The upside comes mainly from shipment growth, not from a fantasy
rerating. Fabrinet already has real demand, visible capacity expansion, and
qualification-based stickiness in AI networking. If it clears component bottlenecks and converts new
datacom ramps into stable volume, revenue can nearly double by 2031. I keep the terminal multiple premium but not extreme because the business remains a high-quality manufacturing service, not a software tollbooth.