Got $3,000? 2 AI Infrastructure Stocks to Buy Before Earnings Season Kicks Into High Gear.

Nvidia and Dell are positioned to benefit from sustained AI infrastructure demand, with Nvidia reporting $62 billion of data center revenue last quarter, up 75% year over year, and Dell citing $43 billion of remaining AI-server orders. Nvidia expects more than $1 trillion in cumulative orders for Blackwell and Rubin chips through 2027, while Dell guided AI revenue to $50 billion this year. The article frames both stocks as attractively valued relative to their growth and suggests upcoming May earnings could act as catalysts.

Analysis

The investable signal is less about headline AI demand and more about who owns the bottlenecks that convert demand into shipped systems. NVDA remains the cleanest operating leverage on compute scarcity, but the second-order winner is the ecosystem around power, networking, and integration: every incremental GPU deployment pulls through more switches, optics, rack infrastructure, and memory, which keeps pricing power broad and delays any normal margin mean reversion. That also means the build-out is still supply-led rather than demand-led; when capacity is constrained, the strongest balance sheets and best allocation to scarce components compound fastest.

The market is likely underestimating the duration of this cycle because it is pricing these names like cyclical hardware instead of capacity-locked infrastructure. For NVDA, the key risk is not demand exhaustion in the next quarter but a slower-than-expected monetization of the order book if inference efficiency improves materially or if platform customers intensify multi-vendor sourcing over the next 6-12 months. For DELL, the bigger hidden lever is mix: backlog conversion can create a very sharp earnings ramp if component shortages ease, but if memory tightness persists, revenue can still grow while margins disappoint, which would cap the rerating.

The contrarian angle is that consensus may be overconfident on a straight-line AI capex expansion. A 50% capex increase by hyperscalers sounds huge, but the market cares about whether that spending is incremental or simply reallocated from other server and networking budgets; if budgets are cannibalized, beneficiaries rotate rather than broaden. Also, any sign that AI workloads are being optimized to do more with fewer chips would pressure the ‘scarcity premium’ embedded in NVDA and, to a lesser extent, the entire rack-scale supply chain.

Net: this is still a momentum-friendly setup into the next earnings prints, but the better trade is to express relative strength versus lagging legacy hardware rather than chase outright beta. The near-term catalyst window is days to weeks around results and guidance; the more durable thesis is 6-18 months if backlog conversion and order visibility remain intact. The most likely failure mode is not a collapse in demand, but a valuation reset from expectations getting ahead of supply-chain execution.

Key Decisions for Investors

• Maintain a core long NVDA into earnings with a 4-8 week horizon; upside is strongest if guidance confirms order visibility beyond the current build cycle, but trim 20-30% if the market starts discounting slower-than-expected gross margin expansion.
• Initiate a long DELL / short legacy hardware pair over the next 1-3 months; thesis is that AI server backlog converts faster than PC/peripherals recover, creating a valuation rerating if execution stays clean. Risk is component scarcity or mix pressure delaying margin lift.
• Use call spreads on NVDA rather than outright shares for event exposure: 1-2 quarter tenor, targeting a post-earnings re-rating if the stock has been range-bound. This caps downside if the market decides the AI trade is crowded.
• Consider a basket long of AI infrastructure suppliers versus semis with weaker AI exposure; the second-order winners from the build-out may outperform if capex broadens to networking, power, and rack systems. Reassess if hyperscaler capex growth slows below expectations.
• Avoid chasing INTC as a sympathy beneficiary; absent evidence of meaningful AI share gains, it remains a lower-conviction way to own the theme and is more exposed to execution risk than to the demand impulse.