AI capital expenditure tsunami reshapes economic resilience, Morgan Stanley raises US stock earnings expectations.

The hot trend of AI infrastructure construction is fundamentally changing the sensitivity of the US economy to prices and financing costs, and driving a substantial upward revision in corporate profit expectations. Andrew Sheets, Chief Cross-Asset Strategist at Morgan Stanley, pointed out in the bank's 2026 Mid-Year Outlook report that the US economy is showing unexpected resilience, with the core DRIVE centered around AI capital expenditure surge. The profit growth forecast for the S&P 500 index in 2026 has been raised from 17% to 23%, and the US economic growth forecast has also been revised upwards from 1.8% to 2.3%. Behind this round of upward revisions is a phenomenon defined by Morgan Stanley as "inelastic demand" - that is, the ability of businesses and consumers to resist higher prices, higher financing costs, and even higher geopolitical risks far exceeds market expectations. Andrew Sheets believes that this feature has profound implications for the investment logic of stocks, bonds, and commodities in the current market environment. AI capital expenditure has expanded sharply, far exceeding previous expectations AI infrastructure investment is the main focus of the report. According to the latest estimates, the capital expenditure of large-scale and super-large-scale cloud computing companies in the US in 2026 will reach $805 billion, nearly doubling from the previous year's forecast of $433 billion, doubling from the actual expenditure in 2025, and increasing to three times the level in 2024. Looking ahead, the bank expects this number to exceed $1.1 trillion in 2027 and approach $1.3 trillion in 2028. Of particular note is that this investment acceleration is happening against the backdrop of comprehensive cost increases. From copper, gas turbines to storage chips, prices of key components have risen significantly, but have not materially inhibited the enthusiasm for AI investments. Andrew Sheets defines this demand feature that is insensitive to prices as "inelastic" and points out that AI investments have a dual attribute of being both a "necessity" and a "highly desirable product" - companies are eager to seize the next generation of core technologies, but also concerned about falling behind in competition. Financing costs have also failed to stem this trend. Tech corporate bond issuance hit a record high in 2026, even as yields continued to rise. Andrew Sheets believes that for such a strategically important priority, whether the borrowing cost is 5.50%, 5.75%, or 6.00%, has become a secondary consideration. Profit upgrades are not limited to AI, economic growth forecasts are also raised simultaneously The spillover effects of AI capital expenditure are directly reflected in macroeconomic data. AI capital expenditure has increased the forecast for US business fixed investment growth in 2026 from 3% six months ago to 7%, more than doubling, which is an important support for the bank's upward revision of the full-year economic growth forecast. On the profit side, the combination of "higher demand plus higher prices" constitutes a positive driver for corporate profit margins. The consensus earnings growth rate for the Korean stock market (which aggregates a large number of AI core suppliers) in 2026 is as high as 235%, with many of the most representative stock investment opportunities in 2026 focused on the AI supply chain. It is worth noting that strong earnings are not limited to AI beneficiaries - the median earnings per share growth rate for stocks in the Russell 3000 index also tracks at 10%. Regarding the economic growth forecast for 2027, it has been raised from 2.0% to 2.6%, showing continued optimism for the mid-term momentum of the US economy. Significant resilience in consumer demand, limited impact of energy price shocks Inelastic demand is not exclusive to AI investments, but is also evident on the consumer side. Despite a sharp increase in US gas prices, consumer driving and spending behaviors have almost not contracted. Gasoline consumption in the US in April 2026 was roughly the same as in April 2025, and retail sales data, excluding oil and gas projects, was better than expected. Air travel data also confirms this assessment. As of April 2026, US airfare prices rose by 20.7% year-on-year, with airlines successfully passing on fuel costs to consumers, and there is no apparent sign of demand shrinking. Having a strong balance sheet is key to supporting this consumer resilience - both the tech sector and the household sector have ample funds, US household wealth is at a historical high, and the savings rate in Europe is also at a historical high. The boundaries of inelastic forces: challenges facing bond markets and central bank policies The above inelastic forces are most prominent in the US, supporting the bank's preference for US stocks over other markets in global asset allocation. At the same time, the continued strong demand for energy is putting oil prices on an upward bias, and the bank believes that being long on oil prices is an effective hedge against its overall cross-asset optimistic view. However, these forces also bring uncertainties to the bond market and central bank policy paths. The baseline assumption is that AI-related categories have relatively limited weight in the inflation basket, and inflation is expected to fall back in the second half of 2026, providing European Central Bank, Bank of England, and Bank of Japan with policy space that is more accommodative than currently priced in by the market. However, Andrew Sheets also warned of tail risks: if the inelastic forces are stronger and more persistent than expected, they may challenge predictions about policy paths and judgments about bond yields moving lower. He boils down the core question to: what the market truly needs to ask may not be whether prices are high enough to affect demand, but whether the demand itself has become too strategic, too indispensable, or too financially strong to care about prices. This article was reprinted from Wall Street Seen, edited by GMTEight: Chen Wenfang.