NextFin News——Transion Holdings将它命名为非洲大部分地区的主导电话品牌和更广泛的新兴市场——中国消费者电子学的罕见成功几乎完全建于海外。 收入和利润在2023年和2024年达到顶峰。 2025年，记忆-价格通胀的激烈循环将净利润大约减半。
储存芯片和抬高售价导致2026年上半年在容易比较的情况下出现反弹；第三季度再次失速。 由于电话内存有时比主处理器成本更高，公司正在推动端侧 AI以寻求定价权。 其香港上市工作于10月中旬完成，但交易很薄，对上海股票的折扣很高，对下一章很难投出信任票。
10月30日的第三季度成绩发布显示前半段修复叙事淡出。 香港线路(HK2636)大多坐落在发行价格附近，发售量轻，营业额低，对A股几乎有50%的折扣——这是经过长征进入A+H结构后的尴尬结果。
如何建立一个低成本的出口冠军
转会是中国手机品牌中的外延。 总部位于深圳，主要在中国进行设计和建设，但销售远离家乡。 创始团队出自伯德；控股股东朱昭江在2006年离开非洲投注之前，曾经营过该公司的国际业务，而非洲的渗透率低，全球品牌参与度低，价格比声望重要。
游戏手册缩小了：实用特征，如多SIM支持、电池寿命长和照相机调频，用于更深的皮肤色调；三个品牌——itel、TECHNO和Infinix——覆盖进入中层；以及2 900多个经销商从国家伙伴到小城镇的分布网。 非洲成为了本土市场；类似的战术在南亚，东南亚，中东和拉丁美洲传播。 在2019年上海STAR上市时，它已经被称为非洲电话王。 IDC从此将它列为世界前三名发货商之一；其非洲销量份额一度超过60%的第三方数据。
2023年带来了利润猛增；2024年带来了创纪录的收入，接近690亿人民币，净利润几乎没有增长，基本利润下降——纯净数额接近上限。 然后供给震荡。
当AI重订记忆市场时
从2025年初开始，DRAM和NAND的价格攀升，因为生产者将能力引导到AI数据中心的高带宽内存和服务器产品，收紧了实际使用的零件电话的供应。 据研究室估计，低端智能手机的存储量大约占40%至50%
成本下降。 瓦费尔制造商和模块房可以最快地重新定价；针对价格高度敏感的买家的品牌不能。 Transsion的混音大部分在150美元以下，2025年智能手机平均售价约为566人民币(约80美元). 那个位置让它暴露了 2025年全年收入跌幅约5%，约660亿元；净利润跌幅约53%，约25.8亿元；毛差跌幅近19%；股价从2025年后期的峰值下跌40%以上。 投资者后来对危机期间的沟通和高层次的股票销售提出了管理压力；答案仍然是公式化的。
以往的暴风雨——货币压力、从专题电话的转变、大流行病的短缺——一直以灵活的来源加以控制。 分析家认为，仅靠普通的定价和频道调整就更难解决。
库存，提高价格，攀登混合
从2025年末开始，Transsion大力建设库存：年末约89亿元，到2026年3月达到142亿元，到6月达到189亿元。 管理称材料大约是库存的一半，主要是记忆。 它曾试图购买更接近来源的货款——与生产商谈判死亡，然后向模块公司付款组装——因此它并不总是购买完全定价的货包。 在移动平均成本计算下，昂贵的新地块与更便宜的新地块混合在一起，因此报告的成本增长滞后。 现金不等待：上半年运营现金流量约59亿元，前9个月运营现金流量约99亿元。
购买价格仍然随市场浮动。 平均内存采购费用从2023年的约50元人民币猛增到2025年的约70元人民币，2026年头4个月达到192元人民币。 记忆成为销售商品中最大的单一成本，领先于处理器和显示器。 有关供应商的小额股份和基石订单意在将Transion保留在前线附近。
在缓冲器买下时，公司融资价格每季都有新车型的浪潮，并削减了最便宜的装置。 价格高于150美元的单位——其内部的中高等级——从先前的一位数字计算到上半年大约16-17 %的销售额。 智能手机货运下降约11%，而平均销售价格上升约30%.
