NVDA vs AMD: The DCA Backtest
The identical plan — $100 on the first trading day of each month — run on both assets over their shared history (Jan 1999 → Aug 2026, $33,200 invested each). Every number below is computed from real daily closes; nothing is annualized away or cherry-picked.
Verdict on this window: $33,200 invested into NVDA grew to $23,745,966, versus $1,812,449 in AMD. Different start dates shuffle this result — the rolling analysis below shows how often.
The same $100/month, side by side
Both positions month by month across the shared window: the dashed line is cash paid in ($33,200 total), the two curves are what NVDA and AMD were worth along the way. Where the curves separate is where the decade's story lives.
Does the winner depend on when you started?
One full-window result can mislead — maybe the winner just had a better final year. So we ran the head-to-head across EVERY possible 5-year start month in the shared history:
Side-by-side numbers
Data through Aug 2026, refreshed daily.
The two ways a semiconductor bet pays
Both charts run through the same crucibles — dot-com, 2008, the 2018 crypto-mining hangover, 2022 — and both nearly died at least once. The difference is what the recovery repriced: NVIDIA rode data-center GPUs into the AI capex boom; AMD's 100x came earlier, from a $2 near-death turnaround few held through. The rolling head-to-head shows leadership flipping with each cycle rather than one side simply winning.
Treat both numbers as tail outcomes, not expectations: these are the semiconductor survivors, measured at the top of an AI spending boom, in an industry whose history is mostly a graveyard. A monthly plan tamed the entry-timing problem; nothing tames the possibility that the next cycle's winner is neither of them.
FAQ
On the shared window measured here, NVDA finished ahead (NVDA: +71424% vs AMD: +5359% on identical monthly buys). "Better" is window-dependent — the rolling head-to-head above tells you how stable that verdict was. Past performance doesn't predict the future; this page tells you what happened, not what will.
Yes, and if the pair is highly correlated the blend will behave like either one alone — check how similar the two curves above are before assuming a split adds diversification. Our portfolio calculator lets you backtest any weighting of the two on the same real data.
Real dividend- and split-adjusted daily closes for NVDA and AMD, compared strictly over their shared history (Jan 1999 → Aug 2026), refreshed daily and validated for gaps and staleness. Fees, spreads and taxes are not modelled — see the methodology page.
Run the plan on a real exchange
Five exchanges where DCA — and tokenized stocks & ETFs — actually work, each with a real discount. Pick one, copy the code, trade cheaper.
Referral links support the site and never affect the math. Availability varies by country — check your local regulations.