Goldman Sachs launches thematic investment portfolio of "resisting the impact of AI": long computing power and security, short replaceable software stocks.

date
08:04 14/02/2026
avatar
GMT Eight
Goldman Sachs has launched a software stock "long-short pair trading" portfolio: going long on companies that are considered difficult to be replaced by AI, with businesses that typically require physical execution, strict regulation, or human responsibility; while shorting companies whose processes are likely to be increasingly automated by AI or replaced internally by the company.
In order to deal with the increasing turmoil in software stocks, Goldman Sachs Group recently launched a new custom stock portfolio that bets on companies that are considered less likely to be impacted by artificial intelligence (AI) compared to other companies. Goldman Sachs launched a software stock "long-short pairing trade" portfolio: to go long on companies that are considered difficult to be replaced by AI, whose businesses typically require physical execution, strict regulation, or human responsibility; and to short companies whose workflows may be increasingly automated by AI or internally replaced by the company. On the long side, Goldman Sachs is bullish on companies that will directly benefit from the increasing adoption of AI, including computing power providers, data infrastructure companies, observability tools, security network companies, large-scale cloud service providers, and AI development platforms. Companies included in this portfolio are Cloudflare, CrowdStrike, Palo Alto Networks, Oracle, and Microsoft. On the short side, traders are targeting companies that are software-driven, but whose workflows may be automated or rebuilt internally as AI capabilities improve. These companies include Monday.com, Salesforce, DocuSign, Accenture, and Duolingo. Media reports indicate that Faris Mourad, Vice President of Goldman Sachs' Custom Basket Team in the United States, wrote in a report to clients: "We expect that as the recent sell-off in software stocks comes to an end, the long part of the portfolio will rebound, while the short part will continue to lag." The launch of this portfolio comes at a time of increasing market concerns about the disruptive impact of AI. Last week, Anthropic launched an efficiency tool for internal legal teams, which led to a sharp decline in legal software and publishing stocks. Subsequently, this sell-off continued to expand. A lesser-known startup, Altruist, launched a tax strategy tool, causing stock prices of companies like Charles Schwab and LPL Financial to drop by over 10% in the past week. Media reports indicate that Wall Street's skepticism towards software stocks has been accumulating for several months, but recent market sentiment has shifted from cautious to clearly defensive. As concerns grow that generative AI may erode traditional business models and compress profit margins, investors have been selling stocks across the entire industry. This sell-off has also reset valuation levels. A year ago, the price-to-earnings ratio of software stocks was around 51 times, making it the highest valued industry in the stock market. But now, the industry's price-to-earnings ratio is around 27 times. Overall profit expectations remain stable. According to media forecasts, the software and services subindustry is expected to achieve about 14.1% profit growth by 2026. Although this growth rate is lower than the overall technology industry's expected growth rate of around 31.7%, which benefits from the expansion of the semiconductor industry, it is still higher than the S&P 500 index's profit growth expectation of 13.7%. This article is reprinted from "Wall Street News," author: Zhao Yuhe; GMTEight editor: Chen Siyu.