Alright folks, pour the coffee—let's talk about how we're playing this next.
Look at the four-quarter curve: revenue climbed from 6.9B in Q2 last year to 7.8B in Q4, then pulled back to 7.5B in Q1 this year. Gross margin fell from 16% to 13%, and I need to address this face-to-face—it's not industry deterioration, it's the "cost" of our own capacity expansion. Depreciation on the two new FCBGA lines in Suzhou and Penang, plus yield ramp-up, is all sitting right there on the income statement.
Here's the logic behind my Q2 and Q3 2026 forecast, built on two core assumptions:
First, the AMD line is our moat. The MI300 series continues to ramp in the second half, and once MI350 hits volume, Suzhou's FCBGA utilization can pull above 85% in Q3. That's the foundation of our revenue base.
Second, traditional CPU + server clients enter peak season in Q3, consumer electronics inventory adjustment is largely behind us, and the lift in overall utilization drives gross margin recovery. But I'm not going to spin a fairy tale about getting back to 16%—the material cost structure of advanced packaging is what it is. 13.5% to 14% is the reasonable "new normal."
Revenue: 7.9B in Q2, 8.2B in Q3, QoQ growth of 6% and 3% respectively, reflecting AI compute demand increment plus seasonal pickup. Net income: 348M and 405M, driven mainly by gross margin recovery from 13.3% to 14%.
The biggest risk—and this one stings—is customer concentration. AMD alone accounts for too high a share of our revenue. Any turbulence on their end—export control escalation, inventory cycle downturn, next-gen product delay—lands directly on us. That 6.34% net margin we posted in Q3 isn't something we can replicate; that was driven by special factors and isn't baked into my forecast.