Wall Street’s artificial intelligence boom has a circularity problem.
For weeks, I have been warning that too much capital inside the US AI complex is moving around a remarkably tight circle. Suppliers finance customers. Those customers use the capital to buy the suppliers’ hardware. The spending becomes revenue, valuations rise, more capital becomes available, and the cycle begins again.
The concern is not that AI demand is fictitious. Far from it. Artificial intelligence is already reshaping industries and will remain one of the defining investment themes of this decade.
The problem is determining how much apparent demand represents genuine end-user consumption and how much is being amplified by financing structures within the ecosystem itself.
Investors need a way to diversify that risk.
China’s emerging AI hardware export boom could provide one.
A new generation of Chinese manufacturers is pushing into international markets with AI-enabled industrial equipment, robotics, data-centre hardware and automated production systems.
This matters because exports provide investors with a different kind of evidence.
A robot shipped overseas and installed on a factory floor is difficult to confuse with financial engineering. Hardware crossing borders, customers taking delivery and overseas order books expanding represent economic demand that can ultimately be tested against customs data, production volumes and corporate revenues.
China is, in effect, entering a third major export era.
The first was built around low-cost manufactured goods. The second was dominated by technologies including solar panels, batteries and electric vehicles. The next could increasingly be defined by intelligent industrial hardware.
And the addressable opportunity is substantial.
Recent analysis of the emerging exporters points to individual overseas markets ranging from roughly $12 billion to more than $200 billion by 2030. With access to the US constrained by trade restrictions, Europe and Southeast Asia are becoming increasingly important battlegrounds.
That creates an investment proposition very different from simply buying another company attached to the American AI infrastructure cycle.
Industrial automation and robotics are particularly important because these markets will expose which Chinese companies possess genuine international competitiveness.
Currency movements can help exporters at the margin, but they will not decide the winners. Execution will.
Manufacturers need overseas distribution, servicing capabilities, local customer relationships and products that can compete against established international suppliers. Companies that build that infrastructure quickly can capture meaningful market share. Those that cannot will discover that a large theoretical addressable market means very little.
This distinction is already visible among listed Chinese automation and robotics businesses. Some have established meaningful scale in Europe, while others are attempting to build service and distribution networks across Southeast Asia.
The investment question therefore becomes considerably more useful than simply asking whether Chinese exports will rise.
Which companies can turn technological capability into contracted overseas revenue?
Production volumes offer another reason investors should pay attention.
Chinese manufacturers are already demonstrating their ability to industrialise technologies that elsewhere remain relatively small-scale. In humanoid robotics, for example, leading Chinese manufacturers have reported shipment volumes measured in thousands of units, while prominent US competitors remain at a much earlier stage of commercial deployment.
Scale matters.
It lowers costs, deepens supply chains, accelerates product iteration and gives manufacturers the operating experience needed to compete internationally.
This week’s World Robot Conference in Beijing should provide another indication of how quickly the ecosystem is developing. Investors should pay particular attention not to spectacular demonstrations or ambitious announcements, but to evidence of commercialisation: orders, overseas customers, production targets, distribution agreements and recurring service revenues.
These are the metrics that separate an investable industrial trend from a technology showcase.
There is also a broader portfolio lesson here.
Many investors have accumulated enormous exposure to the US AI complex through semiconductor companies, hyperscalers, infrastructure providers and the wider technology indices. The extraordinary performance of those assets has made that concentration easy to tolerate.
But concentration becomes most dangerous precisely when it feels most comfortable.
If parts of the American AI ecosystem are increasingly dependent on suppliers financing customers who then spend that money back inside the same ecosystem, investors should want exposure to demand generated through different channels.
Chinese AI hardware exports can offer that distance.
This is not an argument for abandoning US artificial intelligence leaders, nor is it a blanket endorsement of Chinese equities. China carries substantial geopolitical, regulatory, governance and trade risks, and investors need to price those risks appropriately.
It is an argument for diversification within the AI investment theme itself.
There is an important difference between diversifying away from AI and diversifying the sources of AI-related revenue to which a portfolio is exposed.
The second is becoming increasingly important.
Investors should therefore watch the order books emerging from Beijing and from Chinese exporters over the coming quarters.
If manufacturers can translate domestic scale into sustained overseas orders across robotics, automation and industrial hardware, the consequences will extend far beyond China.
They will provide evidence that the next phase of the AI investment cycle is broadening from chips, models and data centres into the physical economy.
And for investors concerned about Wall Street’s increasingly circular AI trade, real machines being sold to real customers in real overseas markets could prove to be a valuable hedge.