The latest batch of global purchasing managers' indices paints a picture that is mixed on the surface but coherent underneath. Taken in aggregate, global growth is improving, and the improvement is being led by manufacturing. The engine driving that manufacturing strength is unmistakable: the boom in capital expenditure tied to artificial intelligence. This spending wave is now spilling over into corporate earnings estimates, which are rising across much of the world as a direct result.
Yet the most important feature of this expansion is not its breadth but its narrowness. While there are early signs of the rally broadening into a wider set of companies, technology remains the overwhelming beneficiary. Recent earnings strength has clustered in names like HPE, Intel, Cisco, and Dell—all technology companies. The concentration is stark in the numbers: technology accounts for nearly half of expected earnings growth in the United States this year and more than 60% of expected earnings growth across emerging markets. This is, in other words, a very focused and concentrated increase, not a generalized boom.
South Korea: Strong Fundamentals, Fragile Foundations
Nowhere is the AI trade more visible than in South Korea. The country's benchmark index has posted a year-to-date gain of more than 100%, vaulting it past markets like Canada and India in terms of global market capitalization. The drivers are real: South Korea is benefiting directly from the AI capex boom and from a shortage of memory chips, which has pushed chip prices higher and lifted earnings expectations dramatically—on the order of 20% to 250% for parts of the market.
But improved fundamentals tell only half the story, and the other half demands caution. Despite these brightening prospects, foreign investors have been net sellers for eight straight weeks. The market's continued ascent is increasingly being powered by domestic retail investors—and they are buying through notably speculative channels. They are using margin and single-stock leveraged ETFs, and their attention is fixated on just two companies operating in the same memory-chip space: SK Hynix and Samsung Electronics.
The signs of froth are hard to ignore. Newly issued single-stock leveraged ETFs raised $4 billion in just four days after inception. Daily turnover in these vehicles runs at multiples of the number of shares actually issued—a hallmark of speculation rather than investment. This combination of speculation and concentration translates directly into elevated risk. It may not take much to trigger a decline, and anyone participating in this market should tread carefully.
China: A Head Start Constrained by Hardware
China presents a different and more nuanced case. There has been considerable commentary on the country's apparent head start in AI, evidenced by the wave of large language model developers that have gone public in venues like Hong Kong over the past year. With Anthropic having filed confidentially for a public listing and OpenAI's listing still awaited in the West, the contrast invites comparison.
China is a genuinely strong contender for AI leadership, and consumer adoption there is robust. The DeepSeek episode demonstrated that Chinese firms can achieve results more cheaply and efficiently than some of their US counterparts, hinting at meaningful cost advantages. But the country faces a hardware bottleneck: computing shortages are causing frequent service disruptions. Closing the gap will likely require both additional investment and further advances in chip and chip-equipment technology before China can realize its leadership ambitions.
There is, however, a silver lining that is rare in this kind of market. Strong consumer demand for AI has handed some Chinese companies genuine pricing power—an unusual and welcome development. That pricing power could supply the capital needed to fund innovation and could eventually push earnings estimates higher. This matters because, unlike most markets around the world, earnings estimates in China have actually been declining this year. Chinese technology has struggled considerably more in this period than it did the year before, which is precisely why a turn toward pricing power and improving estimates would be so significant.
Europe: Lagging and Looking for Validation
Europe occupies the back of the pack in this global tech narrative. The continent lags in technology innovation, and the fundamental problem is one of scale and quantity: it simply does not have enough technology companies, and the ones it does have are not large enough to lift the broader economy and stock market. The point is captured by a telling anecdote—the biggest IPO in London last year was, surprisingly, a tuna cannery. There are bright spots, including chip-related manufacturers tied to lithography equipment such as ASML, but they are too few and too isolated to change the overall picture.
That is why recent developments carry symbolic weight. A major SoftBank investment in France—reportedly its largest—offers a measure of validation for European technology, and a semiconductor rally sparked by STMicroelectronics provided a further glimmer. As Masayoshi Son has framed it, this moment represents an opportunity for Europe to play catch-up with the United States and China. For now, though, investor attention is rationally drawn toward Asian markets that have more participants and a structural advantage, and toward US companies sitting at the genuine forefront of innovation.
Conclusion
The AI capital-expenditure boom is real, and it is lifting growth and earnings across the globe. But its benefits are distributed unevenly and concentrated dangerously. The same forces that have produced triple-digit gains in South Korea have also created conditions ripe for volatility. China holds real promise constrained by hardware limits and softening earnings, while Europe watches from behind, hoping that fresh investment marks the beginning of a catch-up rather than a footnote. For investors, the lesson is to separate the durability of the underlying trend from the fragility of the trades built on top of it—and to recognize that concentration cuts both ways.