HANNOVER, GERMANY — Scores of Canadian companies hoping to compete in international markets are focusing on how artificial intelligence could speed up and improve advanced manufacturing—without needing to hand over sensitive data to the cloud.

“We live in a world where if you talk to all the CEOs of all the AI companies, we’ll all be out of a job in five years. That’s the doom and gloom side of the story,” said Matthew Lowe, the CEO of ZeroKey. “Most of our technology makes humans more efficient at what they do.”

His Calgary-based firm, which was one of about 100 Canadian exhibitors at the international trade fair Hannover Messe last month hoping to expand their business in Europe, is making a bet on how AI and humans can work together. Using trackers and sensors, ZeroKey aims to minimize errors on factory floors and make it easier for humans to complete complex tasks.

In a demonstration of how the system works, Lowe put tracking devices on his wrists and stood at a mock-up of a workstation, where small bins filled with different machinery parts were stacked on either side of a computer monitor. The system showed him what to do next, with nearby sensors detecting the location of his hands with 1.5 millimetre accuracy.

As Lowe stood at the workstation, a diagram on the monitor showed him which part to pick up from which bin, at which stage in the process. When he got it right, the monitor flashed green. If his hand entered the wrong bin at the wrong time, it flashed red to alert him to an error.

The idea is to catch a mistake in real time—fast enough to correct it in the moment, or at least well before the product would ever reach the consumer. A loose bolt, Lowe explained, can be the difference between a safe ride home and a car seat flying out of place during a sudden stop.

Industry Minister Mélanie Joly has said the world is going through another “industrial revolution” driven by AI. The federal government is hoping to make it easier for Canadian firms to compete, including as part of efforts to diversify exports amid the trade war, but it can be challenging for those companies to stand out against far larger firms in the United States and China. The Canadian firms arrived at Hannover Messe just days before Toronto-based Cohere, which had its own booth at the Canadian pavilion, announced its merger with German AI firm Aleph Alpha. Before heading to Germany, Joly had told The Logic that such partnerships could play a role in helping Canada find a way forward amid U.S. protectionism and powerful hyperscalers.

For ZeroKey, and many of the other Canadian firms at the sprawling industrial tech fair in northern Germany, however, the game isn’t to compete with the hyperscalers. It’s to own the final application of AI. Jensen Huang, president and CEO of Nvidia, thinks of the AI industry as a “five-layer cake.” Energy and chips are at the bottom, followed by infrastructure. Then come the models, such as those developed by OpenAI, Anthropic and Cohere. “At the top are applications, where economic value is created,” Huang wrote in March.

Julien Billot, the CEO of supercluster Scale AI, says success for Canadian firms will likely come from that top layer. “If you want to be sovereign,” he said, “you need to own the final customer application.” That is where ZeroKey comes in, along with many of the other Canadian firms that were at Hannover Messe last month promoting physical AI tools for advanced manufacturing.

One of them was Montreal-based robotics company Vention. Its physical AI software product, known as GRIIP, teaches robots to handle unstructured tasks in a factory using perception and reason, without the need for weeks of training and data collection. CEO Etienne Lacroix, whose company expanded to Germany in 2022, noted that unlike generative AI, most physical AI runs on the edge, outside the cloud. “Manufacturers and clients keep their own data. It’s not shared,” he said.

M2M Tech, based in Surrey, B.C., focuses on helping manufacturers and other companies deploy AI solutions. It has a similar sovereign-focused pitch. Its platform uses digital twins, which let companies test and troubleshoot in a virtual environment before they ever build anything on the factory floor or in a foreign-based cloud. Edouard Romeus, chief product officer, said that’s key for manufacturers that have a lot of proprietary data but cannot afford to avoid using AI to boost their productivity. “These manufacturers have no choice but to innovate, or they go out of business.”

As the Liberal government seeks to double non-U.S. exports by 2035, trade expert Danielle Goldfarb has been encouraging Canadians to think beyond physical goods like beef and gold. “The most valuable parts of trade today are really the data, the algorithms, the intelligence that is either sold on its own or embedded in any product,” Goldfarb said at a trade conference earlier this month. 

Jayson Myers, CEO of Next Generation Manufacturing Canada, or NGen, the industry-led non-profit that organized the Canadian mission to Hannover, has said it can be easier for Canadian companies to enter new markets by fitting into existing supply chains. The results of those efforts, however, are not always easy to see—or easily identifiable as Canadian.

One homegrown success story at Hannover Messe was ATS, based in Cambridge, Ont. Its SuperTrak Conveyance system was on prominent display inside the massive exhibition space for German multinational Siemens, where it is used in an AI-powered visual inspection system. Joris Myny, the senior vice-president of digital industries at Siemens Canada, called it a “marvellous piece of engineering,” but just like many Hollywood stars who grew up in Toronto, visitors from around the world would not immediately recognize the technology as Canadian.

Other Canadian exhibitors that spoke to The Logic at Hannover Messe were feeling optimistic about meetings with Siemens, which employs about 4,600 people in Canada. “We don’t see Siemens as a competitor,” said Romeus. “We see them as a partner.”

ZeroKey, which says it has raised some $17 million in funding since it launched in 2016 and has sold its technology to major manufacturers around the world, including automakers, had one of the more eye-catching demonstration booths at the Canadian pavilion inside the cavernous AI-focused exhibition hall at Hannover Messe. Yet a walk through the trade fair showed a myriad of companies from all over the world promoting physical AI solutions and humanoid robots, including many from China.

“Personally, I’m pro-competition—always have been,” Lowe said. “The certain protectionist measures on trade that would limit China’s entry inevitably lead to lack of innovation.” Another big challenge, Lowe said, is homegrown. Unlike the U.S., Lowe said, Canada has a mid-market funding gap, as well as the strings attached to government funding, that can make it harder for a Canadian startup to get the capital it needs to scale up, or survive. “If any of the dots on that path are broken, that company doesn’t make it to that stage.”