The "Tension" of the Dual K-Shaped Economy: Silicon-Based "Devouring" Carbon-Based

```

The wave of the AI industry is reshaping the underlying structure of the global economy, and a deep differentiation characterized by the "silicon-based economy devouring" the carbon-based economy is rapidly accelerating.

Shenwan Hongyuan Securities Research Institute, in its mid-term overseas macroeconomic outlook report for 2026, proposed the "Dual K-shaped Economy" framework, revealing two interwoven and mutually reinforcing main lines of differentiation in current global economic operations.

At the market level, AI-related assets continue to outperform traditional industries. In the US stock market, AI and energy sectors lead, while healthcare and financial sectors lag. The divide between silicon-based and carbon-based is clearly reflected in asset prices.

At the economic level, manufacturing prosperity has continued to recover and accelerate upwards, while the service industry has been accelerating downward since March 2026, with a historic crossover between the two—the former strong due to AI, the latter weak due to a softening job market and energy shocks.

The deep tension of this structural differentiation lies in: The continued expansion of AI capital expenditure is systematically "crowding out" traditional economic sectors through three channels—financing, inflation, and employment—while the global liquidity environment is shifting from an era of "excess savings" to an era of "capital war," further amplifying the intensity and sustainability of this tension.

K-shaped Market: Silicon-based Outperforms Carbon-based, Geopolitics is Not Noise

The differentiation of global asset prices has become highly structural.

The Shenwan Hongyuan report noted that after the hawkish Fed rate cut in October 2025, US stocks peaked temporarily. The early-year HALO event and late-February Middle East geopolitical shocks intensified the adjustment, but in April, with ceasefire negotiations and earnings season exceeding expectations, the market strengthened continually.

Currently, US stocks and oil prices remain the two main anchors for pricing rate-sensitive assets.

At the industry level, AI and energy sectors lead increases, healthcare and financial sectors lag, and the market performance divide between silicon-based and carbon-based is distinct.

It is noteworthy that geopolitics is not merely market noise—the Middle East situation continues to impact energy prices through expectations of oil supply, thereby transmitting to inflation expectations and interest rate pricing, playing a crucial role in asset price volatility.

Dual K-shaped Economy: Two Differentiation Lines Interwoven

Shenwan Hongyuan summarizes current global structural economic differentiation as "Dual K-shaped":

First is the traditional K-shaped differentiation between high-income and low-income groups;

Second is the newly emerging K-shaped differentiation between the AI Economy and the Non-AI Economy. The two main lines overlay and jointly shape the current operational structure of the global economy.

The "Big Reversal" between Manufacturing and Services is the first manifestation of the Dual K-shaped Economy.

Since the second half of 2023, global manufacturing prosperity has been recovering, running above the boom-bust line since the latter half of 2025, and accelerating its recovery since early 2026.

Meanwhile, global services started weakening in late 2025 and have been accelerating downward since March 2026, mainly due to energy shocks and a loosening job market. The report notes that manufacturing's strength is essentially a reflection of the AI industry trend, while the weakness in services reflects cyclical forces.

Traditional K-shaped Differentiation has become increasingly visible since early 2023, as the US labor market softens: the divide between high- and low-income groups is apparent in actual spending, wage growth, inflation tolerance, and wealth accumulation.

According to the Federal Reserve’s Q2 2025 "Consumer Financial Survey," the top 20% of US families hold 71% of US net assets and 87% of company equity and mutual fund assets, while the bottom 20% only hold about 3% of net assets or total assets.

This means that almost all the wealth effect generated by the bull market in US stocks is captured by high-income groups, and naturally turns into strong momentum for high elasticity consumption such as luxury goods, premium tourism, private aviation, and mansion purchases, whereas low-income consumption greatly depends on wage income or government transfers.

New K-shaped Differentiation describes the almost ubiquitous divide between AI and non-AI: prosperity differentiation between manufacturing and services, growth rate differentiation between AI-imports and non-AI imports, scale differentiation between AI investment and non-AI investment—together outlining a structural picture of accelerating expansion in the silicon-based economy and relative contraction in the carbon-based economy.

