CIBC raises Celestica stock price target on AI datacenter growth By Investing.com

CIBC raised its Celestica price target to $480 from $425 and maintained an Outperformer rating, citing improving visibility into AI datacenter spending from Google, Meta, Amazon, and a likely OpenAI-linked ramp. Celestica’s Q1 2026 EPS of $2.94 beat the $2.82 consensus and revenue of $5.51B slightly topped the $5.49B estimate, while supply constraints were largely resolved. RBC also lifted its target to $440 from $400, reinforcing a constructive outlook despite near-term concerns about OpenAI overbuild risk and supply-chain shortages.

Analysis

The key signal is not the upgraded target itself, but the market’s willingness to treat the current pullback as a financing/visibility reset rather than a demand reset. That matters because Celestica sits in the narrowest part of the AI supply chain: when hyperscalers and model providers de-risk their capex plans, allocation tends to shift first into networking, interconnect, and high-mix manufacturing rather than into the broader hardware basket. If that sequence holds, the next leg of upside should show up in estimate revisions before it shows up in revenue, which is why the stock can re-rate months ahead of the actual ramp.

The second-order beneficiary is AVGO, not because of direct revenue leverage from this name, but because the switch silicon and networking stack appear to be the gating item for the next AI datacenter buildout wave. If the DS6000 rollout is real volume and not a one-off qualification win, it implies accelerating demand for the adjacent networking ecosystem and a tighter supply backdrop for components with long qualification cycles. That tends to reward the highest-quality vendors first, while mid-tier ODM/OEM competitors get squeezed on mix and availability.

The main risk is timing. A 2H26 ramp story means the equity can remain sentiment-driven for several quarters, and any headline about overbuild risk, delayed customer orders, or renewed component shortages can keep multiple compression in place even if the fundamental thesis remains intact. The consensus may be underpricing how much optionality sits in estimate drift for 2026-2027, but it may also be overconfident about straight-line execution from first-quarter de-risking to second-half monetization.

For the hyperscalers, the read-through is modest but positive: improving capex visibility generally supports capital allocation confidence, but it also raises the bar for returns on AI spend. If the market starts to believe the buildout is shifting from experimentation to infrastructure hardening, that is a subtle tailwind for GOOGL, META, and AMZN as the capex narrative becomes less discretionary and more platform-defensive. The contrarian angle is that a near-term selloff in CLS could persist even if the 12-18 month thesis is intact, making this more of a patience trade than an immediate catalyst trade.

Key Decisions for Investors

• Long CLS on weakness for a 6-12 month horizon; use the selloff to build a position near support, targeting continued estimate revisions and a re-rating toward peer AI infrastructure multiples. Risk/reward is attractive if 2026 visibility holds, but trim if new supply constraints or customer concentration headlines emerge.
• Pair trade: long CLS / short a slower-growth hardware or contract-manufacturing proxy over the next 3-6 months. The thesis is that AI mix and visibility should drive multiple expansion for CLS while the short leg remains tied to more cyclical, lower-growth demand.
• Add AVGO as the cleaner downstream beneficiary of the networking/switching cycle over the next 6-18 months. If AI datacenter buildouts broaden beyond compute into interconnect, AVGO should capture higher-quality incremental demand with less execution noise than hardware assemblers.
• For GOOGL, META, and AMZN, use any selloff related to capex skepticism to selectively add exposure over the next quarter. The improving visibility into AI infrastructure spend reduces left-tail risk around 'capex waste' and supports the medium-term platform story.
• If you want convexity, consider a 6-9 month CLS call spread financed against upside in a hyperscaler basket. The structure expresses upside from estimate drift while limiting downside if the market keeps punishing hardware names on overbuild fears.