For China's tech giants, the conversation around AI has shifted from celebrating new model launches to confronting the massive bill for those ambitions.
Investor focus is now squarely on whether Tencent and Alibaba can absorb the sharply rising costs of the AI arms race. Tencent, for instance, disclosed a staggering RMB 18 billion in AI-related costs for 2025 and plans to more than double that investment in 2026. Similarly, Alibaba's earnings per share are projected to dip significantly in the near term. The era of hyping up AI capabilities is giving way to the stark reality of its impact on the profit and loss statement.
This financial pressure stems from a perfect storm of factors. First, the competitive landscape has been radically altered. The arrival of powerful, open-source models like DeepSeek V4 has been a game-changer. By offering state-of-the-art capabilities at very low prices, it severely compresses the potential for profit from AI services. This forces major players like Tencent and Alibaba to spend more just to keep up, while the path to monetization becomes narrower.
Second, there's a pricing whiplash happening in the cloud market. After slashing prices in 2024 to win customers, these companies are now forced to raise them for AI-specific computing services due to surging hardware costs. However, with competition so fierce, they can't raise prices enough to fully offset their massive capital expenditures on things like high-end Nvidia GPUs, creating a painful margin squeeze.
Finally, geopolitical and regulatory risks add another layer of uncertainty. The ongoing US-China tech tensions create constant procurement headaches for essential AI chips like the Nvidia H200, leading to delays and higher costs. For Tencent, the threat of the US government forcing it to divest its valuable stakes in American gaming companies like Riot and Epic Games creates a significant overhang on one of its most profitable global businesses.
In short, 2026 is the year the AI bill comes due. While the long-term promise of artificial intelligence remains, the immediate challenge for China's tech leaders is navigating the costly and uncertain path to get there.
- Foundation Models: Large-scale artificial intelligence models trained on vast amounts of data, designed to perform a wide variety of tasks. Examples include GPT-4 and DeepSeek V4.
- Capex (Capital Expenditure): Funds used by a company to acquire, upgrade, and maintain physical assets such as property, plants, buildings, technology, or equipment.
- P/E Ratio (Price-to-Earnings Ratio): A valuation ratio of a company's current share price compared to its per-share earnings. It is used by investors to gauge the relative value of a company's stock.