The AI industry wave is fundamentally reshaping the underlying structure of the global economy, with a deep-seated divergence characterized by a silicon-based economy 'consuming' a carbon-based one accelerating rapidly.

Research from Shenwan Hongyuan Securities has introduced a "Dual K-Shaped Economy" framework in its 2026 mid-year overseas macroeconomic outlook, revealing two intertwined and mutually reinforcing lines of divergence in the current global economic operation.

In market terms, assets related to AI continue to outperform traditional sectors. In U.S. stocks, AI and energy sectors lead gains, while healthcare and financial sectors lead declines, clearly illustrating the divide between silicon-based and carbon-based assets in pricing.

Economically, manufacturing sector sentiment continues to recover and accelerate, whereas the services sector has been declining sharply since March 2026, creating a historic crossover—the former strengthened by AI, the latter weakened by a softening labor market and energy shocks.

The underlying tension of this structural divergence lies in the fact that the ongoing expansion of AI capital expenditures is systematically "crowding out" traditional economic sectors through three channels: financing, inflation, and employment. Concurrently, the global liquidity environment is transitioning from an era of "savings glut" to one of "capital war," further amplifying the intensity and persistence of this tension.

K-Shaped Markets: Silicon Outperforms Carbon, Geopolitics is Not Noise

The divergence in global asset prices has become highly structured.

The Shenwan Hongyuan report notes that U.S. stocks staged a peak following the Fed's hawkish rate cut in October 2025, with adjustments intensifying due to the HALO event early in the year and geopolitical shocks in the Middle East in late February. However, markets strengthened again in April, driven by progress in ceasefire talks and better-than-expected earnings season results.

Currently, U.S. equities and oil prices remain the two primary anchors for pricing interest-rate-sensitive assets.

At the sector level, AI and energy are leading gains, while healthcare and finance are leading losses, showing a stark divergence in market performance between silicon-based and carbon-based industries.

It is noteworthy that geopolitics is not mere market noise—the Middle East situation persistently influences energy prices through oil supply expectations, which in turn transmits to inflation expectations and interest rate pricing, constituting a significant variable for asset price volatility.

The Dual K-Shaped Economy: Two Interwoven Divergence Trends

Shenwan Hongyuan characterizes the current structural divergence in the global economy as "Dual K-Shaped."

The first is the traditional K-shaped divergence between high-income and low-income groups.

The second is the emerging K-shaped divergence between the AI economy and the non-AI economy. These two trends are superimposed, jointly shaping the current operational landscape of the global economy.

The "great reversal" between manufacturing and services is the first manifestation of the dual K-shaped economy.

Since the second half of 2023, global manufacturing sentiment has been recovering, operating above the expansion-contraction line since the second half of 2025 and accelerating its recovery since early 2026.

Simultaneously, the global services sector began weakening in the second half of 2025 and has been declining sharply since March 2026, primarily due to energy shocks and a loosening labor market. The report states that the strength in manufacturing is essentially a reflection of the AI industry trend, while the weakness in services reflects cyclical forces.

Regarding the traditional K-shaped divergence, since early 2023, as the U.S. labor market weakened, the divergence between high and low-income groups has become increasingly evident across dimensions such as actual spending, wage growth, inflation resilience, and wealth accumulation.

According to Federal Reserve data from the 2025 Q2 Survey of Consumer Finances, the top 20% of U.S. households by income hold 71% of the nation's net worth and 87% of corporate equities and mutual fund assets, while the bottom 20% hold only about 3% of net worth or total assets.

This implies that the wealth effect generated by the U.S. stock market bull run has been almost entirely captured by high-income groups, naturally translating into strong momentum for high-elasticity consumption like luxury goods, high-end travel, private aviation, and luxury home purchases. In contrast, consumption by low-income groups remains highly dependent on wage income or government transfers.

Regarding the new K-shaped divergence, the rift between AI and non-AI is almost ubiquitous: the divergence in sentiment between manufacturing and services, the divergence in growth rates between AI-related and non-AI imports, and the divergence in scale between AI and non-AI investment collectively sketch a structural picture of an accelerating expansion of the silicon-based economy and a relative contraction of the carbon-based economy.

Silicon 'Consumes' Carbon: Three Crowding-Out Channels

The "crowding-out effect" of AI on the traditional economy is a central thesis of the report.

