Semiconductor Equipment

Old friend, grab a coffee and let's chat. On the call for these two quarters, let me lay my cards on the table—two core assumptions. First: the capex cadence at domestic foundries won't break stride. I've been tracking the capex curves of our major customers—SMIC, Hua Hong, YMTC, CXMT. The peak windows for equipment acceptance, project milestones, and装机 confirmations are stacked right into the Q2-to-Q3 stretch, so I've built in stair-step revenue growth of 11.8B in Q2 and 13.5B in Q3. Second: the multi-product breakthrough is already showing up in the gross margin mix. Etch, thin-film, cleaning, CPC—all ramping in parallel, with high-margin SKUs lifting as a share of shipments. So I'm modeling gross margin holding around 41%, maybe ticking up slightly, and treating the 37.2% print from 2025Q4 as a year-end, lump-sum provisioning one-off—not letting it set the cadence. Here's the walk on net income: Q2 net profit 1.86B, Q3 2.16B, with net margin climbing from 15.8% to 16.0%. One leg comes from gross margin improvement; the other from non-recurring items—government subsidies and industry fund contributions—running at a normal clip. No Q2 2025-style pulse where net profit at 1.6B ran materially above operating profit. But I won't hide the structural tell: "subsidies propping up earnings" is still part of the fingerprint. One risk, and that's enough: will Q4 repeat the trick—"revenue hits 12B but operating profit shrinks to a mere 500M"? Asset impairments, warranty accruals, cross-period expense recognition—technically clean accounting, but the timing is out of our hands. If 2026Q4 pulls the same stunt, full-year profit release gets scrambled. One more thing sitting in the back of my mind: compliance review on overseas components is tightening fast. If the CF-series vacuum gauges or certain high-end valves get cut off, the multi-product ramp gets strangled at the neck. Same race against localization timelines—we've got to push hard. I'm confident on the operating story. But it's a CFO's job to always run the worst-case ledger first.