Shenwan Hongyuan Group Co., Ltd. has released a research report stating that since April, AI has led technology stocks to significantly outperform the market, driven by better-than-expected corporate earnings and upward revisions in capital expenditures. Historically, the intensity of AI capital expansion may still have room to increase. The report points out that while Federal Reserve interest rate hikes could hinder the AI "bubbling" process, the probability of a rate hike within the year remains low. Financing pressure does not currently pose a constraint, but power bottlenecks and public opinion may become limiting factors. The performance realization of AI-related companies is likely the short-term focus.

Shenwan Hongyuan's main views are as follows:

**Key Consideration: How far has the AI 'bubble' progressed?**

**1. What is the progress of the AI revolution? AI application, job displacement, and productivity boosts are in the early stages.**

Since April, AI has led technology stocks to significantly outperform, supported by strong corporate earnings and increased capital spending. In Q1 2026, "hard tech" companies like chip designers and manufacturers reported high profit growth, correlating with their stock performance since April. In Q1, the year-over-year growth rate of capital expenditures for the "Magnificent 7" (M7) reached 65%, accounting for 33% of the S&P 500, with capital expenditure guidance continuing to be revised upward.

AI is not only driving stock market prosperity but is also becoming increasingly important to the U.S. economy and global trade. In Q1 2026, AI investment contributed 1.1 percentage points to U.S. economic growth, stronger than consumer spending and significantly stronger than traditional investment. Robust U.S. AI investment has also significantly boosted its import demand, reflected in the strength of global manufacturing PMI and the resilience of global trade.

Beyond investment, AI application, job displacement, and productivity enhancement are in their early stages. As of May 2026, the AI adoption rate among U.S. businesses was only 19.8%, far below theoretical levels across industries. AI's boost to U.S. productivity is still comparable to the early stages of the internet revolution. The impact of AI on U.S. employment remains limited, concentrated in high-exposure industries, younger demographics, and higher-paying positions.

**2. How far has the AI bubble progressed? Investment intensity and external financing still have room to grow, with solid financial metrics.**

Historically, the intensity of AI capital expansion may still have room to increase. Since Q4 2022, the share of U.S. AI investment in GDP has cumulatively increased by 1.0%, still having room to grow compared to the internet revolution (1.4%). While capital expenditures of leading tech firms are approaching their operating cash flow levels, the Nasdaq 100's capital expenditure to operating cash flow ratio remains below early 2000 levels.

Although leading U.S. tech firms have begun turning to external financing, concerns about financing pressure are premature. Measured by net debt/EBITDA, the debt pressure on leading U.S. tech firms is significantly lower. In Q1 2000, the combined flow of U.S. corporate bonds, loans, and equity financing accounted for 5.4% of GDP, whereas by the end of 2025, this level was only 1.6%, not signaling excessive financing.

Financial metrics for leading U.S. tech stocks remain robust, with valuations still aligned with performance. During the internet revolution, the market cap increase of leading U.S. tech companies far exceeded profit growth, but in this cycle, market cap gains remain relatively matched with profits. Financial indicators such as cash/market cap, ROE, and net profit margin for the M7 are stronger than those of leading companies during the dot-com bubble, indicating that AI investment is built on a solid financial foundation.

**3. What is the sustainability of AI investment? Financing pressure is not yet a constraint; focus on AI performance delivery.**

While Federal Reserve rate hikes could hinder the AI "bubbling" process, the probability of a rate hike within the year remains low. Under a "weak balance" in employment, the likelihood of a Fed rate hike this year is not high. If the economic impact of oil prices materializes, expectations for rate cuts may rise. The ratio of Nasdaq 100 corporate bond issuance to capital expenditure remains below 40%, significantly lower than the peak of the past decade, indicating that corporate reliance on debt financing is still not high.

When will U.S. AI investment growth peak? Financing pressure is not yet a constraint, but power bottlenecks and public opinion may become limiting factors. M7 capital expenditure growth in 2026 may exceed 60%, raising questions about whether high growth can be sustained into 2027. Potential constraints for AI investment include grid bottlenecks, equipment shortages, and public opposition to data center construction, which could lead to project delays or even cancellations.

The performance realization of AI-related companies is likely the short-term focus. Since 2023, the margin by which major tech firms' earnings exceed expectations and their safety margins have narrowed, potentially raising future market doubts about the rationality of capital expenditures. Core indicators for observing AI investment demand should primarily focus on the performance of leading tech companies, especially the Annual Recurring Revenue (ARR) performance of firms like OpenAI and Anthropic.

**Risk warnings:** Oil price increases exceeding expectations; a more hawkish policy stance from the Fed; U.S. economic slowdown exceeding expectations.