If I Could Only Buy 1 Artificial Intelligence (AI) Stock, This Would Be It

Nvidia CFO Collette Kress indicated the company expects 70% revenue growth in fiscal 2028, ending January 2028, materially above Wall Street's prior growth expectations. The bullish thesis rests on AI-hyperscaler data-center capex projected at $1.3 trillion next year and Nvidia's estimate of $3 trillion to $4 trillion in annual AI data-center spending by 2030. Despite this outlook, the article notes Nvidia trades at 24x forward earnings versus the S&P 500's 24.8x trailing multiple, arguing that substantial AI-driven growth remains underpriced.

Analysis

The actionable issue is not headline valuation but the durability of consensus earnings revisions. NVDA can outperform even without multiple expansion if hyperscaler capex converts into accelerated revenue estimates; conversely, a seemingly modest forward P/E is illusory if customer concentration, product-transition costs, or custom-silicon share gains force FY28 estimates lower. The next 1-3 months hinge on quarterly supply commentary, gross-margin trajectory during platform ramps, and whether the largest cloud customers signal incremental rather than merely reallocated AI budgets.

AVGO is the principal competitive hedge rather than a clean read-through beneficiary. Greater custom-ASIC adoption lowers customers' cost per inference and can reduce the portion of AI spend addressable by merchant GPUs, but it may also expand total compute deployment by making inference workloads economical. A widening NVDA/AVGO relative-performance gap should therefore be interpreted as a market verdict on proprietary GPU system economics versus customer-designed silicon, not simply AI-demand strength.

Consensus appears too willing to extrapolate both demand and margins simultaneously. The structural risk over 6-18 months is that model efficiency, lower-cost inference, and internal ASIC programs increase unit demand but compress accelerator content per dollar of AI capex; that scenario favors networking, memory, foundry capacity and custom-chip suppliers over a pure GPU thesis. This article itself creates no new fundamental catalyst, so avoid chasing a sentiment-driven opening move; require confirmatory capex and order-backlog evidence.

Key Decisions for Investors

• Maintain or initiate a measured NVDA long only on post-earnings confirmation that forward revenue estimates are being revised higher while gross-margin guidance remains stable; target 12-18% upside over 3-6 months versus 8-10% downside if next-quarter guidance or margin outlook disappoints.
• Use a 3-6 month relative-value position: long NVDA / short AVGO in equal dollar amounts only if NVDA demonstrates sustained estimate revisions and AVGO's custom-AI growth outlook does not accelerate. Exit if AVGO outgrows expectations or the relative spread moves 10% against the position.
• For existing concentrated NVDA exposure, buy 3-6 month downside puts around earnings rather than trimming solely on valuation rhetoric. The hedge is justified by asymmetric gap risk from a hyperscaler capex pause, a gross-margin reset, or evidence that ASIC deployments are displacing GPU clusters.
• Set a watch trigger, not a trade, around public 2027-28 capex commitments from Microsoft, Alphabet, Amazon and Meta. Broad-based upward revisions would validate maintaining AI-semiconductor beta; any two major customers guiding capex toward efficiency or internal silicon should prompt reduction of NVDA exposure and reassessment of AVGO relative positioning.