AIToday
AI Stocks & MarketsTop Companies' AI MovesAI Business & IndustryTop Companies AIPublished: Oct 8, 2026, 06:31 JST

Disco posts record 1,239億円 in quarterly shipments, up 60.0%増

Disco posts record 1,239億円 in quarterly shipments, up 60.0%増

3 Key Points

  1. What happened

    Disco said its July–September 2026 preliminary individual shipment value rose 60.0% year on year to 123.9 billion yen ($... no), a quarterly record, with both precision processing equipment and tools staying strong on generative AI demand.

  2. Why it matters

    The record suggests the AI investment wave is reaching back-end cutting and grinding steps, not just front-end lithography and deposition tools, though the market treated the news as already priced in.

  3. What to watch

    Disco's shares fell about 6% the next day as Nikkei called the results 'within expectations' and Kabutan said there was 'little surprise'; the finalized Q2 results are due October 22.

WHO IT HITSThis matters to semiconductor equipment investors and analysts tracking whether AI capex spreads beyond front-end tools into back-end processing, and to HBM supply-chain watchers assessing whether simplified stacking specs lift demand for Disco's molding and singulation equipment.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Context & Analysis

Disco is a global leader in cutting and grinding equipment for critical semiconductor components, and it discloses shipment figures as a preliminary indicator because it books machine revenue only when customers accept delivery—making shipments a better read on actual market conditions. The July–September quarter's 123.9 billion yen in individual shipments beat both the company plan and analyst consensus on a consolidated basis, with estimated consolidated shipments of 149.1 billion yen and sales of 143.9 billion yen. The prior quarter's shipments were 116.5 billion yen, and the year-earlier figure was 77.4 billion yen, underscoring the scale of the increase.

The results come as the market debates HBM specification changes. TrendForce indicated in August 2026 that NVIDIA was considering alternatives to a 12-layer HBM4e configuration for its next-generation 'Rubin Ultra' AI chip, including 8-layer stacks, though final specs remain undecided. Disco argues this would not reduce total production or wafer counts, and could instead increase HBM package counts—lifting demand for its molding and singulation equipment. The company also said the consolidated-to-individual ratio should hold near its usual roughly 1.2x, though a higher-than-expected consumables mix could push shipments to the upper end of the range.

Whether the AI-driven demand translates into sustained earnings growth hinges on whether the back-end processing demand that lifted Disco's shipments persists as HBM specs evolve, and on how the finalized results compare with these preliminary figures when Disco reports on October 22. For equipment investors and HBM supply-chain watchers, the key question is whether this quarter's record marks a durable broadening of AI capex into back-end steps or a one-off spike already reflected in share prices.

FAQ
How much did Disco's July–September shipments grow?
Preliminary individual shipment value rose 60.0% year on year to 123.9 billion yen, a quarterly record, up 6.4% from the prior quarter's 116.5 billion yen.
Why did Disco's stock fall despite record shipments?
The Nikkei reported voices calling the results 'within expectations,' and Kabutan said there was 'little surprise,' prompting profit-taking selling.
How does Disco view HBM specification simplification?
Disco takes a positive view, saying fewer stacked layers would increase HBM package counts and expand use of its molding and singulation equipment, even though total output and wafer counts would not change.
Top Companies AIRead Original Article

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleOpenAI posts 722 AI-generated math papers on GitHub