Nvidia’s (NVDA) Explosive Run May See Limits Going Forward, Warns JPMorgan

Nvidia has rallied about 120% over six months in 2025, but JPMorgan strategist Mislav Matejka cautions that the market may not repeat last year's concentrated semiconductor surge. The Philadelphia Semiconductor Index is up about 9% this month, supported by strong AI-driven demand, including TSMC's 35% revenue jump to a record and Intel's comments on AI chip strength. Wall Street remains bullish on Nvidia with a Strong Buy rating and an average price target of $274.38, implying 31.3% upside.

Analysis

The key shift is not whether semis stay strong, but whether leadership broadens enough to reduce the market’s dependence on a small cluster of AI winners. When a trade moves from single-name momentum to index-wide participation, factor volatility usually falls even if the sector stays bid; that favors the lower-beta infrastructure and memory names more than the most crowded AI hardware leaders. In that setup, earnings revisions matter more than narrative, and the beneficiaries are likely to be suppliers with cleaner second-order exposure to foundry utilization, packaging, and memory pricing rather than the most expensive compute franchises.

The biggest risk to chasing the move in NVDA is not a fundamental collapse, but a multiple-compression event caused by positioning and expectations. If this is a “broader semis” phase, incremental flows can rotate out of the highest-owned winner into laggards, which caps upside for the index leader even as the tape stays constructive. That makes the next 4-8 weeks more about flow discipline than valuation debates: good numbers may not lift the top names if investors use strength to rebalance.

The contrarian miss is that AI spend can remain accelerating while the market still punishes concentration. In other words, the correct trade may be to fade extreme leadership concentration, not to fade AI capex. History says that when semis rally on earnings confirmation rather than multiple expansion, the best risk/reward often comes from owning the “second derivative” names tied to capacity, memory, and supply-chain normalization, while trimming the most crowded proxy for the theme.

Key Decisions for Investors

• Trim NVDA into strength over the next 1-2 weeks; if the stock gaps higher on sentiment, use that to reduce exposure rather than chase. Risk/reward is skewed to flat-to-modestly-upside with elevated downside if flows rotate.
• Go long MU vs short NVDA as a 1-3 month pair trade. Memory names should outperform if semiconductor leadership broadens and AI capex continues, while NVDA is more vulnerable to multiple compression from crowding.
• Add to TSM on any 3-5% pullback for a 2-6 month horizon. It offers cleaner second-order leverage to overall AI demand and foundry utilization with less valuation risk than the U.S. mega-cap leaders.
• Consider a basket long AVGO/TSM/MU against a short basket of the most crowded AI hardware names for a 6-10 week rotation trade. Target 8-12% relative outperformance if breadth improves.
• For tactical upside, use call spreads rather than outright longs in NVDA over the next earnings cycle. A defined-risk structure captures continued bullish sentiment while limiting damage if the next leg is a breadth rotation rather than another vertical move.