Nvidia Is Still the King: 1 Clear Reason the Rubin Upgrade Cycle Means the Stock Is Not in a Bubble

Nvidia is positioned for an upgrade cycle tied to its Rubin architecture, which is expected to cut AI inference costs by 10x and training costs by 4x while carrying about a 25% higher price than Blackwell. Wall Street is projecting 81% revenue growth in FY2027 and 41% in FY2028, with AI hyperscaler data center capex reaching $650 billion in 2026 and potentially topping $1 trillion in 2027. The article argues NVDA remains undervalued despite its run, which is supportive for sentiment but is mainly commentary rather than new hard news.

Analysis

The market is still treating NVDA like a single-cycle hardware beneficiary, but the setup is broader: a higher-cost refresh on top of a structurally larger capex pie. That combination matters because it compresses payback periods for hyperscalers while expanding Nvidia’s average selling price, which is the cleanest way to get both volume and margin leverage at the same time. The second-order winner is the broader AI buildout ecosystem—networking, power, cooling, and foundry-adjacent suppliers should all see incremental pull-through as rack density rises and datacenter budgets shift from concrete toward compute.

The real debate is not whether demand exists, but whether the street is underestimating how quickly orders can re-rate once Rubin qualification starts. If the upgrade cycle begins before 2026 capex guidance is fully rolled in, consensus estimates for FY27/FY28 likely still lag by a meaningful margin, and that can keep forward multiples elevated even after the stock has moved. The flip side is that any delay in platform readiness, memory supply, or customer deployment schedules would hit sentiment fast because positioning is already crowded on the AI long side.

Contrarian-wise, the article’s bullishness likely understates how much of the easy upside is already tied to hyperscaler spending being converted into actual hardware orders. If a larger share of the $650B-to-$1T spend shifts to internal software optimization, custom silicon, or throttled deployment, NVDA’s revenue acceleration could still be strong but less explosive than the market is discounting. The more attractive expression may therefore be relative value: long the AI infrastructure basket most levered to rising rack density, while hedging a portion of outright NVDA beta into the product transition window.

Key Decisions for Investors

• Maintain/scale a tactical long NVDA into the next 4-8 weeks, but use call spreads rather than outright stock to cap downside from any platform-delay headline; best reward if Rubin timing remains intact is another multiple expansion leg.
• Pair trade: long NVDA / short a basket of lower-quality AI beneficiaries that depend on multiple expansion rather than direct compute demand; this isolates the upgrade-cycle winner while reducing factor beta if AI sentiment cools.
• Add exposure to power/cooling/networking beneficiaries over the next 6-12 months, as higher rack density should be a more persistent earnings driver than one-time chip shipments; these names may capture the second-order spend with less valuation risk than NVDA.
• If already long NVDA, consider trimming into strength ahead of the Rubin launch catalyst and re-entering on any post-announcement volatility; the key risk is a classic buy-the-rumor/sell-the-news setup if expectations get too far ahead of execution.