NVIDIA has experienced extraordinary growth over the
past 15 years, with an annualized return of 53.98%, turning a $100 investment
into approximately $63,713 today. This performance far surpasses the market,
driven by its dominance in AI GPUs and data centers. However,
its current valuation of nearly $4.39 trillion poses correction risks.
Historical Performance
NVIDIA has shown volatility
with exceptional years such as 2016 (+233.8%), 2020 (+119%), and 2023
(+233.6%), but notable declines in 2018 (-31.6%) and 2022 (-50.9%). In December
2025, it is trading around $180 USD, with a year-to-date (YTD) increase of 26%
despite recent pressures. This trajectory reflects AI innovation cycles,
outperforming indices like the S&P 500 in compound annual growth.
Key events include the launch
of Blackwell Ultra (H2 2025), Vera Rubin (2026), and Rubin Ultra (2027), along
with $60 billion in global AI infrastructure deals. Investments in sovereign AI
and partnerships such as the one with Synopsys will drive demand for GPUs and
CUDA software. Approvals to
export H-chips to China under the Trump administration will add to the growth
in 2026.
Risks and Valuation
The high valuation implies a 5.9% premium based on
DCF models, with risks of AI saturation, Chinese competition, and dependence on
the data center segment (94% of revenue). Analysts warn of a slowdown in growth
(from 78% year-over-year) and threats from custom silicon. Historical
volatility suggests caution for aggressive entries.
Investment Opinion
The current impulse risk view is valid given the
demanding valuation, but sustained growth is plausible if AI demand persists
beyond 2025, aligning with your short/medium-term technical analysis approach.
For 1-3 year horizons, prioritize corrections for entry; in the long term (5+
years), it remains attractive due to innovation. I recommend monitoring moving
averages (MA200) and volume, similar to previous analyses for stocks like
Toyota.
NVIDIA sustains its long-term growth through its
dominance in the AI chip market, with a 70-95% share thanks to GPUs like the
H100 and the CUDA ecosystem, which generates sustained demand in data centers,
representing 90-94% of revenue. The technology roadmap includes Blackwell Ultra
in 2025, Rubin in 2026, and Rubin Ultra in 2027, offering advancements such as
50 petaflops of FP4 and HBM4 memory for AI models at the trillion-parameter
scale.
Expansion into autonomous vehicles via the DRIVE
platform, adopted by automakers, and edge computing for real-time processing
are driving revenue beyond data centers. Growth in IoT for healthcare,
manufacturing, and smart cities, along with blockchain and AI sovereignty, is
diversifying risks and leveraging emerging markets with a projected CAGR of 29%
in AI chips through 2030.
Financial Strength
Record revenues of $57 billion in Q3 FY2026 (+62%
YoY), with net margins of 56% and a return of $37 billion to shareholders in 9
months, fund continuous R&D. This strength supports innovation despite
increasing competition from AMD and Intel.
Structural AI Demand
The massive adoption of AI across industries is
driving exponential demand for accelerators, exceeding market projections and
positioning NVIDIA as an indispensable leader.
Analysis of the charts and descriptive statistics
table reveals a stock with a history of explosive growth, but with distribution
characteristics that warrant caution due to asymmetry and the risk of extreme
events.
The price chart shows a very
strong upward trend, especially in recent years. The initial price, on January
4, 2010, was $0.46, and the price at the end of this study, on December 12,
2025, is $175.02, showing a 15-year return of 37,947.83%.
This return is extraordinarily high, justifying the interest in the stock, but it also suggests that the price (or return) distribution will be far from normal.
The relationship is mode <
median < mean (0.37 < 9.39 < 22.22); this is characteristic of a
positively skewed or right-skewed distribution, meaning that the tail of the
distribution is longer toward higher values. Most historical prices were concentrated at low
values (near the mode and median), while the mean is strongly pulled upward
by a few extremely high values (the recent prices).
The distribution of NVDA's historical prices differs
significantly from a Gaussian or normal distribution.
The kurtosis value of 891.91 (an extremely high value
compared to 0 for the standard normal distribution or 3 for Pearson's kurtosis)
indicates a leptokurtic distribution. Kurtosis is crucial in investment risk
management because such a high kurtosis value implies that the distribution has
"heavier tails" than a normal distribution. This means that the
probability of observing extreme events or outliers is much higher than would
be predicted under the assumption of normality.
In finance, "extreme events" refer to both
extraordinary gains and catastrophic losses. To mitigate risk, investors should
be aware that NVDA's volatility (measured by a standard deviation of 43.16986)
does not fully capture the risk, as the risk of outliers is very high. Kurtosis
forces investors to use more sophisticated risk models (such as conditional
Value at Risk or heavy-tailed distributions) instead of relying solely on
volatility or the normal distribution.
The very high kurtosis (along with positive skewness)
suggests that, although there have been large gains historically, the current
price is an outlier in itself, and the possibility of future extreme
fluctuations (in both directions) is high. The linear correlation is moderately
high, indicating that a straight line (simple linear regression) explains only
about 69% of the price variation. The correlations of the 3rd and 6th order
polynomials are extremely high (close to 100%). This demonstrates that NVDA's
growth pattern over the 15 years is not linear, but rather follows a curved
trajectory, with periods of slow growth followed by a recent parabolic
acceleration. Nonlinear (polynomial) trend models are much more suitable for
describing the behavior of this stock. The high coefficient of determination
(R²) of the third-order polynomial (98.12%) validates the use of more complex
econometric models for prediction, confirming the exponential growth pattern.
