Nvidia Stock Won't Be Overvalued by 2028: My Case for Buying NVDA Today

Nvidia shares jumped 8.7% after strong earnings and a bullish forecast, but the article argues the valuation is still not “wildly” overvalued. It cites a mid-30s trailing P/E and low-20s forward P/E (near multiyear lows) with PEG around 0.5, versus implied revenue growth to ~$670B by fiscal 2028 (management) and ~$550B by 2028 (analyst-style estimates). Key bear risks highlighted are AI demand slowing, rising competition (AMD and hyperscaler custom chips), export controls/regulatory pressure, and potential margin pressure from cheaper inference hardware.

Analysis

The market is increasingly treating NVDA less like a cyclical chip vendor and more like a toll collector on AI capex. That matters because the real upside is not multiple expansion; it is duration of cash flows. If hyperscaler spending stays disciplined but persistent, NVDA can compound through share gains even if the AI addressable market broadens to custom silicon.

The clearest loser is AMD, but the bigger competitive risk is actually the customers building in-house silicon: they can suppress NVDA’s growth rate without needing to displace it outright. That creates a second-order effect where NVDA may still win unit economics while the rest of the semi stack loses pricing power, especially in interconnect, memory, and lower-end accelerator names. In the near term, any post-earnings pullback is likely more about positioning than fundamentals unless management commentary starts to imply margin normalization.

The key catalyst path is 1-3 months: follow-through capex commentary from the hyperscalers and any change in export-control language. The thesis is falsified if gross margin rolls below the low-70s, if forward growth steps down meaningfully before 2028, or if custom ASIC adoption starts showing up as slower data-center order momentum. The contrarian point is that the stock is probably not wildly overvalued, but it may still be over-earning its implied terminal growth rate; consensus is underestimating how quickly 70% growth math can break if AI spending shifts from training to cheaper inference hardware.

Key Decisions for Investors

• Stay long NVDA on pullbacks over the next 1-3 months; the setup is still favorable as long as forward gross margin holds and hyperscaler capex commentary remains firm.
• Pair trade: long NVDA / short AMD for 1-3 months. AMD is more exposed to share-taking optimism than to near-term earnings power, while NVDA remains the cleaner beneficiary of sustained AI capex.
• Use NVDA call spreads rather than outright equity if adding here; the valuation is supportive, but the easy rerating already happened and upside depends on continued guidance beats.
• Set a downside alert if NVDA loses the low-20s forward P/E support or if a major hyperscaler signals capex deceleration; that would be the first evidence the growth-duration story is breaking.
• If export-control headlines intensify, hedge NVDA with semiconductor exposure reduction rather than broad market hedges, since the direct sensitivity is idiosyncratic rather than macro-beta driven.