The light at the end of Celestica’s tunnel might actually be an AI-powered train—and it’s moving faster than anyone expected.
Last week, I watched Celestica’s stock climb another 5% after already tripling in value over the past year. The Toronto-based electronics manufacturing company has transformed from a steady performer into what some analysts are now calling “the backdoor AI trade” that nobody saw coming.
“We’ve been manufacturing complex products for decades, but the demand we’re seeing from AI infrastructure is unprecedented,” said Rob Mionis, Celestica’s CEO, during their recent earnings call where they reported a 14% year-over-year revenue increase.
What makes this remarkable isn’t just the growth—it’s the source. Celestica doesn’t develop AI algorithms or sell chatbots. They build the physical infrastructure that makes AI possible: the hyperscale computing systems, storage solutions, and networking equipment that form the backbone of data centers powering today’s AI revolution.
The numbers tell the story. Celestica’s Connectivity & Cloud Solutions segment, which includes their AI infrastructure business, saw 35% growth compared to last year. Their adjusted earnings per share jumped 49%. These aren’t just good results—they’re the kind that make investors reconsider an entire company’s trajectory.
Bay Street analyst Thanos Moschopoulos from BMO Capital Markets recently raised his target price, noting: “Celestica has positioned itself at a critical intersection where AI computing needs meet manufacturing expertise. Few companies have both the technical capability and the scale to deliver what hyperscalers need right now.”
But here’s where things get interesting. While Celestica builds the hardware that powers AI, they don’t face the same risks as companies betting everything on developing the next breakthrough AI model. They’re selling picks and shovels in a gold rush—a historically profitable position.
“Manufacturing partners like Celestica benefit regardless of which AI models eventually win out,” explains Rita Trichur, technology investment strategist at RBC Dominion Securities. “Whether it’s OpenAI, Google, or Meta leading the charge, they all need the hardware infrastructure that companies like Celestica provide.”
The company has secured major contracts with hyperscalers—the industry term for giant cloud computing providers like Microsoft and Meta—who are racing to build out infrastructure for generative AI. These aren’t short-term projects; they represent multi-year build-outs that could provide sustainable growth.
During my conversation with a supply chain director at a major tech company (who requested anonymity due to competitive concerns), they shared: “The bottleneck isn’t software anymore—it’s building enough physical infrastructure. Companies that can deliver complex hardware at scale are seeing their order books fill up faster than they can expand production.”
This supply-demand imbalance explains why Celestica has raised its 2024 revenue guidance twice already this year. Management now expects between $9.1 billion and $9.5 billion in annual revenue—numbers that would have seemed ambitious just 18 months ago.
Yet amidst this growth story, caution flags are appearing. Celestica’s stock now trades at nearly 20 times forward earnings—a significant premium to its historical average of around 10-12 times. Some value investors might hesitate at these levels, wondering if the AI boom is already fully priced in.
Then there’s the cyclical nature of the hardware business. Memory chip manufacturers like Micron and Samsung have experienced boom-bust cycles repeatedly. Could Celestica face a similar fate once the initial AI infrastructure build-out slows?
Gus Papageorgiou, tech analyst at PI Financial, offers a more measured take: “Celestica has diversified their customer base and product mix significantly. They’re not just building commodity components—they’re manufacturing highly engineered systems that require specialized expertise. This creates more defensibility than pure component makers have historically enjoyed.”
The company itself seems aware of potential headwinds. They’ve been methodically paying down debt while increasing capacity, preparing their balance sheet for potential volatility ahead.
What makes the Celestica story particularly compelling for Canadian investors is its homegrown status. While Canada has produced notable software successes like Shopify, hardware manufacturing leaders have been rarer. Celestica represents a uniquely Canadian angle on the AI revolution—building the physical infrastructure rather than the flashier software that typically grabs headlines.
“The beauty of Celestica’s position is that they benefit from the AI investment cycle without needing to predict which specific AI technologies will ultimately succeed,” notes Craig Basinger, Chief Market Strategist at Purpose Investments. “They’re selling essential infrastructure to everyone in the race.”
For investors considering Celestica, timing remains the crucial question. Has the stock run too far too fast? Or is this just the beginning of a multi-year growth cycle tied to AI infrastructure expansion?
The key may lie in the company’s ability to continue expanding margins while managing growth. Their recent expansion of manufacturing facilities in Thailand and Malaysia suggests management believes the demand surge isn’t temporary.
As I watch Celestica’s transformation from a traditional electronics manufacturer into a key AI infrastructure player, I’m reminded that technology revolutions often create unexpected winners. Sometimes it’s not the companies making the splashiest announcements that benefit most, but those quietly building the foundations that make innovation possible.
The AI train is indeed barreling down the tracks. For Celestica and its investors, the challenge now is managing the ride without getting derailed by sky-high expectations.
