Goldman Sachs Trading Desk In-Depth Analysis: Behind the AI Capital Frenzy, Monetization Challenges Are Quietly Emerging

```

The wave of nearly one trillion dollars in AI capital expenditure is facing a severe monetization dilemma, and cracks in valuation logic are beginning to emerge.

Rich Privorotsky, head of the One-Delta trading desk at Goldman Sachs, pointed out in a recent client briefing that the AI narrative is shifting from a "race for capability" to a "battle for monetization." The investment return threshold implied by current AI capital expenditure has exceeded levels actually achieved in any comparable technology cycle in history. Meanwhile, the rapid evolution of open-source models and local inference solutions is compressing the space of AI economics, and rational motivations for companies to delay spending are rising.

For the market, the risk transmission path is already clear: Expectations come under pressure first, then impact revenue growth of companies currently benefiting from the AI construction boom. Global semiconductor and hardware-related stocks have around $100 billion in leveraged capital exposure, coupled with heavy retail investor participation. If market sentiment reverses, there is a significant risk of rapid deleveraging.

From Capability to Monetization: Returns Thresholds and Delay Risks of AI Investment

The core contradiction of AI investment is shifting from "Can it be done?" to "Can it make money?"

Privorotsky cited recent research from the Wharton School, noting that to reasonably support the current AI capital expenditure cycle, the required productivity gains are extremely rare in history—even compared to the IT boom, achievements within a similar timeframe are still far inferior. He emphasized that the specific numbers are not the key; what really matters is that the implied investment return threshold of current spending is extremely high.

Currently, private sector annualized AI capital expenditure nears $1 trillion. This spending is creating jobs, supporting economic growth, and generating economic spillover effects far beyond AI itself. However, as the narrative shifts from "every company must immediately experiment" to "returns remain hard to quantify," the likelihood of companies further postponing spending is increasing.

A notable signal is that the Silicon Data Token Spending Index has begun to soften moderately. Meanwhile, increasingly powerful open-source models and local inference solutions are significantly compressing the space of AI economics. Once companies generally expect "AI costs will drop sharply in a year," postponing current spending becomes a rational choice—this is the core risk facing current valuation multiples.

Leverage Risk: The Double-Edged Sword of Convexity

Changes on the supply side are becoming a new source of market pressure.

Privorotsky noted that the large-scale stock offerings expected to occur in the second half of the year have already begun, which has always been a key variable testing the market's capacity. Meanwhile, global semiconductor and hardware-related stocks have about $100 billion in leveraged capital exposed, compounded by substantial retail participation, forming a tail risk that cannot be ignored.

He describes the current situation as a "double-edged sword of convexity": When direction is favorable, capital flows can bring significant upside potential; if sentiment reverses, the deleveraging process will also be extremely intense. Some of last week’s market volatility highlighted this very mechanism.

From a more macro perspective, the fundamental pillars of AI investment remain intact—gaps in power and infrastructure are still significant, and capital expenditure is immense. But as expectations for new frontier model releases heat up, market catalysts and valuation pressures will simultaneously test investors’ resolve.

Risk Disclaimer and TermsThe market carries risk, and investments should be made with caution. This article does not constitute personal investment advice, nor does it take into account individual users’ specific investment objectives, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article fit their own particular circumstances. All investments based on this article are at your own risk. ```