As large corporations figure out what to make of generative systems and other forms of artificial intelligence, Kyndryl is busy installing the plumbing they’ll need to put the technology to work. The IT services firm wants to help customers “turn that science experiment into part of the way they actually run their business,” said CEO Martin Schroeter.  

New York-headquartered Kyndryl manages the data centres and cloud systems on which its giant clients run their technology stacks. The firm spun out of IBM in November 2021. Its low profile—it has never been the subject of a headline in any of the four largest U.S. newspapers—belies its scale. Kyndryl made US$17 billion in revenue in its fiscal year ending March, and employed about 90,000 people.

Canada is one of the firm’s 10 largest markets, with some 2,000 workers based in the country. Clients include major financial-services firms, airlines and government departments, which use Kyndryl to shift data and applications to the cloud and shore up cybersecurity. The company is now targeting industrial customers who want to network their factories and facilities, and businesses in all sectors exploring automation.

Schroeter sees a significant opportunity for Kyndryl in helping prepare clients’ technology infrastructure to implement AI in their operations. “The nature of the work we do is mission-critical,” he said in a late June interview with The Logic from the firm’s Canada headquarters in Markham, Ont. Banks and airlines can’t afford any downtime on their systems. “So the amount of work that has to happen to prepare an infrastructure [for AI] is enormous,” he said.

Kyndryl already helps customers manage the architecture and security of their data, as well as juggle various applications and cloud services. Information-hungry AI naturally layers on top of that infrastructure. For example, Canadian financial-services clients use Kyndryl’s machine learning algorithms for analytics, and to automatically patch potential security vulnerabilities. “In [the] infrastructure world, the AI we use to help our customers run their systems is working … and will keep improving,” said Schroeter.

Then there are large language models and other content-producing systems, which have captivated the tech sector since San Francisco-based OpenAI launched ChatGPT last November. Generative AI is “very much in the science-experiment phase,” according to Schroeter. Some consumer-facing firms are using it to generate buzz. But “nobody invests in enterprise IT because it’s fun or for brand,” he said. “You do it because there’s a payback.” For big businesses to adopt generative AI, its producers will need to demonstrate the financial return.

As viral bots and image makers sweep boardrooms and headlines, tech giants are allying themselves with field-leading startups. Microsoft has invested more than US$10 billion in OpenAI, reselling the smaller firm’s models via its cloud service and adding its technology to products. Oracle has launched a similar integration with Toronto-based Cohere, after joining its US$270-million Series C in June.

Kyndryl isn’t seeking a similar partnership, Schroeter said. Instead, it will work with a customer’s preferred AI tools and cloud services. Post-spinoff, the firm no longer needs to sell IBM’s server and cloud offerings, and now has deals to deploy Amazon Web Services, Google Cloud and Microsoft Azure.

Last September, Kyndryl launched Bridge, a platform for customers to buy and monitor its own and third-party technology integrations and services, including ones which Kyndryl can install and manage for them. (Only Kyndryl products are currently listed in the AI section). Schroeter said Kyndryl can also provide clients with practical insights on adopting the technology from its own experience as a big data company.

“We’ve all gotten on this bandwagon of AI,” said Divya Goyal, an equity research analyst covering technology stocks at Scotiabank, and “Kyndryl has a very dominant role to play.” The firm can help enterprise clients modernize their IT infrastructure and layer in the AI tools and computational power needed to make the most of their data sets. 

Kyndryl doesn’t need an alliance with a model maker in order to win business as companies adopt AI, but its workers must be able to implement the products that the Oracle-Cohere, Microsoft-OpenAI and other partnerships produce, according to Goyal. “Can they actually deploy AI IP?” she said. “That is very important to whether Kyndryl will stay relevant.” 

In April, Kyndryl laid off an undisclosed number of workers, which the firm framed as part of broader cost-cutting measures. Schroeter said he expects the company’s workforce to be “stable” over the next year or two. “Canada has by no means been over-invested in by Kyndryl,” he said, adding that the firm is looking to spend on re-skilling here.