Core Viewpoint - Transsion Holdings is expected to experience a decline in both revenue and net profit in 2025, with net profit projected to be halved, marking the worst performance since its IPO in 2019 [1][5][6]. Financial Performance - The company forecasts a revenue of approximately 65.568 billion yuan for 2025, a decrease of about 31.47 billion yuan or 4.58% year-on-year [2]. - The expected net profit for 2025 is around 2.546 billion yuan, down by approximately 30.03 billion yuan or 54.11% compared to the previous year [2][5]. Market Competition - Transsion remains the market leader in Africa, but faces increasing competition from Xiaomi and Honor, which have shown growth rates of 34% and 158% respectively in the third quarter of 2025 [1][8]. - The overall smartphone market is experiencing sluggish growth, particularly in emerging markets, with a projected global smartphone shipment growth of only 2% in 2025 [7][8]. Cost Pressures - The company attributes its poor performance to rising storage prices, which have increased by over 40% due to demand from AI data centers, impacting product costs and gross margins [7][8]. - Transsion's average smartphone price in 2025 is reported to be 332.1 yuan, with feature phones averaging only 50.1 yuan, indicating a focus on the low-end market [9][10]. Diversification Efforts - To mitigate risks, Transsion is exploring new business areas such as mobility and energy storage, although these segments currently contribute only 8.8% to total revenue [12][14]. - The company has established a mobility division and is expanding into electric two-wheelers and energy storage products, but these initiatives are still in early stages and have not yet significantly impacted overall revenue [11][12].

作者 James Foord ，内容为转载编译，仅为呈现不同市场观点与研究视角，并不意味着本公 众号对文中观点结论认可。 过去12个月，美光暴涨逾250%。 其上涨逻辑简洁且极具说服力。 人工智能模型需消耗海量高速DRAM与NAND闪存，随着越来越多企业部署人工智能基础设 施，美光有望抢占这一领域的大量支出份额。 换句话说，内存需求将凌驾于计算需求之上，因此，未来优先采用"内存优先"设计的架构将更 具优势。 内存芯片市场 关键在于， 当前内存领域存在真实的供给瓶颈——这意味着美光不仅销量提升，还能以更高价 格出货。 这一点从其云内存业务100%的增长率中可得到印证。 【如需和我们交流可扫码添加进社群】 在人工智能推理时代，这一趋势尤为明显——人工智能的迭代升级建立在过往知识积累之上， 而这些知识必须依靠内存存储。 随着人工智能模型复杂度提升，内存占用量的增长速度将超过计算需求。 | Amounts in millions | FQ1-26 | FQ4-25 | Q/Q % Change | FQ1-25 | Y/Y % Change | | --- | --- | --- | --- | --- | --- ...