AI's wild swings: What are the key breakthrough indicators to watch?

```

In the past year, AI chip and storage manufacturers have almost consistently outperformed cloud computing giants (Hyperscalers), becoming the strongest focus line in the global stock markets. However, a report released by JPMorgan on July 1st points out that such an ever-expanding performance gap is difficult to sustain long-term. Whether the AI boom can continue in the future is no longer determined solely by capital expenditure, but by whether AI commercialization capabilities can truly be realized.

JPMorgan believes that the current AI trade is facing two completely different evolutionary paths: one is that AI cloud companies and model firms gradually improve profitability, catching up with chip companies’ growth and thereby pushing the whole AI industry chain to expand together; the other is that excessive profits in chip companies squeeze downstream profitability, forcing cloud computing giants to reduce capital expenditure and in turn hitting chip demand.

In this process, the market needs to focus no longer just on AI Capex, but on indicators such as AI computing power prices and Token prices, which can directly reflect commercialization capability.

AI trading shows growing fractures

JPMorgan points out that since last September, the US semiconductor sector—especially AI chip and storage companies—has almost consistently outperformed large cloud computing platforms, a performance level that warrants caution.

The report believes that a key catalyst for this round of the semiconductor boom is a significant upward revision in cloud computing giants’ capital expenditure expectations.

According to consensus forecasts summarized by Bloomberg analysts, the five largest cloud computing giants (Google, Amazon, Meta, Microsoft, Oracle) are expected to reach $758.1 billion in capital expenditures this year, up about 100% year on year; further increasing to $925 billion in 2027.

Because of this, the AI chip industry chain is considered the most direct beneficiary in the whole AI investment logic. But JPMorgan emphasizes that as different links in the same AI industry chain, the long-term emergence of such large differences in returns between chip companies and cloud giants is itself unsustainable.

Two completely different outcomes for the AI boom

The report believes that this gap will eventually be bridged, with two possible ways.

The first, which JPMorgan favors more, is ongoing improvement in AI commercialization.

As cloud computing giants, AI model providers, and end users gradually increase the monetization ability of AI products, with revenues and profits consistently improving, they will gain an ever larger value share within the AI value chain and thereby catch up with the performance of chip companies. In this scenario, profits expand across the whole AI industry chain, rather than profits simply being redistributed among different links.

The other is a negative scenario about which the market is currently more concerned.

If the semiconductor companies’ excess returns are ultimately built on the foundation of customers’ profits being continually compressed, cloud computing giants, model companies, and end users might all begin to reduce their willingness for capital expenditure. Once capital expenditure slows, it will directly weaken future chip demand, ultimately causing the previously leading semiconductor sector to face greater correction pressures. JPMorgan points out this is exactly what many customers are currently worried about.

Why is the market beginning to worry?

Cloud computing giants’ stock prices have basically been flat for the past year; although profits remain growing, stock prices haven’t risen in tandem, which means valuations have been compressed and the cost of equity financing is actually rising. Meanwhile, their credit spreads have already risen significantly above those of semiconductor companies, meaning debt financing costs are also increasing.

Another data point concerning the market comes from capital expenditure forecasts. According to consensus expectations, cloud computing giants’ capital expenditure growth rate will visibly slow starting in 2027. It is expected to grow 100% year-on-year in 2026, but only 22% in 2027, with further decrease to single digits in the following years.

Although JPMorgan’s own forecast is more optimistic, expecting Capex to reach $1.15 trillion in 2027, higher than the consensus expectation of $925 billion, if the market’s current predictions are realized, the entire AI trade could face prolonged adjustments, affecting not only the stock market but possibly also the credit market.

The real metric to watch is AI computing power price

In JPMorgan’s view, the most important indicator to watch going forward has already changed.

The report points out that the key to determining the future profitability of cloud computing giants is no longer just the scale of capital expenditure, but the price of AI computing. The higher the price of computing power leasing, the easier it is for cloud platforms to increase AI service income, and also to maintain or even expand their profit margins, thereby supporting ongoing large-scale capital investment.

The report shows that AI computing power prices were under downward pressure up to the end of 2025, but saw some improvement between April and May this year, with a slight retreat in June. Meanwhile, large language model (LLM) Token prices continued to rise from late February to late May, cooling off somewhat in June as well.

JPMorgan believes these two price indicators can directly reflect AI commercialization progress and are the key variables for judging whether the AI investment logic can persist in the future.

Risk alert and disclaimer termsThe market has risks, and investment needs to be cautious. This article does not constitute personal investment advice, nor does it take into account the special investment objectives, financial situation or needs of individual users. Users should consider whether any opinions, views or conclusions in this article are suitable for their specific circumstances. Invest at your own risk. ```