Chip Stocks Show Diverging Trends, Says Analyst

Bernstein’s Stacy Rasgon said AI demand is still outpacing supply across semiconductors, with shortages in memory, optics, CPUs and other compute components, supporting continued strength in the sector. He also flagged a valuation disconnect, noting NVIDIA and Broadcom trade at roughly 15x earnings or less while suppliers’ stocks have surged, suggesting the gap may eventually narrow. In premarket trading, Nvidia fell 1.24% to $199.18, Broadcom 1.40% to $400.86, Intel 1.24% to $67.65, and AMD 1.09% to $275.35.

Analysis

The key second-order signal is not simply “AI demand is strong,” but that the bottleneck is migrating toward the most capacity-constrained nodes in the stack. That tends to create a self-reinforcing capital cycle: suppliers with near-term scarcity get rerated first, then customers with pricing power and long-duration demand reassert leadership once investors separate temporary margin transfer from durable earnings growth. In that regime, the market often overpays for the scarcest inputs while underestimating the equity-duration value of the platforms that monetize the end-demand.

The more interesting setup is that the current rotation can persist for several weeks if inventory cover remains tight and every incremental data point confirms supply strain. But over a 3-6 month horizon, the trade becomes fragile if hyperscaler capex growth slows even modestly, because the market is already rewarding upstream names for peak scarcity economics. That creates a classic reversal risk: if order growth decelerates before physical supply normalizes, the highest-multiple beneficiaries of the shortage can de-rate quickly while the largest platforms re-accelerate on relative earnings consistency.

Consensus appears to be missing that the biggest divergence is likely within semis, not between semis and the broader market. If AI revenue continues to compound while supply remains constrained, the right expression is not a blanket long-beta semiconductor basket, but a selective long/short targeting names with pricing power and duration versus names whose upside is mostly tied to transient bottlenecks. The market is pricing the scarcity, but not yet fully pricing the eventual normalization of supply-chain rents.

A secondary risk is policy and customer behavior: prolonged shortages can accelerate custom silicon, multi-sourcing, and architectural substitution, especially if AI labs and hyperscalers decide to internalize more of the stack. That is a 12-24 month threat to merchant suppliers with the most visible scarcity premium. In the near term, though, the trade is still about who captures the margin during the squeeze, not who wins the end-market in the steady state.

Key Decisions for Investors

• Long NVDA / short a basket of constrained-input beneficiaries that have already rerated on scarcity, using a 3-6 month horizon; seek 2:1 downside/upside skew if supply normalization starts to show up in lead times before demand rolls over.
• Add to AVGO on pullbacks as a higher-quality AI monetization vehicle versus pure scarcity trades; target 6-12 month duration, with lower multiple compression risk than the most crowded AI supply names.
• Avoid chasing short-term strength in memory and equipment; if using them, express via call spreads or smaller sizing because the squeeze can persist for 1-2 quarters but reverses fast once capacity comes online.
• Pair long NVDA or AVGO against short INTC/AMD on a relative-value basis only if you are targeting divergence in AI monetization quality over the next 1-2 quarters; stop out if broad semis leadership returns and valuation dispersion narrows.
• Buy 3-6 month downside protection on the most crowded AI supply-chain names after sharp rallies; implied vol should be more attractive than outright shorting if the market remains momentum-driven in the near term.