Silicon-based "Devours" Carbon-based: Three Crowding-out Channels

AI's "crowding-out effect" on the traditional economy is the core theme of this report.

Shenwan Hongyuan points out: when the labor market is tight and liquidity is loose, the upper arm (high-income, AI-related) and lower arm (low-income, traditional sectors) of the K-shaped economy are complementary; when the labor market is slack and liquidity is tightening, they become mutually exclusive. The current macro environment is evolving towards the latter, and the crowding-out effect is transmitted via three channels.

Financing Channel: The sustained expansion of capital expenditure by AI companies depends increasingly on external financing. The report shows that for both representative AI firms and the Nasdaq 100 Index, the proportion of capital expenditure to free cash flow is continuously rising.

Market consensus expects that the capital expenditure of the five major US cloud firms will "consume" free cash flow by Q1 2027.

As reliance on external financing deepens, the sensitivity of AI capital expenditure to macro liquidity will increase significantly, pushing up the neutral interest rate, exerting continuous pressure on mortgage rates, suppressing the housing market boom, and causing US stocks and US bonds to turn negatively correlated—the stronger US stocks, the higher US bond yields.

Inflation Channel: AI capital expenditure and application are already pushing up relevant capital goods and terminal prices.

Researcher Salem Abo-Zaid (2026) estimates that from 2007 to 2025, AI lifted US inflation by an average of about 0.2%, with the upward trend intensifying in recent years. The report notes that the medium- and long-term inflationary effect of AI depends on the game between productivity growth and enterprise markups, and historically, the correlation between technological progress and deflation is neither significant nor stable—thus, one should not expect too much deflationary effect from AI.

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

The most significant substitution is for white-collar employment, especially the 50% to 80% wage percentile for the middle class; key sectors include professional science & technology services (finance and law), enterprise management, information industry (such as software engineers), and office/administration roles.

The report argues that technological progress is not the direct cause of cyclical unemployment swings but does drive job flows and a long-term decline in labor share.

Liquidity Tension: From Excess Savings to "Capital War"

A paradigm shift in the global liquidity environment is the macro background to the tension of the Dual K-shaped Economy.

Shenwan Hongyuan points out: In the post-pandemic era, inflation has returned, coupled with an accelerated restructuring of world order, redefining safe assets. The "global excess savings" era described by Bernanke is being replaced by the "capital war" era.

The rate-cutting cycle may be drawing to a close. The report believes the Fed's policy rate is no longer restrictive and financial conditions are generally loose.

The US economy is currently in a "mid-air refueling" phase; since Q2 2021, the output gap has been positive for 19 consecutive quarters; the labor market is very healthy and the unemployment rate of 4.3% matches the natural rate, with a trend toward tightening at the margin. In this context, rate cuts are not impossible, but conditions are strict, and subsequent "pre-emptive rate hikes" are more probable.

The upward trend in long-term rates is to be continued. Since the second half of 2024, major Western economies have entered the rate-cut cycle, but long-term government bond yields in some economies are rising instead of falling—the UK’s 10- and 30-year gilt yields have exceeded their "gilts crisis" peaks in 2022, while US 30-year yields have reached new highs.

The report notes that pricing logic for long-term rates is more complex and often deviates directionally from policy rates, indicating that monetary policy rates are not restrictive compared to fundamentals.

The QE era is over. The report reviews Fed balance sheet expansion history since 1914, noting true QE-style expansion only occurred four times, all after policy rates dropped to zero or near-zero.

Until rates return to or near zero, major Western central banks will bid farewell to QE, and its restart will await the next crisis.

The dollar may remain strong in the short-term. Since the shale revolution, the US has gradually become a net exporter of crude oil and petroleum products. The correlation between oil prices, trade conditions, and the dollar has shifted from negative to positive. With Middle East geopolitical conflict and the Hormuz crisis favoring US oil exports, this mechanism will support the dollar short-term.

Risk Warning and DisclaimerThe market has risks; investment needs caution. This article does not constitute personal investment advice and has not taken into account the particular investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article fit their specific circumstances. Investing based on this is at your own risk. ```