Optical Chips

Hey everyone, let me lay my cards on the table. After looking at the curve over the past four quarters, I, as CFO of an optical chip company, am actually pretty excited—you think about it, our revenue was only 120 million in Q2 2025, and it shot straight up to 355 million by Q1 2026. That's nearly a 3x jump in four quarters. The net profit growth is even more dramatic, and gross margin climbed from 51.7% all the way to 77.8%. That slope is top-tier even by the standards of the entire A-share tech sector. For this forecast, my two core assumptions are: **First, the "optical-in, copper-out" shift in AI compute infrastructure is still accelerating.** 800G optical modules are entering mass production in 2026, and 1.6T has started shipping in small batches—we happen to be one of the very few domestic players capable of mass-producing 200G/400G EML laser chips. The 77.8% gross margin in Q1 2026 is no accident; it's the result of our product mix shifting upmarket—CW light sources and EML chips for silicon photonics integration now carry a higher share. I'm assuming this mix upgrade continues through Q2 and Q3, so I keep gross margin in the 75-76% range—slightly retreating from the Q1 high, but still an absolutely beautiful level in absolute terms. **Second, the pace of domestic substitution has not been disrupted by external forces.** Under the geopolitical backdrop, major domestic customers (data center operators and big module makers) are showing an increasingly clear preference for sourcing domestic high-speed optical chips. Assuming our InP production line ramps capacity smoothly, I peg Q2 revenue at around 490 million, another 38% sequential increase; for Q3, I model roughly 27% sequential growth, landing near 620 million. I give net profit growth a slightly weaker figure than revenue, because R&D and capacity expansion investments will eat into some of the operating leverage. **The biggest risk—I have to be upfront about it: customer concentration.** Our top customers account for a very high share of revenue. If any major customer adjusts inventory or switches to a second source, quarterly results will swing wildly. On top of that, if overseas incumbents like II-VI and Lumentum ramp up their investment in the China data center market again, pricing pressure will materialize. And then there's supply chain stability for InP substrates and specialty gases—these are all variables hanging over our heads. So these two numbers are drawn on a path based on three premises: "AI demand continuing + domestic substitution deepening + capacity ramping smoothly." If any one of those premises loosens, the curve has to be redrawn. I'd rather discuss this with you as a "optimistic but not unrealistic" scenario chart, not a promise carved in stone. The optical road—we're still walking it.