盈利讯号已明显恶化，Morgan Stanley 指出，超过 2,000 家 A股公司在 2025年4Q 的业绩预告中出现重大疲弱。负面预警比正面多出 14.8%，明显差于 2025年2Q 净负面 4.8% 的水平，其中小型房地产与消费企业承受的压力最大。宏观数据强化了这一趋势：4Q GDP 同比增长放缓至 4.5%，1月生产者价格下跌 1.4%，制造业与非制造业 PMI 均低于预期；同时监管机构在 1月收紧保证金融资，以遏制投机过度。
各行业前景日益分化，这使选股更困难，也削弱了对全面反弹的信心。大宗商品与政策支持主题表现更好，CMOC 报告全年利润增长约 50%，Iflytek 指引增长 40% 到 70%；而 BYD、Great Wall Motor 等电动车相关公司在 1月销售疲弱后走弱。市场一致预期仍指向 2025年 A股盈利约增长 6.5%，此前 2024年为下滑 3%，但这种改善被视为政策与周期驱动，而非结构性改善；在 3月政策会议之前，短期能见度仍有限，且 1月-2月 数据常因假期因素合并统计。
China’s equity rally has lost momentum by February 15, 2026 as weakening earnings expectations and soft demand challenge the usual Lunar New Year boost. The concern is that 9-day holiday spending will not offset broader uncertainty, especially with stimulus support being scaled back. Market performance reflects this caution: the MSCI China Index is up only 0.8% year to date, trailing the MSCI All World Index at 2.8%, and far behind regional peers with South Korea up 31% and Taiwan up 16%.
Earnings signals have deteriorated sharply, with Morgan Stanley noting major weakness in 4Q 2025 pre-announcements from more than 2,000 A-share firms. Negative alerts exceed positive ones by 14.8%, a much worse balance than the 4.8% net negative seen in 2Q 2025, with smaller real estate and consumer firms under the most pressure. Macro data reinforce the trend: GDP growth slowed to 4.5% year on year in 4Q, producer prices fell 1.4% in January, and both manufacturing and non-manufacturing PMIs missed expectations, while regulators tightened margin financing in January to curb speculative excess.
The outlook is increasingly uneven across sectors, making stock selection harder and reducing confidence in a broad-based rebound. Commodity and policy-supported themes are outperforming, with CMOC reporting about 50% full-year profit growth and Iflytek guiding to 40% to 70% growth, while EV names like BYD and Great Wall Motor weakened after soft January sales. Consensus still points to roughly 6.5% A-share earnings growth for 2025 after a 3% decline in 2024, but the improvement is viewed as policy- and cycle-driven rather than structural, with near-term visibility limited before March policy meetings and with January-February data often combined around the holiday period.