Has the market misjudged AI? The backlog of cloud business orders surged 150% to $17 trillion, and Wall Street is calling out: tech stock valuations have plummeted to lows.

Concerns regarding AI capital expenditures have persisted for more than a year, but the latest data is changing this narrative. Over the past year, investors have worried that the tech giants' investment of hundreds of billions of dollars into AI infrastructure might face the issue of "high input, low return." However, with the acceleration in cloud business demand, more signs indicate that AI capital expenditures are translating into stronger orders and future revenue growth, rather than the previously feared over-investment. Data shows that the backlog of cloud business orders from hyperscale cloud service providers surged over 150% year-on-year, reaching approximately $1.7 trillion, with growth significantly outpacing the roughly 80% increase in capital expenditures during the same period. JPMorgan Chase pointed out that this notable gap suggests that the potential revenue returns from AI infrastructure investments are surpassing market expectations, and the pressure on tech giants' valuations may be nearing an end. In this context, JPMorgan noted that the market is reassessing the return potential of AI infrastructure investments, and the valuation multiples of the Magnificent Seven may have hit bottom. AI investment has shifted from a cost to a source of revenue, with cloud giants' demand exceeding expectations. Signals released during the second-quarter earnings season indicate that cloud computing demand is rapidly materializing. JPMorgan analyst Mark Schilsky stated that the growth rates of cloud business backlog orders and net new Annual Recurring Revenue (ARR) are significantly outpacing capital expenditure growth, suggesting that future revenue growth is expected to cover current large-scale infrastructure investments. Amazon.com, Inc. CEO Andy Jassy, in a rare move, raised the long-term outlook for AWS during the second-quarter earnings call. He stated that the company had previously expected AWS to grow into a business with revenues in the hundreds of billions of dollars but now believes this scale will at least double and could potentially reach $1 trillion in annual revenue in the future. Jassy also noted that the demand scale for 2028 is "stunning," while enterprise clients' large-scale adoption of AI inference services remains at an early stage. Management from companies such as Microsoft Corporation (MSFT.US), Alphabet (GOOGL.US), and Meta (META.US) have also conveyed similar signals: the commercialization of AI applications is still in the early stages of expansion, and enterprise demand has not yet reached maturity. Earnings expectations have been revised upward, yet valuations have fallen to historic lows. While the fundamentals of AI are continuously improving, tech stock valuations have experienced significant compression. After the market adjustment in July, the forward price-to-earnings ratio for the S&P 500 Information Technology sector dropped to about 20 times, close to its lowest level in the past year and within the first percentile of historical valuation ranges, below the average of around 23 times over the past decade. This indicates a rare divergence in the tech sector: earnings expectations are continually improving, yet valuation multiples are consistently declining. JPMorgan pointed out that the forward price-to-earnings ratio of large-cap tech stocks (excluding semiconductors) is currently more than two standard deviations below the historical average since 2018. If valuations are restored to one standard deviation below the historical average, there is about 30% upside potential; if they return to the long-term average, potential upside may reach around 56%. Meanwhile, performance of hyperscale cloud vendors relative to the S&P 500 has also dropped to near the bottom of the range over the past three years, and historically, similar positions have often been accompanied by strong mean reversion opportunities. On the other hand, low valuations reflect that institutional investors have not yet fully aligned their allocations with fundamental changes. Data from Deutsche Bank Aktiengesellschaft shows that, despite significant improvements in earnings forecasts for large tech companies, institutional investors' positions in this sector remain at a slightly overweight level, markedly lower than the allocation levels seen during past strong earnings cycles. At the same time, funding has continued to concentrate in the semiconductor sector this year, while holdings in large tech stocks (excluding semiconductors) remain relatively insufficient. JPMorgan believes that if the market's core narrative around AI shifts from "are capital expenditures excessive?" to "are investment returns being realized?", the next phase of tech stock gains may be driven more by internal sector rotation, rather than solely relying on chip stocks to continue rising. From a technical perspective, the MAGS index has rebounded nearly 10% from recent lows, re-establishing itself above the 200-day moving average and approaching the long-term upward trend line since last April. JPMorgan believes that the current 200-day moving average has flattened out, indicating that the market is undergoing a lengthy consolidation phase. Historical experience shows that the longer the sideways trend lasts, the stronger the breakthrough often is after the direction is chosen. This article is reprinted from "Wall Street Insights," author: Li Jia; GMTEight editor: Huang Xiaodong.