Celestica: The Margin Supporting 35x Was Not There In 2023

Celestica says its AI hardware accelerator programs (HPS) are set to lift revenue share from 21% (2023) to 41% (2025), supported by execution on $1.6T programs and ongoing engineering investment. However, margins remain in the single digits and value is perceived as accruing more upstream. The stock trades at 35x earnings and 79x FCF, implying the market needs 50%+ EPS growth and FCF above $1.2B by 2027 for valuation support.

Analysis

Celestica is functioning more like a pass-through of AI infrastructure demand than a true economic bottleneck. The margin profile implies the bulk of value creation is accruing to the silicon, networking, and optics vendors upstream, while CLS absorbs execution risk, working-capital drag, and customer concentration without comparable pricing power. That makes peers with real design content or network attach rates more attractive expressions of the same AI buildout than the manufacturing layer itself.

The key risk is that the market is paying today for a multiyear earnings ramp that still has to clear several hurdles: sustained program wins, no mix dilution, and no customer pause in AI capex. In the next 1-3 months, the stock can remain momentum-supported if order commentary stays clean, but over 6-18 months the valuation becomes fragile if the cadence of 1.6T deployments normalizes or hyperscalers push back on vendor economics. Any sign that margins plateau in the high-single digits or free cash flow undershoots the implied path should compress the multiple aggressively.

Contrarian view: the consensus may be treating revenue growth as if it were the same as durable profit growth. It is not, especially in a low-margin assembler where bargaining power sits with the customer and the product cycle can change quickly. If CLS delivers the required EPS/FCF ramp, the stock can work, but the burden of proof is unusually high at this valuation; absent a meaningful upward revision to margin structure, the risk/reward skews against owning the name outright.

Key Decisions for Investors

• Short CLS on strength; best entry is after any AI-driven gap-up or positive order commentary. Target 3-6 month horizon. Thesis fails if HPS mix drives a clear gross-margin inflection above low-teens and management raises 2027 free-cash-flow guidance materially above $1.2B.
• Pair trade: long AVGO or ANET vs short CLS to express continued AI infrastructure spend while avoiding the low-margin manufacturing layer. Use a 3-12 month horizon; this should outperform if AI capex remains strong but value capture stays upstream.
• If unable to short outright, buy CLS 6-12 month put spreads struck below the current valuation multiple zone to capture multiple compression risk with defined downside. This is most attractive if implied volatility is not already rich.
• Set a watch item on quarterly FCF conversion and receivable/inventory build. A working-capital spike would be the earliest sign that revenue growth is being funded, not monetized.
• Trim or avoid any long position unless the company demonstrates a durable step-up in operating margin; without that, the stock is paying a premium for a value-chain slot that does not appear structurally scarce.