Disco Earnings Q1 FY2027: Net Profit Up 44% — Why the Company That Only Cuts and Grinds Chips Is the AI Boom's Surest Winner

On July 23, DISCO reported its first-quarter results for the fiscal year ending March 2027 (April-June 2026). This is not a company in a glamorous process step like lithography or deposition. It specializes in dicing finished wafers into individual chips (in Japanese) and grinding their backsides thin — and that narrow specialty is capturing the profits of the AI boom more reliably than almost anyone else.

How should the April-June 2026 numbers be read?

Revenue came to 114.31 billion yen, up 27.1% year on year; operating profit was 49.03 billion yen, up 42.2%; and net profit was 34.22 billion yen, up 44%. The drivers were manufacturing equipment for AI chips and sales of consumables such as blades — the cutting edges used to dice wafers (Nikkei, July 23, 2026).

The order of those growth rates is the point. While revenue grew 27%, operating profit grew 42% and net profit 44%. In other words, the more DISCO sells, the higher its margins go. The operating margin for the quarter reached roughly 43% — an extraordinary level in manufacturing, where anything above 10% is considered strong.

Nor is this a one-off windfall. Behind it sits the two-tier revenue structure described below, in which equipment sales are followed by recurring consumables sales (DISCO, Q1 FY2027 earnings report, July 23, 2026).

Why does a company that only cuts and grinds make money in the AI boom?

DISCO has two flagship product lines: dicing saws, which cut wafers into individual chips, and grinders, which thin wafers by grinding their backsides. In process terms, they sit at the entrance to the back end — the stage after circuits have been formed. In both categories, DISCO holds a world-leading share. Our company deep dive covers the business in detail.

This niche machining step has landed at the center of the AI boom. The reason lies in how HBM (high-bandwidth memory), the memory essential to AI servers, is made. HBM is built by grinding memory dies (DRAM) thin and stacking them vertically, many layers high. Grinding wafers thinner than a human hair (in Japanese) is a mandatory step for stacking — and the more layers are stacked, the more grinding and cutting there is to do.

The spread of advanced packaging — the technology that integrates a GPU and HBM on a single substrate — adds another tailwind. More packaging steps mean more singulation, more thinning, more machining, and DISCO's equipment appears at many of those points. Why memory, not the GPU, determines AI performance (in Japanese) makes the company's position easy to see.

The performance race in AI chips therefore translates directly into more machining passes for DISCO. That is why a company that only cuts and grinds is the AI boom's surest winner.

How much does the consumables business matter?

Alongside equipment, this quarter's other named growth driver was consumables such as blades. A dicing blade is an ultra-thin cutting edge made of bonded synthetic diamond, and it wears down with every wafer it cuts. Grinding wheels for grinders wear out the same way.

This changes the quality of the earnings. A machine is sold once; consumables are reordered for as long as the customer's fab keeps running. As chip fabs around the world raise output, DISCO books new equipment orders and repeat consumables orders at the same time. Because the same company makes both the machine and the blade, customers rarely switch to third-party consumables — another prop under the high margins.

The equipment industry is notorious for the boom-bust swings of the silicon cycle. But consumables revenue tracks fab utilization, so a base of income remains even when equipment orders stall. Behind the 44% profit jump sits this two-tier design.

Put differently, the AI boom is a double tailwind for DISCO. First comes new equipment demand as HBM and advanced packaging lines are built. Second comes growing consumables demand as those lines run at high utilization. The first is cyclical; the second gets thicker with every machine installed. Watching quarterly equipment orders alone will cause you to misjudge this company's earning power.

How do you read a company that issues no full-year guidance?

DISCO's results come with one quirk: the company does not disclose full-year forecasts. Demand for semiconductor equipment swings too widely to project a year with confidence, and this has been DISCO's consistent practice.

So what should investors use as a yardstick? The external benchmark is the analyst consensus compiled by QUICK — the average of multiple analysts' forecasts. The full-year revenue consensus for the fiscal year ending March 2027 is 537.24 billion yen (Nikkei, July 23, 2026).

Multiplying this quarter's 114.3 billion yen by four gives roughly 457.2 billion yen, short of the 537.2 billion yen consensus. In other words, the market's forecast assumes quarterly revenue keeps climbing into the second half. In the quarters ahead, the question is whether actual results stay on pace toward that consensus. With no company guidance, the gap between quarterly results and consensus is, in effect, the progress rate.

Key takeaways

• DISCO's April-June 2026 quarter delivered revenue of 114.31 billion yen (+27.1%), operating profit of 49.03 billion yen (+42.2%) and net profit of 34.22 billion yen (+44%).
• Growth was driven by equipment for AI chips and consumables such as blades. The operating margin reached roughly 43%.
• HBM is made by grinding DRAM thin and stacking it, so demand for cutting and grinding rises structurally as AI chips proliferate.
• A two-tier revenue structure — equipment sales plus recurring consumables orders — underpins both the high margins and their stability.
• No full-year guidance, per company practice. The gap to the QUICK consensus (537.24 billion yen in full-year revenue) serves as the de facto progress gauge.