Nvidia is once again in the spotlight - and the stats show that those who held onto results for months, not days, have historically earned the most. Short-term moves after quarterly numbers tend to be significant, but in terms of return and probability of profit, they are clearly speculation rather than an investment thesis.
Since 2016, buying Nvidia just before earnings has yielded only modest median appreciation of tens to low single-digit percentages over a day, week or month, while over the course of a quarter and year, medians have gone into the double-digit tens of percentages. In other words: patience has historically paid off more than trying to hit the first earnings move.
Nvidia's upcoming $NVDA results are again dividing the market into two camps - short-term bettors looking to hit a few percent move in a few hours, and investors willing to give the story time and work with a quarter to year horizon. Yet the historical numbers clearly show that the latter have made the most money: quick reactions after earnings have been substantial and erratic, while longer holdings after earnings have had a much higher probability of success over the past decade.
Since 2016, buying Nvidia just before earnings has yielded a median return of just a few tenths to low single-digit percentages in the first day, week or month, but significant double-digit appreciation over a quarter and tens of percentages over a year. In other words - those who wanted to make Nvidia a 24-hour lottery mostly played the game with little "edge", while more patient investors saw much better numbers and a higher "win-rate" after earnings.
What history says
Yahoo Finance looked at Nvidia's post-earnings performance over the past decade (since 2016) and calculated median returns and probability of profit based on how long an investor held a position after buying before earnings.
Median return after buying just before earnings:
1 day: +0.3%
1 week: +3.3%
1 month: +0.4%
1 quarter: +11.1%
1 year: +87.6%
How many times the return was positive ("win-rate"):
1 day: 55%
1 week: 60%
1 month: 53%
1 quarter: 78%
1 year: 84%
These data show:
short-term play on first response has only slightly better odds than a coin flip
Longer horizons (3-12 months) have historically been much more like "investment" than "speculation"
Rolling average of annual post-earnings returns (calculated over 10 quarters) shows that the peak of AI supercycle euphoria - when average annual post-earnings returns exceeded 150% - is gone and we are now somewhere around 70%. Yet the pattern remains the same: those who gave Nvidia time after the quarterly numbers did best.
What the market expects from the next results
Nvidia reports Q1 fiscal year 2027 after the market closes on May 20. Analyst consensus and broker comments paint another extremely strong quarter:
Revenue:
Consensus of about $78-79 billion, implying year-on-year growth of about 75-80%
Citi and some others reckon that Nvidia may shoot past consensus by around USD 1.4 billion (i.e. sales of around USD 80+ billion)
Earnings Per Share (EPS):
The market is expecting adjusted EPS of around $1.75-$1.80, which would again represent a sharp increase from last year
Data Centers:
The key segment is expected to deliver over USD 70 billion in revenue, according to S&P Global and other sources
Some analysts expect a "beat" here as well, thanks to strong demand for Hopper generation and Blackwell's first shipments
What Nvidia will say to:
Blackwell platform ramp (timing, availability, mix with Hopper)
demand from hyperscalers (Microsoft, Google, Meta, Amazon) and big AI players
the impact of export restrictions to China on the outlook
the evolution of gross margins in an expensive HBM environment and pressure on supply chains
Expectations are thus set high - and historically, it is often at times like this that the outlook has been decided by how much further the bar has moved forward, not just the actual "beat" on the current quarter.
What movement the market has priced in: who is playing with fire
Options market data shows that investors are expecting a relatively large, but not extreme, move:
implied volatility of the current series suggests a move of around 6-7.5% either way on the day after earnings
this is roughly double the normal daily range, but close to what Nvidia has shown in the past around earnings (the long-term average one-day reaction is around 6%)
For short-term traders, this means:
A theoretically attractive opportunity for an options strategy (straddle/strangle)
practically, but also the risk that even with "good numbers" the market will go the other way first - which has happened repeatedly with Nvidia when the outlook was "just" good enough, not brilliant
In other words: anyone who wants to "bet" on one day is playing a game with high volatility and only a slight historical advantage.
Who can profit from the results - and how
1) Short-term traders (days to weeks)
Profit from the fact that implied volatility is high and the market reaction can be violent.
Strategies: short-term options, intraday trading after numbers are released, or quick swings based on guidance and futures reaction.
Risk: historical "edge" after 1 day and 1 week is small (0.3-3.3% median), win-rate 55-60% - so only slightly better than pure chance.
So in practice, these players can make a lot if they hit the direction and intensity of the move, but the stats don't hurt them as much as the hype around Nvidia would suggest.
2) Medium-term investors (about 1 quarter)
Historically the best risk/reward ratio short of a one-year horizon: median +11.1% after one quarter, win-rate 78%.
Strategically, these are investors who accept that the initial reaction may be negative, but are betting that over the course of three months the business story (AI capex, new products, next quarter) will gradually work its way into the price.
They can profit by using any "earnings shocks" (selling off after earnings despite good numbers) as a better entry point and let Nvidia "work off" the next quarterly report.
3) Long-term holders (1 year or more)
Clear winners of historical data: median return +87.6%, 84% of the time in the black after a year of holding since buying before results.
They benefit from the fact that earnings are just milestones in a long AI infrastructure story - demand for chips for datacenters, automotive and other segments is cumulative over time.
Those who have held through multiple earnings cycles have historically weathered even large swings (20-30% corrections) thanks to additional waves of AI investments. The risk is that the current supercycle will slow over time - rolling 1Y returns have already cooled - but there is no indication yet that demand for AI computing power will fundamentally fade in the coming years.
Who has more to lose
Those who combine short horizons with high leverage: Nvidia's short-term reaction stats after earnings show only a slight "edge", but extreme volatility - leverage and all-in bets make it a casino.
Those who overvalue the "perfect quarter": history shows that even with great numbers, the market sometimes looks for an excuse to realize gains ("great numbers, but not great enough").
Those who ignore cyclicality: Nvidia may be a structural winner in AI, but it's still part of the semiconductor cycle - there will occasionally be a period when hyperscaler capex hits limits and growth temporarily slows.