China’s export data provides more tangible validation than market sentiment alone. High-tech product exports reached 5.25 trillion yuan in 2025, an increase of 13.2% year over year. Exports of special-purpose equipment, high-end machine tools and industrial robots rose 20.6%, 21.5% and 48.7%, respectively. Industrial-robot exports also surpassed imports, making China a net exporter in that category, according to China’s government.
The momentum continued into 2026. China’s exports rose 23.9% year over year in dollar terms in July, with Reuters reporting that the global AI infrastructure boom helped drive demand for high-tech goods. Exporters also accelerated shipments ahead of higher US tariffs, which means part of the increase may reflect timing rather than a permanent change in demand.
China is also selling more into developing markets. Exports to Africa rose 25.8% in 2025, while exports to ASEAN increased 13.4%, helping manufacturers diversify beyond traditional markets.
ChangXin Memory Technologies, or CXMT, has become the clearest semiconductor symbol of China’s hard-tech ambitions. The memory-chip maker raised about $8.6 billion in its Shanghai IPO, and its shares rose by more than 500% on debut, briefly giving it a market value of roughly $539 billion.
Investors are effectively backing three linked ideas: AI will sustain demand for memory, China needs domestic DRAM capacity, and policy support will continue to protect strategically important chipmakers. Hefei government-linked investors own 36.8% of CXMT, illustrating how closely public capital is tied to the company’s development.
But the IPO is not proof that CXMT has closed the technology gap. Analysts have highlighted its limited access to leading lithography equipment and the difficulty of competing with established global memory producers. The central test is therefore operational: whether CXMT can improve advanced-memory capabilities, scale production and generate durable profits rather than simply command a strategic premium.
Robotics offers a similar but earlier-stage story. Chinese companies are benefiting from a dense component ecosystem, large manufacturing customers and strong policy interest in embodied artificial intelligence. Unitree has become a prominent symbol of enthusiasm around humanoid and quadruped robots, while the wider robotics market is drawing substantial capital and attention.
The caveat is important: product demonstrations, financing and high valuations do not establish mature deployment economics. Investors still need evidence of repeat orders, production costs, margins and reliable commercial use. Robotics may become a major hard-tech export opportunity, but the available evidence does not justify treating Unitree alone as proof of broad foreign-investor inflows or sector-wide profitability.
China’s hard-tech rally should not be confused with a generalized improvement in foreign capital flows. Actual utilized inward FDI fell 9.5% in 2025 to 747.69 billion yuan, even as the number of newly established foreign-invested enterprises rose 19.1%.
That divergence suggests that companies continue to explore opportunities in China, particularly in selected high-tech and services segments, but are committing capital more cautiously. Concerns include weak demand, regulation, geopolitics, trade barriers and uncertainty over returns.
Meanwhile, Chinese companies are expanding overseas. Outward direct investment rose 7.1% in 2025, and overseas mergers and acquisitions increased by nearly 40%, according to EY. That points to a strategy in which Chinese firms build foreign markets and supply chains directly, rather than relying only on investment into China.
The hard-tech economy is developing against a difficult domestic backdrop. The World Bank projects China’s growth to moderate in 2026 as consumption remains subdued, property-sector adjustment continues and private investment faces low profitability.
The IMF has likewise called for stronger support for consumption and the property sector, with less reliance on inefficient investment. Its recommendations include more effective social spending, further monetary easing and a more efficient property adjustment.
That is the decisive policy test. Continued support for chips, AI, robotics and advanced manufacturing can produce technological progress and export growth. But a durable investment recovery would require that progress to be matched by stronger household demand, clearer treatment of private and foreign firms, credible property-sector restructuring and action against destructive overcapacity.
Global confidence in China’s hard-tech industries is rising for real reasons: AI-linked demand is lifting exports, industrial ecosystems are improving and policy has helped strategic companies reach scale. Yet the data describe a selective technology rally—not an unqualified return of foreign confidence in the Chinese economy.
CXMT captures the opportunity and the risk. Its market debut shows that investors are willing to pay for semiconductor self-reliance and AI exposure. Its equipment constraints and competitive challenge show why future returns will depend less on headlines and more on execution, margins and access to global markets.