Disco Corporation

Q1 shipments were JPY111.1 billion, a record quarterly high, and management said the result was well above its external forecast. They said the overall upward revision versus the prior forecast was about JPY9 billion, including more than JPY5 billion for machinery and equipment and about JPY3 billion for other products; some Q1 shipments were pulled forward from Q2 and some evaluation/demo units were recorded as shipments. Gross profit margin in April-June was 68.1%, below the prior-year period, but management said the gap was mostly FX-related; using last year’s JPY158/USD rate, they estimate margin would have been about 70%. For July-September, they expect gross margin to decline by less than 2 percentage points from Q1, and SG&A is expected to be about JPY27 billion. Management did not give a formal revenue or EPS figure in this transcript. For Q2, they expect IC shipments to fall versus Q1, with no recovery yet in power semiconductors, while memory is expected to be 35%, OSAT 30%, and power semiconductors below 10%.

Management’s core message was that Q1 was unusually strong because of shipment timing and concentrated demand tied to generative AI, especially in ICs and HBM. They stressed that Q2 should look weaker mainly because of timing and the reverse of the Q1 pull-forward, not because generative AI demand is collapsing. On the back half, they were cautious: Q3 looks a bit better than Q2, but Q4 and beyond remain difficult to read because customer ordering behavior is becoming shorter-cycle and less visible.

Koba and Suzuki emphasized the quarter’s mix and margin drivers. Q1 gross margin was 68.1%, and they said the year-over-year decline was largely due to a stronger yen, with JPY144/USD in April-June versus JPY158 a year earlier; they estimated margin would have been about 70% at the prior-year exchange rate. For Q2, gross margin is expected to be down by less than 2 points from Q1 because the forecast assumes JPY135/USD. SG&A for July-September is expected to be about JPY27 billion, and management explained that bonuses are linked to shipments but can be adjusted across quarters through internal fine-tuning.

Analysts pressed on the size and durability of the generative-AI-driven shipment surge, whether Q3 would recover from Q2’s dip, and whether tariffs were affecting timing. Management said the Q1 strength was real but lumpy, driven by generative AI and HBM, and warned not to annualize Q1 or assume a smooth ramp; they said Q2 is weaker mainly on timing, while Q3 should be somewhat better and Q4 is still unstable. On tariffs, they said they cannot yet see a clear effect on investment timing, and North America is only about 10% of projects. Questions also focused on strategic products: management said NAND hybrid bonding is progressing into some customer mass production, PLP discussions are increasing, and WMCM demand is not yet clearly visible.