Nvidia Is Officially the Largest Stock in the World. Is the Artificial Intelligence (AI) Giant Still Cheap?

Nvidia expects AI hyperscaler data center capex to rise to $1 trillion or more in 2027, supporting another year of significant growth after 2026's record $650 billion spend. The stock trades at 23.5x forward earnings, only slightly above the S&P 500's 22x, while Wall Street still sees 41% revenue growth next year. The article argues that 2027 growth is not yet priced in and frames Nvidia as a relatively cheap AI leader despite its $5.1 trillion market cap.

Analysis

The market is treating hyperscaler AI capex as a straight-line expansion, but the more important implication is that spend is migrating from experimentation to industrialization. That usually favors the dominant platform vendor first, then compresses returns for everyone else once procurement standardizes and pricing power shifts from chip performance to supply assurance, interconnect, and software lock-in. NVDA’s setup is therefore less about one more beat and more about whether the 2027 spend cycle becomes a durable multi-year fleet refresh, which would extend the revenue runway beyond the market’s current discount horizon.

The second-order winner is not just NVDA but the entire AI infrastructure stack that can monetize the ripple effects: power equipment, networking, memory bandwidth, and data-center construction all get pulled forward if the capex trajectory is real. The risk is that investors are underestimating how quickly hyperscalers can substitute mix away from premium accelerators once inference economics matter more than training bragging rights. If that happens, the beneficiary set broadens while NVDA’s incremental dollar growth decelerates faster than consensus expects.

The contrarian point is that a multiple near the market’s implies the crowd is already pricing “good, but not exceptional” after 2026; the real debate is not valuation, it is duration. If 2027 orders are already visible, the stock can rerate on confirmation, but if the capex narrative slips even one quarter, the derating could be sharp because expectations are concentrated in a narrow window. That makes this a catalyst-driven long with asymmetric downside if hyperscaler commentary turns more disciplined on ROI, depreciation, or power constraints.

Key Decisions for Investors

• Maintain a tactical long NVDA into the next hyperscaler capex updates, but express it with a 3-6 month call spread rather than stock to limit multiple-compression risk if 2027 guidance is delayed.
• Pair long NVDA vs short a basket of AI adjacency names with weaker platform control over demand capture; the thesis is that capital is consolidating toward the most indispensable supplier, not the broad ecosystem.
• Add to GOOGL on pullbacks as a secondary beneficiary of the AI build-out, but size smaller than NVDA: the risk/reward is better as a cash-generative participant than as a primary capex lever.
• Use any post-earnings spike in NVDA to trim core exposure and re-enter on 5-8% drawdowns; the stock is likely to trade on order visibility rather than near-term earnings, so entries matter more than direction.
• Monitor for a rotation into power and networking names over the next 6-12 months; if AI capex truly moves toward $1T+, the cleaner second-order upside may sit in infrastructure suppliers before it shows up in semiconductor multiples.