SweetValueLab
Write for us

NVDA backtest: ten years of NVIDIA

What a lump sum in NVIDIA would have done over the last decade, measured on daily closes — including the drawdowns along the way, which is the part a headline return never shows.

HoldingsWeight · share
100%
Over the last
Starting with
Adding every month
Compare with

Up to 10 holdings and 20 years. Weights are relative. The comparison line gets the same money on the same days.

Running the backtest…
At a glanceNVDA
SectorSemiconductors
ListedJanuary 1999, NASDAQ
Recent splits4-for-1 (Jul 2021), 10-for-1 (Jun 2024)
PaysA token quarterly dividend

Fund facts as of 2026-09-06. Holdings counts change at each index rebalance.

The return is the easy part

Anyone reading this page already knows NVIDIA went up. What a backtest adds is the shape of the path — and the path is where the story stops being comfortable. The maximum drawdown below is measured on daily closes, so it is the real peak-to-trough a holder sat through, not the tidier figure you get from month-end data.

Read that number as a duration, not a percentage. A fall of that size takes months to happen and longer to recover, and every one of those days offered a reason to sell.

Two splits, and why the curve has no cliffs

NVIDIA split 4-for-1 in July 2021 and 10-for-1 in June 2024. On a raw price chart both look like the stock collapsed overnight. Prices here are split-adjusted, so the curve runs through them continuously — which is what actually happened to anyone holding the shares, since a split changes the share count and the price together and leaves the position untouched.

What this page cannot tell you

  • It was chosen with hindsight. A page for the semiconductor company that did not work out would look very different, and nobody searches for it. Read the result knowing which company you picked and when.
  • One window is one sample. Move the start date above by a couple of years and the annualized figure moves a long way. That sensitivity is a property of a single volatile stock, not a flaw in the test.
  • Dividends are noise here. The payout exists but is negligible against the price move — the exact inverse of the dividend funds elsewhere on this site.

Sources

Sector, listing date and split history from Financial Modeling Prep's company profile and corporate-actions feeds. Prices are their split-adjusted end-of-day closes.

How this backtest works

Daily bars, not monthly

The portfolio is marked to market every trading day, so drawdowns and volatility reflect what actually happened between month-ends — 252 observations a year instead of 12.

Split-adjusted prices, cash dividends

Prices are split-adjusted closes; each dividend is applied as cash on its ex-date at the matching split-adjusted per-share amount. Dividend-adjusted (“total return”) prices are never used, so income is counted exactly once.

A lump sum, or money every month

The weights set the opening trade and then drift, unless you rebalance each January. A monthly amount buys more at the target weights on each month's first trading day, and the comparison line gets the same. No taxes, no commissions.

The window is the overlap of every holding's price history, and a dividend cut is only reported when both the annual total and the average payment fell. Full methodology on the backtest tool.

Questions

Does this account for NVIDIA's stock splits?

Yes. Prices here are split-adjusted, so the curve runs continuously through the 4-for-1 in July 2021 and the 10-for-1 in June 2024. A raw price chart shows two cliffs on those dates that nobody holding the shares actually experienced.

Why is the drawdown so much larger than the S&P 500's?

Because it is one company, not five hundred. A single semiconductor stock carries the industry's cycle with nothing to average it out, and the figure below is measured on daily closes — the real peak-to-trough, not the smoothed month-end version.

Do NVIDIA's dividends matter here?

Barely. NVIDIA pays a token quarterly dividend and the summary below states its share of the total gain. On a position that moved this much the payout is a rounding error — the mirror image of the dividend funds elsewhere on this site.

Page last updated . The backtest above runs on today's data each time the page opens. Open the calculator empty.