The divergence between strong corporate earnings and capital expenditure plans from tech firms and their lackluster share price performance is creating confusion among investors. The narrative around the potential end of the AI boom has evolved from a single scenario to multiple interwoven possibilities, significantly increasing market uncertainty.
According to a report from Nomura Securities strategist Naka Matsuzawa on July 17, the cost-benefit balance of AI investment has become unclear. Market participants are now facing several potential "AI boom end scenarios," leading to widespread doubt and a wait-and-see attitude. Concurrently, a clear divergence is visible within the tech sector: semiconductor stocks have fallen sharply, the MAG7 group is largely flat, while software stocks have shown relative strength. This indicates a quiet repricing of capital across subsectors.
Evolving Scenarios for the AI Cycle's End
Previously, the market primarily evaluated three potential inflection points for the AI investment cycle: first, that overheated AI investment would squeeze hyperscaler cash flows, leading to a slowdown in spending; second, that high memory prices would increase investment costs, triggering a contraction in AI expenditure; and third, that rising raw material costs would exacerbate inflationary pressures, prompting central banks to adopt a more hawkish stance.
However, the Nomura report highlights that the market is now alert to a fourth scenario—where high memory prices lead to overheated semiconductor investment, subsequently causing memory prices to fall. In essence, the cost-benefit that AI brings to the broader economy has become blurred. Semiconductor stocks, which previously benefited from soaring memory prices, appear to have reached an inflection point.
The coexistence of these four scenarios makes it increasingly difficult to clarify the overall economic cost-benefit equation of AI. Matsuzawa believes it is precisely this multi-path uncertainty that is causing market participants to become skeptical and hesitant.
Tech Sector Divergence Amid Strong Fundamentals
A core market contradiction lies in the failure of robust tech company profits and aggressive investment plans to translate into stock price gains, a divergence particularly pronounced recently.
Within the sector, semiconductors and the MAG7 have shown weakness, while software stocks have remained relatively resilient. In the broader U.S. market, defensive and consumer-related stocks have performed better, while technology, capital goods, and bank stocks have generally faced pressure. The VIX index has risen to 16.7, whereas volatility indicators in the bond and foreign exchange markets continue to decline, suggesting the current uncertainty is primarily concentrated in the equity market.
Matsuzawa notes that even if hyperscalers report strong earnings and capital expenditure plans over the next two weeks, predicting market reactions across subsectors remains difficult due to the interference of these multiple scenarios—strong data may not lead to a broad-based rally.
Bond Market Yet to Price in an AI Slowdown
A notable signal is that if the market truly began pricing in the end of the AI boom, bond buying should emerge as investors position for future interest rate cuts. This, however, has not yet occurred.
According to the Nomura report, the U.S. Treasury yield curve is experiencing a bear flattening. The 10-year real yield has rebounded to 2.31%, while the 10-year breakeven inflation rate continues to fall to 2.23%. Market expectations for Fed rate hikes have intensified, with pricing for a 3-basis-point hike in July, a cumulative 14 basis points by September, and 27 basis points by December. The 2-year forward OIS rate, a proxy for the terminal rate, has risen to 3.84%.
This indicates that the bond market's current pricing logic remains dominated by rate hike expectations, not a shift toward betting on future cuts. Based on this, Matsuzawa judges that the market's pricing for an end to the AI boom is incomplete. Investor doubt is still in a fermentation stage rather than having formed a clear consensus.
Comments