AI Trading's Next Act: Goldman Sachs Bets on Revenue Revolution Despite Market Woes

Since 2025, AI trading has experienced significant fluctuations, particularly with DeepSeek's rise, which has shaken the U.S. stock market. These developments have led investors to question whether the AI trading boom is over. However, analysts from Goldman Sachs, including Ryan Hammond and David J. Kostin, expressed optimism in their March 6 report, suggesting that despite short-term turbulence, the continuous advancement in AI technology and resulting profitability will eventually lure investors back.

Goldman Sachs maintains that AI trading is poised for expansion, with the "third phase of AI trading"—companies achieving revenue growth through AI technologies—being more alluring than the "second phase". They categorize AI stocks into four phases: the initial phase includes AI chip stocks led by Nvidia, the second covers AI infrastructure stocks, the third involves revenue-generating companies utilizing AI, and the fourth benefits from productivity enhancements driven by AI.

The report identifies two primary factors impacting AI trades: downgraded economic growth forecasts and adjustments in investment positions. Concerns arose regarding companies' capacity to fund AI initiatives amidst slowing economic growth. Overinvestment in second and third phase stocks has exacerbated this by creating overbearing positions.

Since February 19, Nvidia and AI infrastructure stocks have underperformed the S&P 500 equal-weight index by 16 and 9 percentage points, respectively. Stocks related to AI revenue generation underperformed by 7 points, and those tied to productivity enhancement fell by 3 points compared to the index.

Nonetheless, Goldman Sachs views AI’s long-term outlook favorably. They believe technological progression and profitability growth will eventually reignite investor interest in AI stocks. The short-term market challenges require "position clearing" or improvements in economic data to pivot the current situation. Last Friday, their sentiment index dropped to -0.4, indicating that positions haven’t bottomed out sufficiently to support a tactical upswing.

They emphasize the appeal of third-phase AI stocks over those in the second phase. As big companies reduce capital expenditure on AI and costs decrease, investment focus is expected to shift from infrastructure to technology application and revenue generation. Even after a sell-off, second-phase AI stocks maintain relatively high valuations compared to their historical averages, whereas third-phase stocks show relatively lower valuations.

It's noteworthy that third-phase AI stocks have witnessed positive sales revisions this year, with median sales expectations for 2026 rising by 0.3%. In contrast, both second and fourth-phase AI stocks experienced negative revisions, with 2026 median sales forecasts reduced by 0.3%. Within Goldman's compilation of third-phase AI stocks, 40 companies are highlighted, with 27 being software enterprises. Companies projected to see rapid growth in sales over the next two years include Palantir, Cloudflare, SentinelOne, and GitLab.