Wall Street splits on the next move in AI semiconductors, with JPMorgan urging a dip-buy and Morgan Stanley favoring hyperscalers

A fresh debate among Wall Street strategists is reviving a familiar question for the AI trade: is the next leg best played through chipmakers, or through the hyperscalers that buy and operate the data centers driving AI demand. According to a market report published July 6, JPMorgan is leaning toward buying weakness in AI-related semiconductors, while Morgan Stanley is pushing investors to tilt toward the hyperscalers instead.

The split appears to center on how each house is interpreting the market’s recent selloff in chips. JPMorgan’s view, as characterized in the report, is that the semiconductor dip may represent a buying opportunity because AI chip demand has not broken down. In other words, the stock-price weakness is seen more as sentiment and positioning than as evidence that customer demand for AI infrastructure is fading.

Morgan Stanley, in contrast, is described as aiming at a different part of the supply chain. Rather than emphasizing semiconductors directly, the bank is said to want investors in hyperscalers, the large cloud and platform operators that run the AI workloads and place the bulk of the orders for data-center hardware. The implied logic is that if hyperscalers continue to build out AI capacity, their earnings and cash generation may offer a steadier route than semiconductors that can swing more on expectations and product-cycle timing.

The divergence fits a broader pattern taking shape early in the second half of 2026, with multiple firms weighing whether AI’s momentum is shifting from growth-at-any-price to a more earnings-validated phase. One research-style summary indexed by BigGo Finance described JPMorgan and other global banks as generally treating the pullback as a chance to reposition rather than a sign of systemic risk. It also cited JPMorgan’s European equity strategy leadership describing a persistent “buy the dip” approach amid geopolitical turbulence, while acknowledging the market’s sensitivity to macro variables like energy prices, inflation expectations, and currency moves.

In that same summary, the report said other banks were also picking at the internal composition of the AI complex, including calls that memory chips could become especially important as AI systems evolve. While the JPMorgan-versus-Morgan-Stanley dispute focuses on whether investors should buy chip weakness or rotate toward platform owners, the wider takeaway is that strategists are increasingly mapping AI exposure to where they think earnings durability will show up first.

What is not clear from the available public excerpts is the level of granularity each firm is using for its “buy” posture. The July 6 market report references JPMorgan’s overall stance and Morgan Stanley’s preference for hyperscalers, but it does not provide specific stock lists, target prices, or explicit valuation metrics in the materials available here. It also does not spell out what would invalidate each thesis, such as a slowdown in hyperscaler capex (capital spending) plans or a change in supply constraints for advanced chips.

For investors, the practical difference between the two views is straightforward, but not necessarily easy to implement. A chip-focused approach bets that AI infrastructure buildouts remain strong and that chip stocks will eventually re-rate as demand expectations stabilize. A hyperscaler tilt bets that even if chip stocks remain volatile, the platforms that control the workload demand and have the scale to absorb supply-chain timing will deliver more resilient earnings trajectories.

Next, the debate is likely to keep hinging on concrete indicates from the data-center cycle, including how hyperscalers describe future infrastructure spending and whether semiconductor demand indicators remain firm. If hyperscalers report steady or rising AI capex while chips continue to look cheap on sentiment, the JPMorgan-style dip-buy narrative may gain support. If, instead, hyperscaler guidance cools or AI hardware spending looks more constrained than the market is pricing, the Morgan Stanley-style rotation story could become more persuasive.