Nvidia's 1,200% Returns Over the Past 5 Years Make It the Third-Best Stock on the S&P 500. These Stocks Have Done Better

Vertiv has risen about 1,300% over five years and Comfort Systems USA nearly 2,000%, both benefiting from AI-driven data center demand. Vertiv reported Q1 2026 sales up 30% and net income up 137%, while Comfort Systems posted $9.1B in annual sales and a backlog just under $12B. The article is bullish on the businesses but cautions that high forward P/Es of 55 for Vertiv and 47 for Comfort Systems may limit near-term upside.

Analysis

The market is effectively repricing the AI buildout as a real-asset boom, not just a semis story. FIX and VRT are levered to the same capex cycle as NVDA, but with a different latency: revenue now reflects projects already committed, while equity performance is still discounting several more years of backlog conversion. That makes both names structurally attractive, but also vulnerable to the classic infrastructure cycle problem — orders are sticky until customers finish their first wave of capacity, then growth can decelerate sharply even if end-demand for AI remains intact.

The second-order winner set is broader than the article implies. Electrical contractors, switchgear, cooling, and power-quality vendors should continue to see pricing power as the bottleneck shifts from chips to grid interconnect, thermal, and facility build-out. The hidden risk is margin compression from scale: as these companies grow faster, they often have to add labor, capacity, and working capital ahead of revenue recognition, so reported earnings can outrun free cash flow in the near term and then mean-revert when project mix normalizes.

Consensus is probably underestimating how much of the good news is already embedded in valuation. A 30-50x forward multiple is defensible only if AI data-center capex remains elevated through 2027; any pause in hyperscaler spending, utility delays, or project deferrals can compress these stocks 20-35% quickly because the market has moved from paying for growth to paying for duration. NVDA is now the cleaner quality exposure: lower forward multiple, stronger pricing control, and less direct exposure to execution risk in construction and power delivery.

The contrarian view is not that the AI infrastructure theme is wrong, but that the easiest money has likely shifted from pure buildout beneficiaries to picks-and-shovels adjacencies with less narrative saturation. Investors chasing FIX/VRT are implicitly buying perfect execution and uninterrupted capex; the better risk/reward may be in names that benefit from the same spending wave without the same valuation and sentiment crowding.

Key Decisions for Investors

• Avoid initiating new outright longs in FIX at current levels; if already long, trim into strength and use a 3-6 month horizon to reassess after next backlog and margin print, because upside is now more dependent on capex duration than on execution.
• Prefer a relative-value long NVDA / short VRT pair for the next 6-12 months: if AI capex broadens but multiples compress, NVDA should defend better on earnings quality and valuation support while VRT remains more exposed to multiple reset.
• Use pullbacks to build a smaller tactical long in VRT only on 10-15% drawdowns; the risk/reward works best if you believe hyperscaler capex stays >2025 levels for another 2+ years, otherwise valuation leaves little cushion.
• Look for long exposure in adjacent beneficiaries with less multiple stretch — electrical grid, switchgear, and power distribution names — as a cleaner way to express the same AI infrastructure thesis with lower sentiment risk.
• For hedge funds with options capability, consider put spreads on FIX or VRT 2-3 months out financed by upside calls: this expresses the view that the stocks can stay strong, but the market has already priced in most of the next leg of growth.