AI will drive Nvidia higher by more than 40% from here, says Wells Fargo

Wells Fargo raised Nvidia's price target to $315 from $265, implying 44% upside from Monday's close, and reiterated an overweight view. The bank cited continued compute demand outpacing supply, Blackwell AI platform adoption, and a projected AI infrastructure pipeline of more than $1 trillion by 2027. Nvidia is described as trading at under 20x P/E on durable 2027 estimates, with 57 of 61 analysts maintaining buy or strong buy ratings.

Analysis

The key takeaway is that this is not just a multiple-expansion story; it is a capital-intensity compounding story. If compute demand remains above supply through the next several quarters, the real economic moat shifts toward vendors that can finance, package, and deploy full-stack systems at scale, which should continue to concentrate share in NVDA while pressuring smaller GPU-adjacent suppliers that lack system-level leverage.

The second-order winner is likely the infrastructure ecosystem around power, networking, and thermal management rather than chip fabs alone. As AI clusters get built in larger blocks, the bottleneck moves from silicon availability to grid interconnects, liquid cooling, optics, and memory bandwidth, which should create a rising tide for select equipment and component names even if unit growth in GPUs normalizes later in 2025. That also means the market may be underestimating how sticky revenue becomes once customers have committed to multi-quarter deployment schedules.

The main contrarian risk is that the trade is becoming crowded precisely because it has a clean narrative and visible analyst support. If delivery cadence slips, gross margin expansion slows, or hyperscaler capex pauses for even one quarter, the stock can de-rate quickly because expectations are now anchored to a very high bar for continuous execution. On a 3-6 month horizon, the issue is not demand collapse but any sign that supply catches up faster than pricing power or that customer mix shifts toward lower-margin inference deployments.

Key Decisions for Investors

• Stay long NVDA into the next 1-2 quarters, but favor call spreads over outright stock to reduce headline and multiple-risk; risk/reward remains attractive if AI deployment remains supply-constrained.
• Add a basket long in AI infrastructure beneficiaries with cleaner second-order leverage than semis alone: long VRT/ANET/GLW on a 3-6 month horizon, because bottlenecks in power, networking, and optics should persist even if GPU growth moderates.
• Pair trade: long NVDA vs short a broad semiconductor basket or a slower-growth legacy hardware name over 2-4 months; the trade is that NVDA should keep compounding earnings revisions while the rest of semis face more normal cyclicality.
• Use any 5-8% pullback in NVDA to add, but trim if the market starts pricing in flawless 2027 assumptions more than 12 months early; at that point upside becomes more dependent on sentiment than fundamentals.
• Avoid shorting WFC on this note alone; the more actionable read-through is for sell-side estimate revisions in AI capex beneficiaries, not for a direct bank trade.