Nitto Boseki Co., Ltd.
Information Technology · Electronic Components
Near-monopoly on production-grade T-Glass - the low-CTE / low-dielectric glass cloth that reinforces AI-server IC substrates and high-layer-count PCBs. As GPU substrates grow larger (NVDA-class), core layers need ~2x the T-glass and more layers overall, sending demand vertical against a supply base of effectively three producers - Nittobo dominant on the top tier.
This is the materials-layer bottleneck of AI chip packaging, the part of the stack almost nobody covers.
- Structural bottleneck, not a fad: T-Glass lines at full capacity since H2 2023; top-tier glass $80-100/kg; substrate / PCB makers wait in line. End-demand pulled by $NVDA, $MSFT, $GOOGL, $AMZN; courted by Apple.
- Operating business compounding: FY27 guide operating profit +25%, net sales +16%.
- Multi-year scarcity window - the ¥150B Fukushima capacity-triple does not fully relieve supply until 2027-28.
- Valuation is full on NORMALIZED earnings. Reported FY26 net income +225% (¥41.8B) was inflated ~¥37B by one-off items; management guides FY27 net DOWN to ¥17B as the one-off rolls off → roughly 45-54x forward. The "cheap" trailing P/E is an optical illusion.
- The bull thesis IS the shortage - and management is actively un-bottlenecking it (¥150B triple + Nan Ya handling ~20% of output by 2027). T-Glass de-rates from scarce specialty toward tight commodity once supply catches demand.
- TSE / JPY name: FX + access friction, odd-lot mechanics under one round lot (~$14k), and no clean public-equity proxy (Nittobo IS the pure-play).
No major news in the last 7 days for 3110 - only listicles and opinion pieces, which we filter out by default. .
Taxonomy note: belongs in a future "AI Substrate / Glass-Cloth" bottleneck bubble (with Taiwan Glass + the CCL / ABF-substrate names). Mapped to semi-equipment as the closest existing home until that bubble has ≥3-4 constituents.