なぜフジクラはわずか3カ月で利益予想を2倍以上に修正したのか…背景にあるAIデータセンターブーム｢3つの変化｣ | ビジネス | 東洋経済オンライン

Fujikura lifted its profit forecast by more than double — twice — in three months, jolting the market. For Taiwanese readers this isn't a single Japanese earnings footnote: Fujikura makes optical fiber and connectivity gear, sitting at a choke point in this wave of AI data-center construction, and Taiwan's electronics and components supply chain overlaps heavily with it. Understanding why Fujikura doubled means understanding which segment the AI money actually reaches.

Companies refresh guidance at fixed points, because a revision is a promise; revising up twice — sharply — inside three months means demand is outrunning the company's own model. This isn't "steady growth"; it's "pushed by sudden demand it keeps underestimating" — a higher-quality signal than a target shouted at year-start, because reality forced it. Toyo Keizai names three shifts: sudden demand, short-lead-time delivery, and geographic diversification.

Sudden demand: AI data centers are built far faster than telecom infrastructure, with fiber and cable orders arriving big, urgent, concentrated. Short lead times: whoever ships fastest captures pricing power. Geographic diversification: customers spreading sourcing to hedge concentration — both a "second-source" opportunity and a "new rival nurtured" threat for Taiwanese firms. Investment-wise, Fujikura is a classic "picks and shovels" play: in a gold rush the steady money is often in shovels, not gold. Its profit surge confirms, from the supply side, how real the AI capex actually is — closer to ground truth than chip-maker visions.

For Taiwanese readers: treat Fujikura as a thermometer — forced, repeated upgrades mean demand is still accelerating. Distinguish "structural beneficiary" from "one-off rush order," beware double-ordering, and look past headline growth to order visibility and customer concentration. Watch its next guidance, order visibility, and how the chain re-regionalizes.