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

Nvidia shares rose 8.7% after strong earnings and a bullish forecast, despite the article arguing the stock still looks “richly valued yet not absurdly priced.” The piece cites fiscal-year revenue of $250B+ (+65% YoY) and data center/AI accelerators at ~$190B+ (~90% of sales), with expectations for 2028 revenue of ~$670B under management’s ~70% growth outlook. Key risks highlighted are slowing AI demand, faster competition (AMD and hyperscaler custom chips), export restrictions/regulatory pressure, and margin pressure.

Analysis

The market is no longer pricing NVDA like a hypergrowth story stock; it is pricing it like the operating system for AI spend. That matters because the multiple is now being defended by cash-flow durability rather than narrative, which makes the stock less fragile than many assume — but also more sensitive to any sign that hyperscaler capex is maturing faster than headline revenue growth suggests. The real beneficiaries are the companies whose business models remain attached to a sustained GPU bottleneck; the immediate losers are inference-leaning alternatives and custom-silicon programs that need time to prove they can displace an incumbent platform.

The key debate is not whether AI demand exists, but whether the mix shifts from scarce training capacity to lower-margin inference fast enough to compress gross margin before volume can offset it. If that happens, NVDA can still grow revenue and underperform on EPS, which is the path most likely to surprise the consensus on the downside over the next 2-3 quarters. Watch for any slowdown in order visibility, a softer read from cloud capex, or tighter export rules — those would be the first falsifiers and would likely hit AMD first on relative-share expectations, then NVDA on valuation.

Contrarian view: the market may be underestimating the duration of AI infrastructure spend, but it may also be overestimating how linear the monetization path is from here. In the next 6-18 months, the stock’s biggest risk is not a demand collapse; it is margin normalization plus crowding in the long book, which can compress the multiple even if the fundamental story remains intact. That argues for expressing bullishness with defined risk rather than chasing strength outright.

Key Decisions for Investors

• Prefer a tactical long NVDA over 1-3 months only on pullbacks, not after a post-earnings gap, because the near-term upside is already partially de-risked; use the stock as a quality-duration exposure to AI capex.
• Pair trade: long NVDA / short AMD into the next 1-2 earnings cycles. AMD is the cleaner relative loser if investors keep rewarding platform dominance and discounting share-take claims that have yet to show up in margin structure.
• If seeking convexity, buy a 6-12 month NVDA call spread instead of outright calls. The thesis is still intact, but the moved-up starting point makes defined-risk upside more attractive than paying full premium.
• Set a de-risk alert if NVDA gross margin guidance falls 200-300 bps from current expectations or if management commentary implies slower hyperscaler capex in the next print; that would be the first sign the multiple should compress.
• For longer-horizon investors, treat NVDA as a core AI infrastructure winner, but hedge with a small short in the semiconductor basket on rallies if you think the market is overpaying for the second-tier AI semis' catch-up potential.