Nvidia jumps 6% in premarket trading after blockbuster earnings boost AI confidence

Nvidia shares rose about 5.95% premarket after CFO Colette Kress reiterated fiscal 2028 revenue growth of ~70% (Feb 2027–Jan 2028), supported by “much greater” AI demand—though the company remains supply-constrained by TSMC and memory-chip shortages. Nvidia also reported $40.3B in AI Clouds, industrial, and enterprise (ACIE) sales in the quarter, up 138% year over year, reinforcing a diversifying customer base beyond hyperscalers. Analysts flagged a potential near-monopoly “threat” from hyperscaler custom chips, but management’s supply-chain visibility and stated AI inflection point calmed investor capex/return fears.

Analysis

Near term, this is a squeeze catalyst for the entire AI compute complex, but the cleaner expression is still NVDA because it controls the scarce bottlenecked layer of the stack. The more interesting second-order winner is memory/HBM: if systems remain supply-constrained, the incremental dollar of AI capex shifts toward components with the tightest availability, which supports pricing and mix for memory suppliers even when unit growth at the chip designer is capped by packaging and wafer capacity.

TSM is a more nuanced beneficiary. Tight foundry capacity preserves utilization and pricing power, but it also means the market cannot fully monetize the demand step-up, so the stock may lag on a relative basis if investors rotate toward the company with the stronger operating leverage to demand. Over 1-3 months, the key catalyst is whether hyperscaler budgets and order books inflect higher; if they do not, this becomes a sentiment pop rather than a durable rerating.

The contrarian risk is that the market is overpricing the near-term threat from custom silicon. Custom ASICs can take share at the margin in inference, but they do not displace the software moat, ecosystem lock-in, or full-stack training demand quickly enough to matter in the next several quarters. What would break the thesis is any sign of capex discipline from the large cloud buyers, easing lead times at TSM, or a rollover in memory pricing that signals the supply chain is no longer tight.

Key Decisions for Investors

• Long NVDA on pullbacks over the next 1-3 weeks; use the post-print momentum while supply constraints keep revisions biased upward. Risk/reward is favorable as long as the stock holds the earnings gap and next-quarter guide does not soften.
• Pair trade: long NVDA / short TSM for 1-3 months. Thesis: NVDA captures the economic rent from demand visibility, while TSM’s upside is capped by capacity constraints and lower operating leverage to the end-demand surprise.
• Watch/accumulate MU or a memory basket on confirmation of HBM tightness. If AI system shortages persist, memory pricing can stay stronger for longer than consensus expects, creating a secondary winner with less headline risk than NVDA.
• If custom-silicon headlines accelerate, express it with a relative short in AMD versus long NVDA rather than an outright NVDA short. The moat is more likely to compress competitor share than to impair NVDA’s near-term multiple.