Skip to content
World clockEU--:--UK--:--USA--:--CN--:--PLDEFRIT中文EN

portal about AI and technologyevents · analysis · interviews · technical background

Search
LIVE
›

The AI accelerator crunch has moved to a 90% monopoly in glass cloth

A single Japanese company, Nittobo, controls roughly 90% of the global supply of the specialist glass-fiber cloth inside every advanced AI chip package. Its new capacity will not reach the market until mid-2027.

TechnologyAnalysisRachel NwosuPublished: 27 September 20264 min readSources 3
The AI accelerator crunch has moved to a 90% monopoly in glass cloth

Nittobo's T-glass is a low-CTE (coefficient of thermal expansion) glass cloth. It sits in the organic core of IC substrates, the interconnect layer between a chip and its printed circuit board. It keeps large, high-heat packages dimensionally stable. That matters more as AI processors grow bigger and run hotter. Tom's Hardware reported on 9 March that the material sits inside every advanced AI chip package.

Prices have risen between 20 and 30%. Lead times for downstream materials such as copper-clad laminates have stretched from a normal 8 to 10 weeks to beyond 20.

"With T-glass supply even more constrained now, suppliers are no longer providing lead times," said Bill Ho, an analyst at Yuanta, according to Tom's Hardware.

The bottleneck is not easy to design around. Bilal Hachemi tracks the IC substrate supply chain at Yole Group. He told Tom's Hardware that T-glass "has specific dielectric and CTE values that work better for the AI chips, especially for the organic core." Cheaper E-glass goes into lower-end chips such as microcontrollers and older mobile processors. For massive 2.5D and 3D packaging, and so for advanced AI accelerators, T-glass is the preferred option.

Nittobo is tripling capacity at its Fukushima plant in Japan. Tom's Hardware reports that new supply will not arrive on the market until mid-2027. New production lines cannot be spun up overnight. The material needs specialized electric melting furnaces running between 1,600 and 1,700 degrees Celsius. The process melts silica-rich glass, spins it into yarn and weaves it into an ultrathin cloth. Hachemi also pointed to the industry's thin margins. "Any increase in demand for build-up materials, ABF material, or T-glass can cause potential shortage, because it's against the basics of this industry," he said.

Bigger packages, more cloth

Hyperscalers are ordering ever-larger chip packages that consume more T-glass per unit. Nvidia data cited by Tom's Hardware shows interposer sizes grew from 814mm2 for the Hopper architecture to 1,700mm2 for Blackwell, a 109% increase. The forthcoming Rubin and Feynman generations scale further. Bank of America estimates Nittobo's electronic materials segment will nearly double sales from 40.9 billion yen ($266 million) in 2025 to 87.7 billion yen by March 2028. Operating margins approach 48%.

"Demand for T-glass cloth seems likely to grow more than originally expected," said Takashi Enomoto, a research analyst at Bank of America, per Tom's Hardware. "The focus has been on growth in demand for thick T-glass for use in GPU and CPU semi packages, but now ultra-thin T-glass demand is likely to rise on a shift from E-glass to ultra-thin T-glass in leading-edge devices."

The scramble for allocation has already produced an unusual sight. Hachemi told Tom's Hardware that Nvidia reaching out directly to an upstream material supplier is unprecedented in that part of the supply chain. "For the first time, we are seeing Nvidia, the end customer, reaching out to the upstream material suppliers to secure the capacity and make sure they will get it," he said. The risk is that once Nvidia locks down its share, rival chip buyers fight over the remainder.

Nittobo is not relying on Fukushima alone. It is doubling raw yarn capacity at its Taiwan plant and importing yarn back to Japan for cloth manufacturing. It has also struck a collaboration deal with Nanya Plastics to outsource some weaving. Tom's Hardware reports that by 2027 roughly 20% of Nittobo's glass cloth is expected to be woven by Nanya. Partnering with one of its biggest competitors shows how tight the market has become.

The wider accelerator pipeline is not standing still. Microsoft's Maia 200, presented at Hot Chips 2026, is built on TSMC 3nm with a 750 Watt TDP. It has 7TB/second of HBM bandwidth across 6 HBM stacks and 10,000 TFLOPS of FP4 performance, according to ServeTheHome's write-up of the session. Those packages still need substrate material that only a handful of suppliers can make.

Nvidia is meanwhile pushing AI deeper into its own design flow. Tim Costa is vice president and general manager of computational engineering at Nvidia. He told journalists that AI and accelerated computing are moving "from productivity tools into being foundational engineering infrastructure," The Next Platform reported on 27 July. Cadence claims its AI Super Agents can deliver 40-times faster Register-Transfer Level validation cycles. Synopsys says its agentic workflow can deliver validated RTL 50 times faster than other platforms.

None of that shortens the queue for T-glass. The substrate crunch is a reminder that the AI buildout depends on materials few people have heard of. They are made by companies with almost no competition.

Comments 0

Sources

3
  1. 01Shortages of crucial chip packaging material threatens AI accelerator supply chainsEN
  2. 02Microsoft's Maia 200 AI Accelerator at Hot Chips 2026EN
  3. 03Nvidia Accelerates Chip Engineering With AI AgentsEN

All figures and quotations in this text come from the sources listed below.

Content prepared by the editorial team with AI assistance.

Rachel Nwosu

Rachel Nwosu

AI, models and technology

Rachel Nwosu covers AI, models and technology for FLASH24, working from public model documentation, benchmark releases and repository histories rather than press summaries, and she skips announcements that arrive without reproducible numbers. She checks training-data claims against dataset cards and reruns reported metrics where code is available. She spends much of her week interviewing researchers and engineers, tracking model launch calendars, and comparing vendor benchmarks with independent evaluations. Outside the desk she runs 3D printers, restores old computers, and tests how models learn from internet junk. She does not publish benchmark figures she cannot trace to a source.

Newsroom →

Comments

0
  1. No comments yet — be the first.

Write a comment

Comments are public. We do not publish abuse, spam or advertising.