How much has the entire US stock AI sector fallen in the past month?

US AI-related stocks are experiencing a significant summer correction, with valuation pressure visibly easing from previous highs.

According to Zhuifeng Trading Desk, citing Barclays' latest research report released on August 4, since tracking began in February this year, the share prices of the 12 "AI concept stocks" under its coverage have fallen an average of about 25% from their respective peaks, with the implied AI P/E ratios of nearly half of these stocks dropping to historical lows within the tracking period. Meanwhile, the portfolio's average implied AI valuation multiple shifted down by another 17% over the past month, equivalent to a compression of about 9 valuation multiples.

Barclays stated that since its last report, the stock prices in its AI classification have generally shown a trend of small to moderate declines. In contrast, the S&P 500 Index fell by 1% during the same period (July 6, 2026, to July 31), and the Nasdaq 100 Index fell by 5%.

Although the AI sector is under pressure, related stocks have still risen an average of about 60% year-to-date. However, this gain has continued to narrow from previous highs of approximately 70% to nearly 80%. Compared to the S&P 500 Index's 9% year-to-date rise, the AI sector's excess return remains significant, but the correction trend is clear.

Barclays analysts believe that although capital expenditure expectations for hyperscale cloud service providers continue to be revised upward and the quarterly reports of related companies have generally performed reasonably well, actual earnings were insufficient to support further valuation expansion because stock prices had already risen significantly before the earnings reports. This triggered a decline in stock prices and a contraction in implied valuation multiples.

Divergent Declines in Individual Stocks; Optical and EMS Sectors Under Greatest Pressure

Among the AI concept stocks tracked by Barclays, the most severe declines were concentrated in the Electronic Manufacturing Services (EMS) sector and the optical sector. Over the past month, Fabrinet fell 13%, Celestica fell 5%, Jabil fell 7%, and Flex fell 18%; in the optical sector, II-VI fell 13%, while Corning plunged 29%.

Looking at a longer timeframe, from February 11 this year, when Barclays began tracking, to the peak interval ending July 31, the declines from respective highs varied significantly among companies: Corning fell 46% from its peak, Fabrinet fell 42%, II-VI fell 40%, Flex and Super Micro Computer both fell 30%, Dell fell 13%, and Arista Networks had the smallest decline at just 4%.

The Barclays report pointed out that the continued sluggishness of Super Micro Computer's stock price was mainly due to the company's own issues rather than changes in the fundamentals of the AI industry.

Implied AI Valuation Multiples Generally Compress, Some Stocks Approach Historical Lows

Barclays' analytical framework isolates the "core business valuation" of individual stocks to reverse-engineer the market's implied independent valuation multiple for their AI businesses, thereby measuring changes in the AI premium.

The latest data shows that over the past month, the implied AI valuation multiples of most companies contracted significantly. Corning's implied AI valuation multiple dropped sharply from 155x to 76x, II-VI from 75x to 64x, Flex from 36x to 27x, and Fabrinet from 30x to 26x. The average implied AI valuation multiple for the overall portfolio decreased from 52x in the previous period to 43x.

In terms of absolute levels, there is now clear divergence: the implied AI valuation multiples of most EMS companies (Fabrinet, Celestica, Jabil, Flex) are concentrated in the 20x to 30x range; meanwhile, some companies that previously lagged in gains, such as Arista Networks and Hewlett Packard Enterprise, have implied AI multiples in the relatively high range of 45x to 70x. Due to specific company issues, Super Micro Computer's implied AI multiple remains at the bottom, currently at only about 2x.

Barclays pointed out that although they have fallen significantly from previous highs, the implied AI valuation multiples of nearly half of the AI concept stocks are at stage lows within the tracking period. However, these multiples remain significantly higher than the "core business" valuation levels of each company, meaning the market still assigns a considerable premium to AI businesses.

AI Business Fundamentals Remain Strong; Cloud Revenue Growth Forecasts Revised Upward

Notably, this valuation correction does not stem from deteriorating fundamentals. The Barclays report shows that as hyperscale cloud computing vendors continue to raise capital expenditure expectations, the bank has revised up its forecast for the compound annual growth rate of cloud AI revenue (2023–2027) for the aforementioned AI concept stocks from an initial estimate of about 53% to about 65%, which is also higher than the previous forecast of about 62%.

Regarding recent growth expectations for individual companies, Dell's CY26 cloud AI revenue growth forecast is as high as 144%, Cisco's is 127%, and Corning's is 90%, all belonging to the high-growth camp. Barclays estimates that this batch of AI concept companies under its coverage accounts for approximately 20% of global cloud/AI capital expenditures, with related businesses showing substantial revenue growth.

Non-AI Sectors Quietly Strengthen, Signs of Capital Rotation Emerge

In sharp contrast to the widespread correction in AI concept stocks, non-AI companies tracked by Barclays saw their stock prices rise by low double-digit percentages on average during the same period (July 6 to July 31), significantly outperforming the S&P 500.

Specifically, Garmin rose 27%, Nutanix rose 27%, Telesat rose 25%, Motorola Solutions rose 9%, and Apple rose 4%. Barclays believes that as investors enter the second half of the year, their interest in previously "overlooked" companies is rebounding. These companies have solid fundamentals but failed to simultaneously enjoy the benefits of stock price increases and valuation expansion during the first half of the year's AI rally, thus becoming targets for capital seeking "laggard" opportunities.