The opportunity is strong but not unconstrained. Celestica already has real AI demand, so the bull case is less about discovering product-market fit and more about converting current
hyperscaler relevance into broader rack content, steadier contract economics and sustained mix improvement. I assume it keeps a premium to standard contract manufacturers, but not a silicon-style multiple, because its advantage is execution and embedment rather than ownership of the core chip or power choke point.