反弹没有持续
2026年上半年对一个薄弱的基数看似强劲：收入增加22%，净利润增加46%，总比值超过22%。 与2024年相比，收入仅出现倾斜，利润仍然低得多。 第三季度抵消了大部分增长：收入逐年下降四分之一以上；九个月的收入仅比一年前高一点。 头条净利润比核心利润保持得更好，核心利润急剧下降；非经营收益在可归属数字中占很大份额。 公司抽走现金和金融资产，增加短期借款，但仍大量用于商品。 到9月，库存从年中开始略有减少，但仍远高于2025年底的水平。 管理称，随着内存价格的上涨速度减缓，它将更有选择地储存。
AI的故事遇到了怀疑的市场
上市将Transion从它最熟悉的低成本高密度频道游戏中拉出，并进入Samsung，小米，Honor等在品牌，技术和融资报价上更强的层次。 风险在底部正在减少，而不保证上面的径。
对投资者的叙述以端侧 AI为中心：一大笔香港IPO收益用于AI研究；拟运行于中低层芯片的紧凑模型；当地语言语音以及实用相机和呼叫功能；硬件与Transsion自己的软件平台之间更紧密的联系；以及辅助配件、电器和能量存储——相对于近3亿月用户的移动间网络基础。 IPO前会议挤满。 对交易的要求没有。 报价在幅度底端定价，严重依赖基石投资者——包括供应链伙伴——并在下期开放。 贸易一直保持安静，对A股有宽大的折扣。
这种差距捕捉到疑点：本地化AI可能会改进产品，但是它没有被证明是用来固定利润模型的，它仍然被捆绑在存储成本曲线最糟糕的部分的手机上。 转销仍然有分布范围，有本地产品感，在低价点有规模。 这些优势受到部分通货膨胀和较强硬对手的压力。 物价进一步上涨，需求可能破解；条纹规格和产品削弱。 AI已不再是一个附带项目，而是决定公司能否超越纯体量。 现在，记忆力膨胀仍在决定节奏。
* (本文首次发表于TMTPost App. Series: Capital Profile. 作者：温一道。 )*
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
NextFin News — Transsion Holdings made its name as the dominant phone brand across much of Africa and a wider set of emerging markets—a rare Chinese consumer-electronics success built almost entirely overseas. Revenue and profit peaked in 2023 and 2024. In 2025, a fierce cycle of memory-price inflation cut net profit roughly in half.
Stockpiling chips and lifting selling prices produced a rebound in the first half of 2026 on an easy comparison; the third quarter lost speed again. With phone memory sometimes costing more than the main processor, the company is pushing on-device AI in search of pricing power. Its Hong Kong listing, completed in mid-October, has traded thinly and at a steep discount to the Shanghai shares—hardly a vote of confidence in that next chapter.
Third-quarter results released on October 30 showed the first-half repair narrative fading. The Hong Kong line (HK2636) has mostly sat near its issue price, with light volume, low turnover, and almost a 50% discount to the A-share—an awkward outcome after a long march to an A+H structure.
How a low-cost export champion was built
Transsion is an outlier among Chinese phone brands. Headquartered in Shenzhen, it designs and builds largely in China but sells far from home. The founding team came out of Bird; controlling shareholder Zhu Zhaojiang had run that firm’s international business before leaving in 2006 to bet on Africa, where penetration was low, global brands were thinly engaged, and price mattered more than prestige.
The playbook scaled: practical features such as multi-SIM support, long battery life, and camera tuning for darker skin tones; three brands—itel, TECNO, and Infinix—covering entry to mid-tier; and a distribution web of more than 2,900 dealers from national partners down to small towns. Africa became a home market; similar tactics spread through South Asia, Southeast Asia, the Middle East, and Latin America. By its 2019 Shanghai STAR listing it was already called Africa’s phone king. IDC has since ranked it among the world’s top three shippers; its African volume share once exceeded 60% in third-party data.
2023 brought a profit surge; 2024 brought record revenue near 69 billion yuan, with net profit barely growing and underlying profit down—signs that pure volume was approaching a ceiling. Then the supply shock hit.
When AI reordered the memory market
From early 2025, DRAM and NAND prices climbed as producers steered capacity toward high-bandwidth memory and server products for AI data centers, tightening supply of the parts phones actually use. Research houses estimate memory now accounts for roughly 40–50% of the bill of materials in low-end smartphones—and more than 60% in sub-$100 devices—up from the low teens in calmer years.
