Goldman Sachs: AI trading shows differentiation, market begins to examine capital returns BlockBeats News: On June 25th, Goldman Sachs strategists believe that AI trading on Wall Street is entering a more complex phase: the market still believes in the AI investment cycle, but no longer puts all AI companies in the same valuation framework. Over the past year, investors have been most willing to buy direct beneficiaries on the AI infrastructure chain. NVIDIA, TSMC, and some semiconductor equipment and server suppliers have benefited from the continued increase in capital expenditures by large cloud computing companies. As long as Amazon, Alphabet, Meta, and Microsoft continue to purchase chips, servers, and data center capacity, hardware companies' revenue expectations will continue to receive support. But the stock price performance of the hyperscaler itself, which bears these expenses, is not as strong. The market is rewarding the 'receiving party' but remains cautious towards the 'spending party'. Investors are increasingly concerned about whether these hundreds of billions of dollars in AI investments can ultimately be converted into profits, free cash flow, and shareholder returns. This is what Goldman Sachs calls AI trading like a 'stretched rubber band'. On one hand, hardware suppliers' orders and profit expectations are constantly being raised; On the other end, large technology platforms are facing increasing capital expenditure pressure. As long as the demand for AI continues to grow, this structure can still be maintained. But if the market begins to doubt investment returns, or if cloud giants hint at a peak in AI spending growth, related stocks may be re priced. Goldman Sachs is not bearish on AI, but believes that AI trading has entered the stage of return verification from thematic investment. The market is no longer just asking 'who participates in AI', but is starting to ask 'who can truly make money from AI'. For Nvidia, TSMC, and the AI device chain, the biggest risk is not the disappearance of demand, but the fact that demand growth will no longer continue to exceed expectations. For Amazon, Alphabet, Meta, and Microsoft, short-term pressure comes from excessive capital expenditures; But if the cost of AI decreases or AI products bring clear revenue, they may actually become the next beneficiaries. The larger variable is the AI cost curve. If China, Japan, or other regions can develop and operate high-performance models at lower costs, the current high capital expenditure path of large US technology companies may be questioned. The market previously assumed that leading AI would inevitably require more chips and larger data centers, but the improvement of model efficiency and the development of alternative chips may weaken this logic. Therefore, the AI mainline is not over, but the buying logic is becoming more refined. The key to the next stage is no longer just whether there is a demand for AI, but who can turn AI investment into real cash flow.