Key takeaways
- Semiconductor demand keeps growing: The global push toward AI infrastructure, cloud computing, and advanced electronics is creating massive demand for the companies that design and manufacture the hardware powering these trends, and Canada has real exposure to this theme.
- Celestica is a legitimate player: Celestica has transformed itself from a traditional contract manufacturer into a company with serious exposure to high-growth end markets like hyperscaler data centers, and its financial results have reflected that shift in a big way over the past couple of years.
- Valuation and concentration risk matter: When a stock runs as hard as Celestica has, you need to pay close attention to how much of the growth is already priced in and how reliant the business is on a handful of major customers, because losing even one contract can change the story fast.
Canada doesn’t have a deep bench of semiconductor stocks. That’s not a knock on the market, it’s just the reality of where the country’s strengths lie. Most of the chip supply chain lives in the U.S., Taiwan, South Korea, and the Netherlands. So when Canadian investors want exposure to semis, they usually end up buying U.S. names or grabbing a semiconductor ETF. Fair enough. But there are a few Canadian-listed companies that touch this space in meaningful ways, even if they aren’t designing chips themselves.
Firan Technology Group is a name I find genuinely interesting here. It manufactures printed circuit boards and electronic assemblies, which are the backbone of pretty much every piece of electronic equipment in existence. These aren’t glamorous products. Nobody’s writing breathless headlines about PCB fabrication. But without these components, nothing works. Not defense systems, not aerospace avionics, not the industrial equipment that keeps critical infrastructure running.
What draws me to a company like this is the customer base. When your revenue is tied to defense and aerospace programs, you’re not competing on price with commodity manufacturers overseas. You’re competing on precision, certification, and reliability. Those are moats that take years to build and are extremely hard to displace. The switching costs for customers are real.
The Canadian tech sector has a handful of these hidden compounders, companies that aren’t household names but have been quietly building shareholder value through operational excellence rather than hype. Firan fits that profile. It’s a small cap with a niche position in a critical part of the electronics supply chain.
So does the business actually hold up under scrutiny? That’s what I wanted to figure out.
In This Article
- Celestica Inc. (CLS.TO)
Celestica Inc. (TSX: CLS)
Celestica Inc. is a leading global provider of electronics manufacturing services (EMS) and supply chain solutions...
Competitive Edge
- Celestica's HPS business creates switching costs that pure EMS peers like Flex or Jabil lack. Designing custom rack-scale platforms for hyperscalers like Meta and Microsoft embeds Celestica into the product development cycle, not just the manufacturing step.
- The AI infrastructure buildout is a multi-year capex cycle. Celestica sits at the intersection of networking (400G/800G switches) and compute (GPU server assembly), giving it exposure to both the training and inference phases of AI deployment.
- Dual-listed on TSX and NYSE with USD reporting, Celestica benefits from CAD-denominated labor costs in some facilities while billing in USD. This natural currency hedge improves margins when CAD weakens.
- Celestica's global manufacturing footprint across Thailand, Malaysia, and Mexico provides tariff diversification that is increasingly valuable as US-China trade tensions push customers to restructure supply chains away from China.
By the Numbers
- CCS segment income grew 58.4% YoY on 41.6% revenue growth, meaning CCS margins expanded from 7.4% to 8.2%. This operating leverage in the highest-growth segment is the most important margin story in the business.
- Consensus estimates imply revenue roughly tripling from $12.4B trailing to $45.9B by Y4, with EBIT estimates growing from $1.04B to $2.37B. The forward P/E of 30x compresses to roughly 16.6x on Y2 estimates of $14.88 EPS, which is cheap if the AI infrastructure build-out sustains.
- Communications revenue surged 80.6% YoY to $7.1B in FY2025, now representing 57% of total revenue versus 28% in FY2023. This concentration in hyperscaler/data center spend is driving the entire re-rating thesis.
- ATS segment income grew 17.3% YoY despite revenue growing only 1.5%, implying ATS margins recovered from 4.6% to 5.3%. This signals cost discipline in the slower segment rather than chasing unprofitable volume.
- Trailing EBIT margin of 8.4% is structurally higher than the EMS industry norm of 3-5%, reflecting the shift toward higher-value Hardware Platform Solutions (HPS) original design work rather than pure contract assembly.
Risk Factors
- SBC of $74.3M represents roughly 5.7% of trailing net income (implied ~$1.3B from P/E and market cap), and buybacks of $98.3M barely exceed SBC. Net share reduction is minimal, so the buyback yield of 0.28% is mostly offsetting dilution, not creating shareholder value.
- P/FCF of 70.4x versus P/E of 36.4x reveals a massive gap. FCF yield of 1.4% against earnings yield of 2.7% means roughly half of reported earnings are consumed by working capital or capex. FCF conversion trend is flagged at -1, confirming deterioration.
- Enterprise revenue declined 18.9% YoY to $2.06B after three consecutive years of 30%+ growth. This sharp reversal suggests enterprise IT spending is pulling back, and the Communications surge is masking weakness elsewhere.
- P/B of 16.5x means the market is pricing in enormous intangible value. If hyperscaler capex cycles moderate, the premium over tangible assets leaves significant downside risk with little book value support.
- Revenue concentration is intensifying dangerously. Communications went from 34% of CCS revenue in FY2023 to 78% in FY2025. A single large hyperscaler pausing orders could create a revenue cliff given EMS contract structures.
The Canadian semiconductor space is tiny. One name. That’s what we’re working with here. And honestly, I think that tells you something important about how you should approach this as a Canadian investor. You’re not picking from a deep pool. You’re making a very specific bet on a very specific business.
That can actually be a good thing. When there’s no basket to hide behind, you’re forced to understand exactly what you own. With Firan, the question isn’t whether semiconductors as a sector will grow. Of course they will. The question is whether this particular company, at this particular size, with its particular customer relationships, can keep converting its niche position into growing earnings. That’s a narrower, more honest question to ask yourself, and the answer should drive your decision more than any broad sector enthusiasm ever could.
