giants including Ford Motor Co., Lockheed Martin Corp., Unilever NV and Dyson Ltd. relied on its software to get a handle on supply chains amid pandemic-related disruptions.
Chief marketing officer Jay Muelhoefer said in March that customer usage in the first quarter was up 33 per cent from the fourth quarter last year. Kinaxis was one of the rare companies to keep its prepandemic financial guidance for 2020 intact – forecasting revenue of US$211-million to US$215-million, up from US$191.5-million in 2019, and an adjusted operating profit margin of 20 per cent to 23 per cent – as others withdrew forecasts.
In buying Rubikloud, Kinaxis will increase its 800-person ranks by 10 per cent, including about 30 machine learning (ML) and artificial intelligence (AI) specialists. Boosting its ML and AI capabilities has been a focus recently as Kinaxis worked to develop software to help customers automatically “heal” their supply chains in response to unexpected issues. “This moves us ahead significantly,” Kinaxis chairman Ian Giffen said in an interview.
The deal also gives Kinaxis entry into the retail sector, where Rubikloud has specialized. Kinaxis previously focused on clients in aerospace and defence, automotive, consumer products, high-tech electronics, industrial and life sciences.
“We look at this as a very strategic, technological acquisition,” Mr. Sicard said. “They are an entry point to a market vertical that we’re not in. [Rubikloud has] solved some use cases that we simply do not have.”
Rubikloud was one of a slew of AI startups to emerge from Canada in the mid-2010s, fuelled by global interest in breakthroughs in self-teaching algorithms largely pioneered in Canada. Rubikloud promised to use AI to help large merchants automate complex decisions, enabling them to run simulations to improve the accuracy of forecasting, ensure merchandise was stocked and allocated as needed and to better model the impact of promotion plans.
The startup secured US$39-million from foreign billionaires Li Ka-shing and Len Blavatnik and Intel Capital, as well as Canada’s iNovia Capital and First Ascent Ventures. It also secured multimillion-dollar contracts with a handful of retail giants in the United States, Canada and Britain.
But Rubikloud CEO and co-founder Kerry Liu said in an interview his company reached a point this year at which it had to decide whether to raise significant funds to build a global sales force or partner with a company that could provide that. As Rubikloud explored its options, it kept running into Kinaxis in bidding processes, as they competed for different supply chain management tasks for consumer packaged goods giants. The companies discussed partnering on bids; that evolved into takeover talks.
Mr. Liu, who will become executive vice-president of strategic innovation with Kinaxis, said selling was “bittersweet,” but added “being acquired by a Canadian homegrown organization ... was a very important element. … I am very committed and excited … [to be] part of the scale-up of a multibillion-dollar Canadian company.”
The deal is the latest in a string of takeovers of Canadian AI startups for relatively modest sums compared with the ambitions of entrepreneurs and funders who joined the AI gold rush. Those include Layer 6 Inc., which sold to Toronto-Dominion Bank for US$100-million-plus in 2018, and Dessa, which sold to Square, Inc. in February for US$36.6-million.
Steve Irvine, CEO of Toronto AI startup Integrate.ai Inc., said AI companies have struggled to scale up because the software they developed to help clients root out money-making opportunities hidden in their data requires “a lot of bespoke modelling work to build, tune, monitor and retrain” their algorithms, while extensive “data cleaning” is needed to make their customers’ information usable.
AI companies have also been challenged in attempts to reuse algorithms trained on data of certain customers on others. “This makes it labour intensive and more linear than exponential in its growth potential,” he said.
Your time is valuable. Have the Top Business Headlines newsletter conveniently delivered to your inbox in the morning or evening. Sign up today.