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

Nvidia reported fiscal Q2 adjusted EPS growth of 120% year over year to $2.22 and revenue growth of 106% to $96.2B, while guiding fiscal Q3 revenue to roughly $108B, up 12% sequentially. Management expects about 70% revenue growth in fiscal 2028, although gross margin is projected to decline from 75% to 72%-73% amid higher memory costs. The bullish $400 five-year stock case relies on EPS reaching $20 by 2031, supported by a $235B repurchase capacity that could retire about 4% of shares, but remains exposed to a post-2028 AI-chip cycle downturn and concentrated hyperscaler demand.

Analysis

The key underwriting issue is not near-term accelerator demand but the durability of hyperscaler return on invested capital once capacity constraints ease. A shift from supply-limited to demand-clearing conditions would expose both customer concentration and the risk that internally designed ASICs from AMZN, GOOGL, META, and MSFT absorb the lower-value inference workload; Nvidia can retain leadership while still face lower unit growth and mix-driven gross-margin pressure. The cited buyback is less supportive than its headline size implies: net dilution, rather than gross authorization, determines EPS accretion, and repurchases executed at elevated multiples can amplify downside if the cycle rolls over.

Over the next 1-3 months, consensus estimates likely remain supported by backlog conversion and tight supply, limiting the case for an outright directional short absent evidence of order deferrals. The 6-18 month catalyst path is more asymmetric: memory and advanced-packaging cost inflation can reduce incremental margins before revenue decelerates, while hyperscaler capex commentary is the earliest observable demand signal. The contrarian view is that the market may be underwriting a smooth transition to mature, recurring AI infrastructure spending; semiconductor cycles typically turn when lead times normalize, not when reported revenue first weakens.

A sustained long case requires customer AI monetization to broaden beyond training spend and net share count to decline materially. The thesis is falsified for bulls if hyperscaler capex guidance slows sequentially, gross-margin guidance fails to stabilize after the projected cost reset, or net share count remains flat despite repurchases; any two would justify a lower terminal multiple and materially reduce the probability of the implied long-duration EPS path.

Key Decisions for Investors

• Maintain NVDA as a trading long only through the next earnings/capex-readout window; do not add solely on buyback optics. Reduce exposure if gross-margin guidance moves below the low-70% range or if two major hyperscalers guide AI capex lower sequentially.
• For a 6-12 month normalization hedge, consider a modest long TSM / short NVDA pair. TSM retains foundry-volume exposure if accelerator demand remains healthy, while NVDA carries greater risk from ASP, mix, and valuation compression; reassess if Nvidia demonstrates durable margin stabilization or TSM guides leading-edge utilization lower.
• Use defined-risk downside protection rather than an outright NVDA short: buy 12-18 month put spreads financed partly by selling upside calls only against an existing long position. The structure targets the post-supply-constraint transition while preserving participation if backlog conversion remains stronger for longer.
• Set an alert around disclosed net diluted-share-count trends at the next two reports. If annualized net reduction remains near zero despite repurchases, remove buyback-driven EPS growth from estimates and tighten valuation discipline on any rally.