Last week’s selloff in semiconductor stocks has sparked cautious optimism on Wall Street that now might be a good time to re-enter the sector. The VanEck Semiconductor ETF (SMH) fell 8.9% last week, its biggest weekly decline since a 15% drop in April 2025. The Philadelphia SE Semiconductor Index (SOX) fell nearly 10%. Despite concerns that the new frontier model will force some consolidation at the application level, analysts believe that hardware and infrastructure-level fundamentals are essentially intact, there is ample headroom for capital spending, and memory chip bottlenecks are unlikely to disappear in the near term. “We believe there is a lot of gas left in the tank due to AI capital spending…and (gigawatt) power capacity growth well beyond 2028-29 East…all of which is driving semiconductor demand with a large supply gap,” Mizuho’s Vijay Rakesh said in a note to clients on Saturday. Morgan Stanley’s Joseph Moore wrote on Monday that the recent decline in U.S. memory stocks “creates an attractive entry point.” “We believe the best value in the market comes from the computing names, particularly NVDA and AVGO, but given this slowdown, memory is catching up quickly and we think this could be a good entry point for stocks,” he said. .SOX 1M Mountain Philadelphia SE Semiconductor Index Past Month Performance Nvidia and Broadcom stock prices have fallen nearly 4% and 10%, respectively, over the past month. “Semiconductor companies should start bidding soon,” JPMorgan’s Mislav Matejka said on Monday. “In our view, fundamentals are likely to remain constructive and with no significant supply additions expected until 2028, it would be premature to price in a shift at this point.” Evercore analysts became slightly more cautious over the weekend. They describe SOX’s 20% decline over the past four weeks as a “mid-cycle correction,” suggesting fundamentals remain positive, but believe SOX will fall further before rebounding. “It could take another 2-3 weeks and another 10-15% (to complete play),” Evercore’s Mark Lipasis wrote on Sunday. “The median mid-cycle adjusted SOX rebound rate is 36% over 20 weeks.” Last week, Chinese company Moonshot AI released Kimi K3, a frontier AI model with a partially disclosed or “open weight” setting, sending chills down the spines of the U.S. tech industry. Dean Ball, head of strategic futures at OpenAI, spoke of this in the context of “complete AI communism” in a social media post on Friday. But almost everyone on Wall Street expects consolidation at the AI software level, which should create a small number of dominant players and corporate configurations, without necessarily penalizing the hardware level. Peter Fenton, general partner at venture capital firm Benchmark, told CNBC over the weekend: “Given that open-weight models offer high performance at significantly lower costs, we’re going to see an interesting, almost thermonuclear, battle to provide[computing power]to support the emerging demand for open-weight models.”
