Glass fiber cloth is a huge bottleneck for GPUs, ASICs and 800G/1.6T switches, and demand is shifting to alternatives as Nittobo, with 90% of T-glass and 70% of NER-glass, lacks capacity until mid-2027. Taiwan Glass is among the strongest near-term AI plays, with low-CTE output ramped in Q1 2026 and Nvidia/AMD certification. Fulltech is moving to T-glass, and Nan Ya Plastics is integrated via Nittobo. Unitika is most compelling, with asymmetric upside if Qualcomm or hyperscalers sign long deals. … Show more

Unitika is more than a meme stock: its ultra-thin, low-CTE glass fabrics keep large AI dies from warping in advanced package substrates, and supply is short. Leader Nittobo's order books are reportedly full well into 2026, with new capacity not before H2 2027, and Qualcomm has visited Unitika seeking another source. Only a handful of firms can supply NE-glass for AI-server laminates. Now restructured, Unitika nearly doubled Q3 operating profit to ¥9.0B; May 14 results may quantify AI revenue. … Show more

Unitika may be too small after its restructuring, or too early in its new cycle, to benefit from AI glass demand, so the earlier thesis was too aggressive. Glass fiber and cloth sales fell 9% year over year to ¥33.7 billion, while operating profit rose 436.5% to ¥1.6 billion on one-offs. Unlike the prior quarter's mention of vigorous AI-related E-glass demand, the latest results had no explicit AI language. Its cleaner balance sheet keeps capacity and qualification updates worth watching. x.com 

The move from organic ABF substrates to glass is becoming a necessity for Nvidia's largest packages, and Nittobo, Schmid, LPKF and Ibiden sit at its choke points. Nittobo has a near-monopoly on low-CTE T-Glass fiber amid a shortage; LPKF's LIDE drilling and Schmid's inline systems ride US and EU sovereign AI pushes; Ibiden bridges ABF and glass core. Timing is the risk: TSMC pushed panel-level CoPoS to Q4 2030, so Rubin Ultra may use hybrid glass stiffeners, and Nittobo has already run 9x. x.com  … Show more

LPKF Laser is a bottleneck supplier for glass substrates, the next-generation AI packaging material Intel, Samsung and TSMC are piloting. Glass cracks when drilled, and LPKF's LIDE is one of very few processes that structure glass at production scale without cracking, so no glass ecosystem gets built without a tool like it. Management says the question is now when the ramp begins, not whether. Besi trades at 77x EBITDA on advanced packaging; upstream names trade at fractions of that. x.com 

LPKF Laser is a small-priced option on glass substrates becoming the next generation of AI chip packaging, with 4-10x upside if glass wins. It owns LIDE, the only working laser process that drills precise holes in glass without cracking it. Intel has committed to high-volume glass by 2030, Samsung is piloting and TSMC is evaluating. At a 293M EUR market cap it is pre-revenue inflection with no meaningful EBITDA, which is the point of the bet. x.com 

Corning is the wrong instrument for glass substrates, the material set to replace organic substrates in AI chip packaging. Corning, AGC and Schott control about 90% of semiconductor-grade glass, and Corning works with TSMC with CHIPS Act funding. But substrates are a small part of a conglomerate whose growth comes mainly from AI optical fiber, trading at 90x earnings. The option value materializes in 2027-28; 2026 is for qualification, with yields at 75-85% and costs 2-3x organic. x.com 

LPKF's earnings release shows its future in glass advanced packaging, even as current results look weak. Leading chipmakers are deploying its LIDE technology, and follow-on orders confirm it as key for crack-free glass processing. LPKF is expanding into glass package singulation and laser bonding of multilayer glass stacks, and working on co-packaged optics on glass. @ren_stocks notes revenue fell to €17.1M with a -40% EBIT margin, but the case rests on 2027 LIDE glass-core mass production. … Show more

Schmid Group is moments from dilution, with almost no room for error in its cash runway. That financial position is critical to any comparison of glass-substrate plays, even as a post the author replies to argues upstream suppliers gain pricing power after Intel's CFO called glass a margin headwind. x.com