Nvidia Stock Can Top $400 in 5 Years If One Assumption Holds Up

Nvidia reported fiscal Q2 2027 adjusted EPS growth of 120% year over year and revenue growth of 106% to $96.2B, while guiding fiscal Q3 revenue to about $108B, up roughly 12% sequentially. Management expects approximately 70% revenue growth in fiscal 2028, though gross margin is projected to decline from 75% to 72%-73% as memory costs rise. The company has $235B available for repurchases through fiscal 2028, potentially retiring about 4% of shares at current prices; the article argues a $400 share price by 2031 is achievable if EPS reaches $20 and post-2028 earnings remain resilient, but flags cyclical chip demand and customer concentration as material risks.

Analysis

The relevant debate is no longer whether NVDA can compound near-term earnings, but whether the market is underpricing the probability of a post-buildout utilization reset. The valuation can appear inexpensive on forward estimates while still being vulnerable: a 20-30% reduction in data-center revenue expectations, combined with a normalization of gross margin and a de-rating toward mature semiconductor multiples, would produce materially negative equity returns even if absolute earnings remain well above prior-cycle levels. The key leading indicators over the next 1-3 months are hyperscaler capex guidance, GPU lease pricing, lead times, and evidence that inference workloads are monetizing fast enough to sustain customer ROI.

Rising memory input costs create a second-order split within the AI stack. NVDA retains substantial pricing power while supply remains constrained, but margin pressure shifts incremental economics toward HBM and memory suppliers, particularly MU, and potentially packaging/advanced-foundry bottlenecks rather than system OEMs. Conversely, server assemblers and cloud customers face a less attractive return profile if accelerator pricing remains firm while memory and power costs rise, making AI capex discipline—not demand headlines—the more important medium-term risk to NVDA.

The contrarian view is that the market may be treating buybacks as a structural EPS floor when they are primarily a capital-allocation choice that can be offset by stock compensation and cannot protect the multiple during a semiconductor inventory correction. A durable five-year bull case requires recurring software, networking, and inference revenue to reduce dependence on concentrated customer procurement cycles; absent that transition, the stock should trade more like a high-quality but cyclical capital-equipment beneficiary than a secular software compounder. Over 6-18 months, evidence of broad enterprise inference demand would falsify the cyclical-risk thesis; a cut in aggregate hyperscaler capex plans or declining accelerator utilization would validate it.

Key Decisions for Investors

• Maintain NVDA as a core long only with a 6-18 month horizon, but avoid adding aggressively into strength until the next hyperscaler earnings cycle confirms capex durability. Add on a 15-20% drawdown only if management reiterates supply-constrained demand and gross-margin pressure remains contained; thesis fails on a material fiscal-year revenue-guide cut or evidence of excess channel inventory.
• Express the memory-cost spillover via a 3-6 month long MU / short SMH pair rather than a directional AI-beta purchase. MU benefits more directly from tightening high-bandwidth-memory economics, while the short leg limits broad semiconductor-factor exposure; exit if memory pricing rolls over or NVDA margin pressure proves transient.
• For downside protection around the next major cloud-capex reporting window, buy NVDA put spreads 3-6 months out rather than outright puts. The setup targets the asymmetric risk that any slowdown in META, MSFT, AMZN, or GOOGL AI infrastructure spending triggers simultaneous earnings-estimate and multiple compression; size premium at risk to a modest fraction of expected long-book gains.
• Set an alert on aggregate hyperscaler capex guidance and accelerator utilization metrics: two consecutive quarters of decelerating capex growth, or meaningful reports of GPU lease-rate declines, should trigger a reduction in NVDA exposure before consensus earnings revisions catch up.