Why J.P. Morgan Sees Opportunity in AI Semiconductors After the Pullback

J.P. Morgan’s Sept. 28, 2026, view is a case for reconsidering the AI trade after a pullback—not for buying every technology stock. The bank said investor positioning had become cleaner and valuations had fallen, making renewed interest possible, particularly in semiconductors. It also warned that technology may not return to its earlier levels of success, even while describing the fundamental case as constructive.17

Why semiconductors stand out

Semiconductors are a direct part of the infrastructure being built for AI: data centers need chips, memory and other hardware. In a separate analysis, J.P. Morgan identified high-bandwidth memory as one area where supply bottlenecks can give companies pricing power and support profits.12 The bank has also described AI infrastructure spending as spanning chips, memory, servers, power systems and new data centers.7

That creates a potential link between continued investment in AI infrastructure and demand for semiconductor products. But it is a conditional argument: spending plans, supply constraints and pricing can change, and none guarantees that a particular stock will rise. The Sept. 28 report, as described in the available coverage, does not specify current chip prices or quantify the supply outlook.17

Why the pullback may matter

J.P. Morgan’s immediate rationale is that a less-crowded trade and lower valuations may make it easier for investors to re-engage. That is a change in the setup, not proof that semiconductor shares are cheap by any particular measure: the report does not provide valuation multiples or a target price.17

The bank’s caution also matters. A constructive view of AI fundamentals is not a prediction that technology stocks will repeat their previous gains. Investors still need to assess whether spending and demand translate into durable company earnings.17

What the comparison with software does—and doesn’t—show

The report identifies semiconductors as a particular opportunity, but the available coverage does not establish a specific short position in software or supply a direct semiconductor-versus-software performance comparison. It also does not give the two sectors’ current year-to-date returns. So the evidence supports describing the call as selective interest in chips—not as a proven instruction to sell software.17

Software has its own monetization test. J.P. Morgan’s separate software-sector report says AI is pushing companies toward usage-, consumption- and outcome-based pricing, while enterprise buyers increasingly expect measurable productivity gains and clearer returns on spending.15 That makes evidence of customer value and revenue generation important when assessing software firms, but it does not by itself show that AI will damage the sector’s profits.

The takeaway for investors

The argument for semiconductors rests on two factors: the AI investment cycle creates demand for physical infrastructure, and a pullback may have improved positioning and valuations. J.P. Morgan’s view is best read as a selective, conditional re-entry—not a guarantee of gains, a claim that every chip stock is undervalued, or a definitive verdict against software.1217

The available reporting leaves key comparisons unresolved: it does not quantify current chip supply and pricing, provide a current year-to-date semiconductors-versus-software return comparison, or show how much AI spending is already appearing in earnings. Those details would be needed to evaluate the relative trade more fully.