Wall Street is shifting its focus from valuation expansion to earnings realization. JPMorgan values earnings trends and position improvement, Goldman Sachs focuses on the easing macro environment, and Jefferies is betting on an AI-driven earnings growth cycle. The AI bull market seems to be looking for an “earnings baton-passing window”—when valuation compression slows, as long as earnings per share continue to grow, stock prices do not have to rely on the price-to-earnings ratio rising back to high levels in order to climb. Whether this logic can be realized will depend on whether the expansion of AI complex-task workloads can exceed the impact of declining unit inference prices and ultimately translate into cloud service revenue, chip orders, and sustainable profits.
The JPMorgan strategy team led by Mislav Matejka believes that the U.S. stock market’s seven major tech giants (Mag 7) have already undergone a significant valuation reset, and their forward 12-month price-to-earnings ratio relative to the broader market has fallen to about one standard deviation below the historical median, at a ten-year low. Morgan Stanley data also shows that the valuation premium of the Magnificent Seven relative to the other 493 stocks in the S&P 500 is only about 10%, the lowest in more than a decade, while they as a whole still enjoy an annual earnings growth advantage of about 45%.
JPMorgan points out that declining crowding in tech sector positioning, strong earnings, and more realistic valuations together create an opportunity to reposition, with a particular preference for semiconductors. Since June, forward 12-month earnings per share forecasts for semiconductors have been revised up by about 30%, while software earnings expectations have lacked corresponding improvement. Based on this, the bank favors a relative trade of “going long semiconductors and shorting some high-momentum software.” JPMorgan also acknowledges that higher AI capital expenditure increases capital intensity, rising financing needs, and pressure on free cash flow do indeed correspond to a certain valuation discount, but a considerable portion of the adjustment has already occurred, and strong earnings from hyperscale cloud providers may continue to support share prices.
Meta Muse and OpenAI’s GPT-6 Astra have ignited a massive AI agent wave, and AI FOMO trading has made a comeback. Agents transform a single request into multi-stage work such as retrieval, planning, tool invocation, and code execution, driving accelerators such as GPUs and TPUs to handle model computation, CPUs to run browsers and database queries, HBM and server DRAM to carry model data and context, enterprise SSDs to store task records and persistent states, and high-speed networking and optical interconnects to support distributed data exchange. Nasdaq’s official weekly report confirmed that Muse’s early performance and the AGI discussion sparked by Astra jointly boosted AI FOMO trading, with the Nasdaq 100 index hitting a record high on September 22 and rising about 3% cumulatively last week.
Goldman Sachs strategist Mark Wilson believes that a year-end rally does not necessarily require waiting for the midterm elections to end; if inflation slows and growth cools moderately without falling into recession, further tightening may be difficult to fully deliver. Goldman Sachs economist Jan Hatzius emphasized that upside risks to the growth market have diminished, while strategist Ben Snider believes that some industries have excess earnings but that an overall earnings bubble has not yet formed. Jefferies expects that, driven by both the AI investment frenzy and earnings beating expectations, the S&P 500 index will soar to 8,000 by the end of 2026 and further reach 9,000 in 2027, with its baseline forecast based on earnings per share of $373 and a price-to-earnings ratio of 21.5 times. At Monday’s U.S. stock market close, AI computing power-themed stocks such as AMD (AMD), Micron (MU), and SanDisk (SNDK) came under pressure as oil prices and U.S. Treasury yields rose, but Nvidia (NVDA) rose against the trend after announcing a record $150 billion increase in its share buyback authorization.
