Memory manufacturers Team Group and Apacer Technology reported record profits in 2025, driven by rising DRAM and NAND Flash prices and growing demand for high-capacity, high-bandwidth memory products.
Apacer took a cautious sales approach in the fourth quarter of 2025, posting consolidated revenue of NT$3.137 billion (approx. US$100 million), down 2.7% sequentially. Continued increases in memory prices, however, boosted product pricing and profitability, raising quarterly gross margin by more than 12pp and lifting after-tax net income 90% quarter-over-quarter to NT$417 million, with earnings per share of NT$3.25.
For the full year 2025, Apacer's consolidated revenue rose 42% year-over-year to NT$11.124 billion, a near 15-year high, while gross margin increased to 20.74%. Annual after-tax net income doubled to NT$860 million, producing a record EPS of NT$6.7.
Team Group saw stronger sequential gains in the fourth quarter, with revenue of NT$6.348 billion, up 52.49% quarter-over-quarter and 94.31% year-over-year. Gross margin reached 22.5%, up 9.22pp sequentially and 12.12pp from the same quarter in 2024. Operating margin improved to 16.35%, rising 13.71pp sequentially and turning positive year-over-year.
After-tax net income jumped 307.5% quarter-over-quarter to NT$815 million, with EPS of NT$9.59. For the full year 2025, Team Group reported revenue of NT$20.428 billion, up 2.46% year-over-year. Gross margin rose to 14.26%, operating margin to 6.9%, and annual after-tax net income doubled to NT$1.11 billion, a 110.62% increase year-over-year, with EPS of NT$13.06.
Both companies highlighted AI-driven demand as a key factor behind their performance. Team Group pointed to accelerated AI infrastructure deployment by cloud service providers and structural shifts in upstream memory production as drivers of market conditions.
The company plans to maintain strategic partnerships with memory wafer foundries, focus on edge and high-end computing platforms, and expand into industrial control and system integrator markets. Both firms emphasized a shift in industry demand toward server DRAM and enterprise-grade SSDs as AI workloads increasingly move from training to large-scale inference.
Article edited by Jack Wu
