【AI + Bubble】Goldman Sachs Partner Warns: Tech Stocks’ Soaring Rally Slightly Hints at the “2000 Dot-com Bubble Atmosphere” as Market Structure Becomes More Fragile

U.S. tech and semiconductor stocks continue to stay strong, drawing Wall Street’s attention to concerns that the market may be overheated. Rich Privorotsky, a partner in Goldman Sachs’ sales and trading division, said plainly that the current U.S. stock market trend has “a slight whiff of the 2000 bubble.” Prices are in a “crazy chase upward” (Upward Parabola) state, and market sentiment has moved into an extremely euphoric range.

Privorotsky admitted that betting bearish too early during the bubble-formation stage is often futile, and that predicting a “top” is even more of a high-risk act. Therefore, even in trend trading, it is not advisable to fight the trend head-on. At the same time, he reminded that the market structure is becoming increasingly fragile. The vertical rise in semiconductor and memory stocks behind it reflects a “triple resonance” pattern—rising spot prices, rising volatility, and investors chasing bullish options. However, the mechanical buy-side “space” for volatility-control strategies has clearly narrowed. Once the market encounters shocks from inflation, interest rates, or geopolitical tensions, downside pullback pressure could be amplified.

AI computing power demand is questioned

Privorotsky did not deny the long-term demand for AI computing power, but questioned the logic of current valuations. He believes the real debate is not whether AI computing demand will grow, but whether all of that demand will concentrate on high-end GPUs and large data centers, or whether a substantial portion will shift to cheaper distributed and edge computing.

As open-source models, small models, and improvements in local CPU inference capabilities advance, some low-value-density tasks may not necessarily require the most expensive cutting-edge computing power. If computing power compression continues and model efficiency keeps improving, the assumption that market demand for high-end chips will expand without limit may face a repricing.

Privorotsky pointed out that although major U.S. stock indices keep making fresh highs, the internal structure of the market shows signs of weakness. Data shows that market breadth has fallen back to recent lows, meaning this rally is increasingly dependent on a handful of AI and semiconductor leaders. Meanwhile, cracks have appeared on the consumer side, reflecting that pressure in the real economy is accumulating.

He described the current period as having entered a structurally high-volatility environment, where holding Gamma (volatility sensitivity) could continue to deliver returns. For investors, even if chasing gains in the short term may still work, the risk-reward profile has become clearly asymmetrical.

Macroeconomic undercurrents not gone: inflation re-ignites, geopolitical deadlock

On the macro level, there are also signs that global inflation pressures may be reigniting. Privorotsky mentioned that China’s latest CPI and PPI data came in above expectations, which could signal that global re-inflation risk is heating up. In addition, geopolitical situations such as the deadlock between Iran and the United States, oil-price volatility, and a casino market that shows the probability of an economic recession in the UK exceeding 50%—these risk factors are currently being obscured by AI narratives, but they have not disappeared.

He wrapped up with a tone of black humor:

“Hope the prices of storage chips keep going up… because the energy situation feels like it’s still very far from being truly resolved.”

		Chasing trades at the open for quick gains or loss-taking
	





All eyes on the “Xi–Trump meeting”: will the stock market move across the board? For Hong Kong stocks, will rally momentum still rely on Alibaba and Tencent’s earnings?