Celestica: 2027 Inflection Is Getting Bigger, AI Growth Still Underestimated

Celestica raised its 2026 outlook to $20.5 billion in revenue and $11.30 in EPS as Q3 growth nears 69% year over year. Component lead times above 52 weeks are prompting binding customer commitments, improving demand visibility through 2028-29. Multiple AI-infrastructure ramps—including 10 1.6T programs, CPO, AMD Helios and OpenAI Jalapeño—support a sustained growth runway beyond 2026.

Analysis

The investable question is whether CLS can convert unusually long customer commitments into a durable earnings multiple re-rating rather than merely a cyclical revenue spike. If backlog is tied to design wins across several AI architectures, its manufacturing footprint becomes a scarce integration asset: customers face meaningful switching costs once systems, thermal design, supply qualification and delivery schedules are locked. That favors sustained mix improvement and working-capital efficiency, but the market will require evidence that incremental revenue carries higher operating margin rather than being pass-through component content.

The second-order beneficiaries are optical interconnect and power/thermal suppliers—COHR, LITE, CIEN and Vertiv (VRT)—because denser AI systems require more difficult integration even if server unit growth moderates. Jabil (JBL) and Flex (FLEX) are the more direct competitive read-throughs, but CLS may retain relative upside if its programs are concentrated in higher-complexity, lower-substitutability compute racks rather than commoditized assembly. Conversely, supply scarcity can become a balance-sheet risk: inventory and customer prepayment disclosures matter more than reported backlog, since a normalization in lead times could unwind both pricing power and working-capital support.

Over the next 1-3 months, the catalyst is confirmation through order-book conversion, gross-margin progression and cash generation—not another headline guidance raise. Over 6-18 months, the key risk is AI capex becoming concentrated among a few buyers whose deployment schedules slip after initial cluster build-outs; that would expose the market's assumption that every announced platform translates into recurring hardware content. Falsify the constructive view if CLS reports sequential margin dilution, material growth in inventory or contract assets relative to sales, or reduced visibility beyond the next two quarters.

Key Decisions for Investors

• Maintain or initiate a long CLS position on post-earnings consolidation rather than chase a gap higher; size against a 12-18 month horizon. Target a 20-30% upside case from earnings/multiple expansion, with thesis review if the next two reports fail to show operating-margin expansion and operating cash flow conversion.
• Run a relative-value pair: long CLS / short FLEX, sized beta-neutral, for the next 1-3 quarters. The thesis is that complex AI rack integration earns superior margin and customer stickiness; exit if FLEX demonstrates faster AI-related margin expansion or CLS's gross-margin premium narrows.
• Add a basket hedge through a modest short SMH or SOXX against the CLS long if portfolio AI exposure is already high. This isolates execution and share-gain exposure while protecting against a broad hyperscaler-capex de-rating; remove the hedge if order conversion and cash flow validate company-specific differentiation.
• Set an accounting/liquidity alert before increasing exposure: require disclosure of customer deposits, inventory turns and contract-asset growth. A material deterioration in cash conversion despite revenue growth would indicate that supplier constraints are shifting financing burden onto CLS, reducing the quality of the earnings upgrade.