[Big Bank View] Guotai Securities: AI U.S. stock’s estimated P/E isn’t overheated; the market is still working to find flaws in the AI narrative

HSBC (00005) Global Investment Research stated that the current forward PE ratios of AI trades do not show any signs of overheating.

HSBC Research published a report, stating that AI trades have been frequently compared to the tech bubble of the late 1990s to early 2000s for months. Although with the rise of agentic AI driving even greater demand for data centers and chips, bubble rhetoric has largely been replaced by supply bottleneck issues, bearish voices still keep coming.

HSBC Research mentioned that last week, the optimistic narrative on AI was once again questioned. From leveraged single-stock ETFs to the overly hawkish stance of the Federal Reserve, many reasons could explain tech stocks weakening again. Micron’s (US: MU) earnings report poured cold water on these narratives, providing concrete evidence that the AI market backdrop remains vibrant and healthy. However, excluding the post-earnings rally of Micron, overall tech stocks did not rebound as expected.

HSBC Research stated that it is cautious about how 'market narratives' drive short-term market movements, 'even if fundamentals do not support changes in investment strategy.' For example, last year the 'AI bubble' rhetoric caused U.S. semiconductor stocks to fall about 15%. But as AI bottlenecks become more evident than bubbles, semiconductor-related ETFs saw a stronger rebound.

HSBC Research said: 'For us to become more worried currently, the market would need to display excessive optimism and positioning. By our measure, current market sentiment remains firmly neutral. A shift in market expectations toward a more dovish U.S. interest rate stance could be another catalyst driving stock market strength – and last week’s data has already prompted the market to take a cautious step in that direction.'

A simple re-rating of valuations back to historical median levels would be sufficient to sustain the strength of AI trades

HSBC Research also mentioned that current forward PE ratios of AI trades do not show any signs of overheating. For example, NVIDIA (US: NVDA) currently has a 12-month forward PE ratio slightly below 20x, at its lowest in a decade. In contrast, Monster Beverage (US: MNST) has the same metric at a 10-year high, with a 12-month forward PE ratio near 40x, forming a stark contrast.

Therefore, even if valuations simply undergo a re-rating back to their historical median levels, that would be sufficient to support the continued strong development of AI trades.

HSBC Research further pointed out that another opposing view is that market expectations for future earnings are too optimistic, and the reality check will eventually arrive. However, the main issue with this view is that for many of these stocks, the realized earnings growth over the past year has already outpaced their price performance. As a result, the trailing PE ratios of Meta (US: META), Amazon (US: AMZN), Microsoft (US: MSFT), NVIDIA, Broadcom (US: AVGO), and Micron have all declined over the past 12 months. Many other sectors in the market are not the same.

Furthermore, earnings growth expectations may not yet have caught up with strong earnings momentum. HSBC Research noted that for many companies at the forefront of the U.S. AI wave, market expectations for their full-year 2026 earnings growth are flat or even lower compared to the annual growth rate seen in the 12 months ending Q2 2025. This remains the case even though the earnings revisions ratio for these companies has rebounded in recent months. 'Therefore, while the market's narrative about AI is still trying to find fault, the "pain trade" may be exactly that AI continues to strengthen in the second half of the year and brings upside surprises beyond expectations.'