Back to News
Market Impact: 0.42

Rothschild Redburn initiates Celestica stock with buy on AI growth

SNDKCLSAVGOAMDCM
Analyst InsightsCorporate EarningsCompany FundamentalsCorporate Guidance & OutlookTechnology & InnovationArtificial Intelligence
Rothschild Redburn initiates Celestica stock with buy on AI growth

Celestica reported Q1 2026 EPS of $2.94 versus $2.82 expected and revenue of $5.51 billion versus $5.49 billion expected, despite component shortages. Rothschild Redburn initiated coverage with a Buy and a $460 target, while CIBC lifted its target to $480 and RBC to $440 on strong demand for customized Ethernet switching and AI data center hardware. Management also pointed to a significant AI-native customer ramping in early 2027, reinforcing the growth outlook.

Analysis

The tradeable signal is not just that CLS is winning share, but that the market is re-rating who captures AI infrastructure margin as architectures become more standardized. If the control plane shifts toward open-source networking software and merchant silicon, the economic moat moves away from pure switch OEMs and toward integrators with execution, supply access, and qualification depth; that is a structural tailwind for CLS and a relative headwind for higher-multiple “software-like” networking names. Broadcom is the other understated winner: every incremental design win that standardizes on its silicon broadens pricing power and cements it as the toll collector in the stack. The second-order pressure point is on legacy OEMs and any vendor whose premium multiple depends on proprietary differentiation. If hyperscaler procurement keeps pivoting toward less differentiated hardware, the market will increasingly value gross-margin durability and manufacturing scale over brand or software ecosystems. That can create a valuation air pocket in names that screen well on growth but are exposed to switching commoditization over the next 6-18 months. The key risk is timing: the bull case is predicated on 2026-2027 capex cycles and rack-scale deployments actually translating into sustained orders, not just headline wins. Any slip in hyperscaler budgets, supply-chain normalization that reduces urgency, or a design cycle delay at the large AI customer expected in early 2027 would compress the multiple quickly because the stock is already discounting a high-visibility ramp. Near term, the biggest reversal risk is not demand collapse but execution friction — component availability, qualification slippage, or a narrower-than-expected product mix that limits margin expansion. Consensus may still be underappreciating how much of this is a Broadcom-led ecosystem story rather than a pure CLS story. If Tomahawk Ultra becomes the default choice for standardized AI switching, the upside is broader than one vendor’s revenue line: it can pull forward AVGO content, force competitors into price competition, and extend the capex cycle by lowering integration risk for hyperscalers. That makes the move in CLS look directionally justified, but the cleaner expression may be via a pair that isolates the standardization theme rather than outright chasing a crowded single-name long.