The histogram, along with the frequency polygon, is
the visual representation of the analyzed distribution.
The chart shows a high concentration of frequency in
the lower class intervals (0-22.10 and 22.10-100.22), and a sharp drop in the
upper intervals (100.22-200.22 and 200.22-300.22), even though the final prices
are at the upper end.
The histogram corroborates the positive skewness and
leptokurtosis. In a normal distribution, the highest frequency would be
centered around the mean and distributed symmetrically. In the case of NVDA,
the large mass of data is concentrated in the historically low prices, while
the few recent, high prices are found in the right "tail," confirming
the positive skewness and kurtosis of the distribution.
From a purely statistical and econometric
perspective, the analysis reveals that NVDA is a very high-risk, high-reward
(high-alpha) investment, where volatility does not capture all the risk. The
growth momentum and the high fit to the polynomial curve suggest that the
upward trend may still be dominant. The extremely high kurtosis is A severe
warning sign. In the context of financial markets, risk is concentrated at the
extremes. Such a high kurtosis value suggests that the possibility of a drastic
correction ("left tail") is real and should not be underestimated.
The current price is an extreme historical event, increasing the vulnerability
to a violent reversal.
Technical analysis, based on indicators and
oscillators, offers a unanimous conclusion of strong selling pressure,
reinforcing the caution suggested by the extreme kurtosis and skewness of the
statistical analysis.
The results are clear and consistent with an
overbought situation or an imminent reversal. The RSI (14), with a value of
33.035 indicating selling pressure, although not yet at the oversold threshold,
is approaching it, indicating that the price has weakened. An RSI below 50
suggests that selling pressure is greater than buying pressure.
The MACD (12, 26) at -1.62, also indicating a sell
signal, is negative, meaning the short-term moving average (12) is below the
long-term moving average (26), a fundamental bearish crossover.
The STOCH (9, 6) at 20.866, also indicating a sell
signal, and the Williams %R = -95.841 (oversold), indicate that the price has
fallen to the lower end of its recent range. In the context of a long-term
uptrend, this could signal a temporary correction and a potential opportunity,
but only if the underlying trend holds. At this time, the immediate signal is
bearish pressure.
The technical indicators (RSI, MACD) show the current
momentum. A negative MACD and an RSI below 50 are classic signs of weakening
upward momentum.
Previous statistical analysis showed that the current
price ($175.02) was an extreme event (tail of the distribution). The fact that
the cumulative probability has covered 99.98% of the historical data confirms
that the price is at the 99.98th percentile; that is, 99.98% of NVDA's
historical prices were at or below this point.
Statistically, after reaching a point so far from the
historical mean, the probability of the price reversing toward the mean or
experiencing a sharp correction increases dramatically. This reinforces the
extreme kurtosis warning that the stock is highly susceptible to sharp
movements, with a downward move being the most likely from this position.
The closing price ($180.93) is above all key pivot
points ($176.45). However, the current
price is approaching support levels (S1, S2, S3), indicating a decline from the
previous close, which corroborates selling pressure.
The beta of 2.28 indicates that NVDA stock is 2.28
times more volatile than the broader market (S&P 500 index, typically beta
= 1). This quantifies the systematic risk and confirms that NVDA is a high-risk
asset.
Given the current data, this is not an opportune time
for a new purchase. The strong sell consensus from moving averages and
indicators, combined with the statistical risk of a heavy tail (kurtosis),
suggests that the stock is in an active downside correction or imminent
consolidation phase.
If you do not own the stock, avoid buying until the
indicators show a rebound (RSI above 50, MACD crossover) (positive or the
short-term moving averages cross above the long-term moving averages).
If you own the stock, consider protecting your gains
or reducing your position, as the risk of a significant pullback is extremely
high, validated by both the 99.98% probability and the technical warning.
P/E Analysis: NVDA's P/E ratio of 43.6x is
significantly higher than the historical average for the semiconductor sector,
which tends to reverse when it exceeds 25x. This suggests that, based on its
current net earnings, NVDA is expensively valued or overvalued relative to the
overall market and the sector's history.
A
high P/E ratio is only justified if investors expect future earnings per share
(EPS) growth. The market is paying a substantial premium today for those future
earnings. Such a high P/E ratio makes the stock extremely sensitive to any
disappointment in future results. If NVDA fails to meet high growth
expectations, the price could fall sharply, which is already implied by the
kurtosis warning (risk of extreme events).
Although
high, some market analyses suggest that NVDA is fairly valued when compared to
direct competitors with explosive growth (such as AMD, which has traded at even
higher multiples) or when compared to its own projected fair P/E ratio, which
could be higher.
EV/EBITDA
is a preferred metric for some analysts because it ignores the impact of taxes,
interest expenses, and depreciation/amortization (D&A) providing a clearer
picture of operating profitability. It is especially useful for comparing
companies with different capital structures and tax regimes.
