What if you bought Nvidia in 2016, when it began turning into an 'AI company'?
This uses real data to show what would have happened if you had invested a lump sum in Nvidia (NVDA) on May 27, 2016 — the launch day of the Pascal-based GeForce GTX 1080, when Nvidia began being re-rated from a gaming company into a deep-learning and AI-infrastructure company — and held until now. As a single stock, be mindful of survivorship bias.
⚠️ Know the risk first
Even after entry, Nvidia fell sharply from its peak in late 2018 and again in the 2022 bear market. The final return only holds if you held through those large drawdowns without selling.
⚠️ This asset is an individual stock chosen as of today, so past returns can look better than they really were (survivorship bias).
What happened that day
On May 27, 2016, Nvidia's Pascal-architecture GeForce GTX 1080 went on sale. The Pascal generation sharply raised deep-learning training performance and is often cited as the starting point for Nvidia's re-rating from a gaming company into an AI-infrastructure company.
Why this date
The buy date is May 27 — when the Pascal GPU became a tangible product on sale — rather than the May 6 announcement, computing the question 'what if you bought the day the hardware of the AI era entered the world.'
Investment conditions
Asset · Nvidia (NVDA) · lump-sum on the 2016 Pascal launch day, then held long-term
Method · Lump-sum (all at once)
Period · 2016-05-27 ~ 2026-07-01
Amount · $7,407
As of · 2026-07-01
Key metrics
These results do not reflect taxes. Based on historical data.
Buy and final prices are shown in the asset's local currency (US & crypto $, Japan ¥, Korea ₩). Total invested and final value are in Korean won (₩).
Risk & recovery
As important as returns. This service does not hide maximum drawdown or loss periods.
Maximum drawdown (MDD)
-66.3%
Largest drop from peak
Longest loss period
0months
Months in loss: 0
Recovery period
8months
Growth over time
Invested principal (dashed) and portfolio value (solid). Values below match the calculation.
Total invested $7,407 → Final value $1,300,656 (+17458.9%), Maximum drawdown (MDD) -66.3%
Why this period and asset
May 27, 2016 is the day Nvidia's Pascal-architecture GeForce GTX 1080 went on sale. At the time Nvidia was still seen as 'a graphics-card company,' but the Pascal generation sharply boosted deep-learning training performance and became essential hardware for AI researchers. Long before ChatGPT (2022) brought the AI boom mainstream, many now look back on this moment as a 'dawn of AI.' Keep in mind, though, that even after this the stock swung sharply several times, including a late-2018 plunge and the 2022 bear market.
Interpreting the result
If you bought at this point and never sold, the final value and annualized return (XIRR) below come out very large. But that number is only the result of 'if you never sold.' Nvidia fell hard from its peak in late 2018 and again in 2022, and if fear had pushed you to sell in those stretches, the story changes completely. Investing the same money gradually each month instead of all at once would have reduced the risk of buying everything at one point — use the 'lump-sum vs monthly' link below to compare the methods.
Caveats & limits
Be especially mindful of survivorship bias. Nvidia looks clearly like the top beneficiary of the AI boom in hindsight, but in 2016 it was one of many semiconductor and tech stocks, and no one could be sure it would turn out this way. A flashy-looking result here may be the illusion of picking only a survivor that grew enormously. Figures use adjusted close (dividends and splits reflected) and exclude exchange rates, trading fees, and taxes. Always review the maximum drawdown and loss periods.
Event fact sources
- NVIDIA GeForce Newsroom — GeForce GTX 1080 announcement (announced 2016-05-06, on sale 2016-05-27)
- VideoCardz / Electronic Design — NVIDIA Pascal GTX 1080 launch coverage (2016-05)
Requested date vs actual trading date
If the event date is a market holiday, the fill uses the next trading day's close. The 'effective trading date' below is the date actually used.
Data sources & limits
- Price data source: Yahoo Finance / FinanceDataReader
- Collected on: 2026-07-28
- Effective trading date: 2026-07-01
- Price basis: Adjusted close (reflects dividends and stock splits)
- This asset is an individual stock chosen as of today, so past returns can look better than they really were (survivorship bias).
- Trading fees and taxes are not reflected — figures are pre-tax.
- Based on historical data; does not guarantee future returns.
Frequently asked questions
Why use May 27, 2016 as the buy date?
That is the day the Pascal-based GeForce GTX 1080 actually went on sale. The Pascal generation greatly improved deep-learning performance and is often cited as the starting point for Nvidia's re-rating into an 'AI-infrastructure company.' We use the actual on-sale date, not the May 6 announcement.
Were there loss periods after buying?
Yes. Nvidia plunged in late 2018 on weak crypto and gaming demand, and fell sharply again from its peak in the 2022 bear market. You can see those stretches in the maximum drawdown metric below. The final return only holds if you held through those declines without selling.
Could you have known in 2016 that Nvidia would soar on AI?
No. It looks obvious in hindsight, but at the time Nvidia was one of many semiconductor and tech stocks. Picking the past while already knowing the outcome makes a single stock look safer and flashier than it really was — that is survivorship bias.
Are exchange rates and taxes reflected?
No. Exchange rates, trading fees, and taxes are not reflected — figures are pre-tax. In real U.S.-stock investing, USD/KRW movements and taxes would additionally affect results. This page only shows historical data and recommends no stock.
Related scenarios
📋 Results are based on historical data; past returns do not guarantee future returns.
📋 This service is provided for educational purposes to help you understand investing, not as investment advice.
⚠️ Calculated using today's representative assets, which may differ from the market composition at the time.