Clarifying Market Concerns on AI Risks: Is It the Size of the Pie or Its Distribution?

Over the past month, semiconductor stocks have collectively come under pressure, bringing risk hedging for AI-themed investments back into focus. However, a more fundamental question is often skipped in various hedging discussions—what exactly are the AI risks the market is worried about?

According to Zhuifeng Trading Desk, Goldman Sachs' Global Market Commentary released on July 23 provided an analytical framework: AI investment faces two fundamentally different types of risk. One threatens the size of the pie—a shrinkage in the total value created by AI; the other changes how the pie is sliced—the total value remains unchanged, but winners and losers are reshuffled.

The transmission paths of these two types of risk across asset markets are distinctly different, and the available hedging tools are completely different. Although strong earnings reports may temporarily overshadow valuation concerns, and the AI investment boom is expected to continue, Goldman Sachs still advises investors to keep a close eye on potential challenges—the key is to identify the nature of these challenges.

$26 Trillion vs. $9 Trillion

This distinction is urgent because AI pricing is becoming increasingly dependent on optimistic assumptions.

The market capitalization growth of AI-related stocks (including private companies) since November 2022 is approximately $26 trillion, or about $23 trillion after deducting baseline returns. Under Goldman Sachs' baseline assumption, the present discounted value (PDV) of capital income that U.S. corporations can obtain from AI-driven productivity improvements is only about $9 trillion. Even under the most optimistic combination of assumptions—higher productivity growth, faster adoption rates, and a larger capital share—this figure is only about $28 trillion.

In other words, the current market capitalization growth of AI stocks is approaching the upper bound that can only be supported if "everything is perfectly realized."

In the Goldman Sachs framework, any downward revision in expectations for the five variables driving the total value of AI—productivity growth, adoption speed, capital share (corporate monetization capability), international share, and discount rate—will shrink the pie. Macro shocks (monetary tightening, rising oil prices, weakening employment) will also transmit through profit expectations and discount rates. Given the high valuations, concentrated positions, and large financing needs of AI stocks, macro shocks may disproportionately impact the AI sector.

An easily overlooked boundary: the "pie" here refers to the AI value accessible to the U.S. corporate sector. Even if the total global economic value of AI remains unchanged, as long as value flows from U.S. corporations to consumers or non-U.S. producers, the pie shrinks for U.S. stocks.

The distribution pattern is already changing. So far, AI value has mainly flowed to U.S. AI companies and some key Asian enterprises. In the past 6-9 months, the market has especially rewarded the supply side—semiconductor, memory, and other infrastructure providers. Since late 2025, memory chips have seen significant price hikes due to shortages caused by accelerated AI demand, essentially a terms-of-trade shock: chip producers benefit, while consumers suffer. While the pie is growing, the way it is sliced is also changing rapidly.

The Same Event, Two Completely Different Shocks

How do we determine whether an AI risk belongs to "pie size" or "pie slicing"?

Some judgments are intuitive. Slower adoption and limited use cases are typical examples of a shrinking pie. A decline in market financing willingness, which pushes up the discount rate, is also a shrinking pie.

But many situations are far less clear.

Difficulties in monetizing AI products and model competition lowering innovation costs? Revenue flows from corporations to consumers, shrinking the pie in the hands of U.S. corporations—but corporate consumers also benefit, and lower costs may even accelerate adoption. Semiconductor capacity expansion or improved chip efficiency? This hurts producers and benefits consumers; the total pie may remain unchanged, but the slicing changes. AI disrupting traditional industries? This is essentially distributional, but if the winners are not listed in the U.S. market, the pie also shrinks.

Goldman Sachs cited a key example: hyperscale cloud vendors cutting capital expenditures.

The same action, if driven by pessimism about AI investment returns or a tightening financing environment, means the pie shrinks; if driven by the discovery that existing infrastructure is sufficient or by finding more efficient utilization methods, the pie remains unchanged, merely shifting from suppliers to the cloud vendors themselves.

Different sources lead to completely different market consequences.

The Market Has Already Given Its Answer

Five market events over the past 18 months clearly demonstrate this distinction.

First, let's look at three aggregate shocks: the DeepSeek incident (January 2025), widespread AI concerns (February 2026, core skepticism about the ability to sustain high capital expenditures), and non-farm payroll-driven interest rate spikes (June 2026). Market reactions were highly consistent: the S&P 500 fell by 1.5%, 1.7%, and 2.6% respectively; the Cboe Volatility Index surged by 20.5%, 20.9%, and 39.7%; AI baskets and semiconductors significantly underperformed the broader market; credit spreads widened; defensive stocks rose against the trend; and the yield on the 10-year U.S. Treasury note declined by 9 basis points in the first two instances. In the third instance, interest rates themselves were the source of the shock, resulting in the opposite direction. The first two movements aligned with lowered growth expectations, while the third aligned with a hawkish policy shock.

Now, let's look at two distributional shocks: Alphabet announcing a secondary offering to finance AI capital expenditures (June 2) and Meta Platforms announcing the construction of a cloud business to sell excess computing power (July 1).

The picture was completely different. The S&P 500 barely moved, U.S. Treasury yields remained stable, and the Cboe Volatility Index did not budge.

But beneath the surface, there was intense turbulence: On the day of the Alphabet event, semiconductors rose 5.8% while hyperscale cloud vendors fell 2.4%; the Meta Platforms event saw the exact opposite direction, with hyperscale cloud vendors rising 2.5% and semiconductors falling 6.4%. Winners and losers switched violently, offsetting each other at the index level, leaving macro assets almost unaffected.

The classification of DeepSeek deserves separate explanation. Breakthroughs that lower innovation costs may redistribute value to consumers (including corporate consumers), but simultaneously reduce the overall AI revenue share of the U.S. corporate sector—globally, the slicing changes; for U.S. stocks, the pie shrinks. Therefore, it is classified as an aggregate shock.

Volatility data tells the same story. Currently, the average implied volatility of S&P 500 components is at the 99th percentile of the past 15 years, while implied correlation has dropped to its lowest point in 15 years—individual stocks are diverging sharply, but the consensus pricing on the overall value of AI has not shaken. Over the past 6-9 months, apart from macro events such as the Iran war, only aggregate shocks have pushed up implied correlation and index volatility.

The more dominant distributional shocks become, the more limited the pressure on index volatility.

Who Can Hedge, and Who Cannot

For investors holding broad-based U.S. equities, the core threat is aggregate shocks. Fortunately, aggregate shocks have clear macro transmission paths—stock indices and interest rates are the most sensitive, while foreign exchange and commodities provide weaker signals; except when interest rates themselves are the source of risk, aggregate AI concerns typically lower U.S. Treasury yields. Macro hedging tools are available, and diversification across countries and industries can absorb distributional volatility.

Investors holding AI-specific exposures or overweight positions are in a more tricky situation. Aggregate risks can still be hedged macroeconomically, but distributional shocks also impact portfolios yet are difficult to protect using macro assets—distributional shocks self-cancel at the macro level, and the correlation between other assets and core AI positions is unreliable.

If you only hold the broader market, the size of the pie is your problem, but the slicing of the pie is not. But if you are heavily invested in AI, both risks will hit you—and the latter is harder to guard against.