Cost flows downhill. Wafer makers and module houses can reprice fastest; brands aimed at highly price-sensitive buyers cannot. Transsion’s mix sits mostly under $150, with 2025 smartphone average selling prices around 566 yuan (about $80). That positioning left it exposed. Full-year 2025 revenue slipped about 5% to roughly 66 billion yuan; net profit fell about 53% to 2.58 billion yuan; gross margin hit a listed low near 19%; the share price dropped more than 40% from its late-2025 peak within months. Investors later pressed management over communication during the crisis and over high-level share sales; answers stayed formulaic.
Past storms—currency stress, the shift from feature phones, pandemic shortages—had been manageable with flexible sourcing. This one, analysts argue, is harder to fix with ordinary pricing and channel tweaks alone.
Stockpile, raise prices, climb the mix
From late 2025, Transsion built inventory aggressively: from about 8.9 billion yuan at year-end to 14.2 billion by March and 18.9 billion by June 2026. Management said materials were roughly half of that stock, mostly memory. It has tried to buy closer to the source—negotiating die with producers, then paying module firms to assemble—so it is not always purchasing fully priced packages. Under moving-average costing, expensive new lots blend with cheaper ones, so reported costs rise with a lag. Cash does not wait: operating cash flow ran about −5.9 billion yuan in the first half and −9.9 billion for the first nine months.
Purchase prices still floated with the market. Average memory procurement cost jumped from roughly 50 yuan in 2023 to about 70 yuan in 2025 and 192 yuan in the first four months of 2026. Memory became the largest single cost in goods sold, ahead of processors and displays. Small stakes and cornerstone orders in related suppliers were meant to keep Transsion near the front of the line.
In the time that buffer bought, the company raised prices with each quarterly wave of new models and cut back on the cheapest devices. Units priced above $150—its internal mid-to-high tier—rose to about 16–17% of sales in the first half, from single digits before. Smartphone shipments fell about 11%, while average selling price rose about 30%.
A rebound that did not last
First-half 2026 looked strong against a weak base: revenue up 22%, net profit up 46%, gross margin above 22%. Compared with 2024, revenue had only edged up and profits remained far lower. The third quarter erased much of the growth: revenue down more than a quarter year on year; nine-month revenue barely higher than a year earlier. Headline net profit held up better than core profit, which fell sharply; non-operating gains accounted for a large share of the attributable figure. The company drew down cash and financial assets and increased short-term borrowing while still spending heavily on goods. By September, inventory had been trimmed slightly from midyear but remained far above end-2025 levels. Management said it would stockpile more selectively as the pace of memory price increases slowed.
The AI story meets a skeptical market
Moving upmarket pulls Transsion out of the low-cost, high-density channel game it knows best and into a tier where Samsung, Xiaomi, Honor, and others are stronger on brand, technology, and financing offers. The risk is losing volume at the bottom without securing the rung above.
The narrative to investors has centered on on-device AI: a large slice of Hong Kong IPO proceeds for AI research; compact models meant to run on mid- and low-tier chips; local-language speech and practical camera and call features; tighter ties between hardware and Transsion’s own software platform; and adjacency into accessories, appliances, and energy storage—set against a mobile-internet base of nearly 300 million monthly users. Pre-IPO meetings were crowded. Demand for the deal was not. The offer priced at the bottom of the range, relied heavily on cornerstone investors—including supply-chain partners—and opened below issue. Trading has stayed quiet, with a wide discount to the A-share.
That gap captures the doubt: localized AI may improve the product, but it is unproven as a fix for a profit model still tied to handsets stuck in the worst part of the memory cost curve. Transsion still has distribution reach, local product sense, and scale at low price points. Those advantages are under pressure from component inflation and tougher rivals. Raise prices further and demand may crack; strip specifications and the product weakens. AI is no longer a side project—it is the wager on whether the company can grow beyond pure volume. For now, memory inflation is still setting the tempo.
*(This article was first published on the TMTPost App. Series: Capital Profile. Author: Wen Yidao.)*
Auto-fetched and translated by AI Atlas. All rights belong to the original author.