Shenwan Hongyuan points out that during periods of tight labor markets and loose liquidity, the upper arm of the K-shaped economy (high-income, AI-related) and the lower arm (low-income, traditional sectors) are complementary. However, during periods of slack labor markets and tightening liquidity, they become mutually exclusive. The current macroeconomic environment is evolving towards the latter state, with the crowding-out effect transmitted through three channels.

Financing Channel: The ongoing expansion of capital expenditures by AI firms is increasingly reliant on external financing. The report shows that the ratio of capital expenditures to free cash flow has been rising for both representative AI firms and the Nasdaq 100 index.

Market consensus expects that the capital expenditures of the five major U.S. cloud providers will "consume" their free cash flow by Q1 2027.

As firms deepen their reliance on external financing, the sensitivity of AI capital expenditures to macro liquidity will rise significantly. This will exert sustained pressure on mortgage rates by pushing up the neutral rate, thereby suppressing housing market sentiment and causing U.S. stocks and Treasuries to turn negatively correlated—the stronger the stock market, the higher Treasury yields.

Inflation Channel: AI capital expenditure and application are already pushing up prices for related capital goods and end-products.

Researcher Salem Abo-Zaid (2026) estimates that AI contributed an average of about 0.2% to U.S. inflation between 2007 and 2025, with a rising trend in recent years. The report notes that the medium-to-long-term inflationary effect of AI depends on the interplay between productivity gains and corporate cost-plus pricing. The historical correlation between technological progress and deflation is neither significant nor stable, suggesting excessive expectations for AI's deflationary effects are unwarranted.

Employment Channel: The substitution effect of AI on employment is accumulating, with the structural effect currently outweighing the aggregate effect.

The most significant substitution is occurring in white-collar jobs, corresponding to the middle-class group at the 50th to 80th wage percentiles, with key impacted industries including Professional, Scientific, and Technical Services (finance and law), Enterprise Management, Information Technology (e.g., software engineers), and Office and Administrative Support.

The report argues that while technological progress is not a direct cause of cyclical unemployment fluctuations, it does drive the flow of jobs and contribute to the long-term decline in labor's share of income.

Liquidity Tension: From Savings Glut to 'Capital War'

The paradigm shift in the global liquidity environment forms the macroeconomic backdrop for the tension in the dual K-shaped economy.

Shenwan Hongyuan notes that the return of inflation in the post-pandemic era, coupled with the accelerated restructuring of the global order and the redefinition of safe assets, means the "global savings glut" era described by Bernanke is giving way to an era of "capital war."

The rate-cutting cycle may be nearing its end. The report suggests that the Fed's policy rate is no longer restrictive, and overall financial conditions remain accommodative.

The U.S. economy is in a state of "mid-air refueling," with the output gap positive for 19 consecutive quarters since Q2 2021, a very robust labor market, and an unemployment rate of 4.3% roughly matching the natural rate, even showing marginal signs of tightening. Against this backdrop, rate cuts are not impossible but require stringent conditions, and the probability of subsequent "rate hikes front-loaded" is higher.

The uptrend in long-term rates is not over. Since the second half of 2024, major Western economies have entered a rate-cutting cycle, yet long-term government bond yields in some economies have risen instead of falling—UK 10-year and 30-year gilt yields have exceeded the peaks of the 2022 "gilts crisis," and U.S. 30-year Treasury yields have also reached new highs.

The report points out that the pricing logic for long-term rates is more complex and often directionally deviates from the policy rate, indicating that monetary policy rates are not restrictive relative to economic fundamentals.

The QE era has ended. The report reviews the history of the Fed's balance sheet expansion since 1914, noting that true QE-style expansion has occurred only four times, each after the policy rate fell to or near zero.

Until interest rates return to or near zero again, major Western central banks will have bid farewell to the QE era. A restart of QE may require waiting for the next crisis.

The U.S. dollar may maintain a relatively strong pattern in the short term. Since the shale revolution, the U.S. has gradually become a net exporter of crude oil and petroleum products, shifting the correlation between oil prices, terms of trade, and the dollar from negative to positive. Against the backdrop of Middle East geopolitical conflict, a crisis in the Strait of Hormuz inherently benefits U.S. crude exports, a mechanism that will support the dollar in the near term.