Goldman Sachs: Institutions sold tech stocks at a record pace last week, AI trade shifts from storytelling to revenue realization.

BlockBeats news, October 24: Rich Privorotsky, Goldman Sachs' head of European fixed income and equity trading, said that last week the bank's prime brokerage book saw the largest technology stock sell-off in its history. The previously highly crowded "long AI, short everything else" trade is unraveling, with investors beginning to demand that AI companies prove massive capital expenditures can translate into actual revenue.

Privorotsky noted that the market expects AI capital expenditures to reach $1.1 trillion next year, but high spending alone is no longer enough to support valuations; cloud service providers also need to prove that orders can convert into revenue and improve free cash flow. Meanwhile, AI infrastructure faces real-world constraints, with an expected power shortfall of 40 to 50 gigawatts next year, and highly leveraged cloud service providers are also facing financing pressure.

He believes the world is shifting from "savings glut" to "spending glut," with AI investment and government fiscal deficits together pushing up the cost of capital and weakening market risk appetite. Investors are currently worried about both insufficient computing power supply and future overcapacity, and the AI trade is facing a dual test of monetization capability and return on investment.

Despite rising short-term risks, Privorotsky still pointed out that spot computing power prices continue to climb and demand remains strong; some electrification, civil aerospace, and transportation assets that suffered heavy sell-offs are also beginning to show investment value.