NVDA Px Targets vs. Upside Px Probabilities via OpenBB

December 10, 2025

We are approaching year-end amid widespread discussion of a potential “AI Bubble”. Let’s run the OpenBB Workspace facilitated analysis we posted on for AAPL this last summer (video embedded at end of note)¹ for NVDA, to see if there is any unusually large disconnect between the market and fundamental based expectations for this core AI-related name.

• If analyst price targets for NVDA are unusually low relative to NVDA’s model based price probability percentiles that could indicate that NVDA’s stock price is still overextended, despite sitting ~15% below all time highs.
• Conversely, if analyst price targets are unusually high relative to the model price probability percentiles, it could signal that sentiment for NVDA remains too strong, leaving price vulnerable to further downside.

Setting up the analysis:

Just as we did in August for AAPL, we bring VecViz’s Vector & Sigma VaR / OaR History widget and the OpenBB provided “Price Target” widget into the Workspace and add them to the context for OpenBB’s co-pilot.

OpenBB’s AI Co-Pilot output follows below:

We easily converted the table generated by the co-pilot to a line chart using the table settings. The first below chart depicts the % of NVDA Analyst Price Targets published in the 6 months prior to each date that are above that date’s Vector Model (blue) and Sigma (green) 63d forward 95th %tile to the upside (i.e., 95% OaR) from 6/30/2023 upward. The second chart does the same for the 63d forward 99%tiles to the upside (with the Vector Model in orange and Sigma in blue)

Assessing the most recent 6 months of NVDA price targets relative to current 95th and 99th %tiles to the upside.

We can tell from the charts above that NVDA’s price targets not exceptionally high or low (on a historical basis going back to mid 2023) relative to the price probability percentiles, though they are a bit elevated on a historic basis relative to the Vector Model 99%tile (the orange line).

For the sake of clarity we asked the OpenBB co-pilot to place the most recent reading for % of Targets above each threshold into historic context statistically via the “Z-Score”:

We toggled the chart to a table and exported the data to csv to check the math in Excel²:

The key point we note here³ is that analyst price targets are at most at 1.44 standard deviations above average relative to price probability percentiles. On average, across the four thresholds, the Z score was 0.85. A bit elevated but far from exceptional.

Analyzing what the current level of recent Price Targets in excess of the price probability %tiles implies for NVDA going forward.

We then asked the co-pilot to generate forward 21d (1 month) and 63d (3 month) returns for NVDA:

We checked the rolling forward 21d and 63d returns for accuracy , just as we checked the Z-Score calculations, and they were likewise correct.

Then, we plotted them against the Price Target – Price Probability breach data. Below we display perhaps the most interesting chart – the 3 month forward return vs. the % of Price Targets in excess of Sigma’s 95%tile to the upside.

Forward price returns align fairly linearly with the % of price targtets in excess of Sigma_95U, with a correlation of 0.44. Corresponding correlations for the other price probability %tiles to both the rolling 21d and 63d forward returns, were lower, but some were also respectable. All are given in the table below:

Given the tight link of Sigma forward price probability levels to spot ticker price and the stickiness of analyst price levels it occurred to us that some part of the 0.44 correlation was simply mean reversion / dip buyer’s alpha that has prevailed in the broader market since October 2022.

To test this hypothesis we asked OpenBB for rolling price/ 50d moving average of closing price to incorporate into the analysis as a “mean-reversion” factor.

We likewise checked the output for this ratio, and it was correct (despite our typos in the prompt).

Then, we regressed the NVDA 21d and 63d forward returns on this rolling ratio of NVDA’s price to its 50d moving average, and reran the correlation analysis on the residuals of that regression. The correlation of the Target vs Quant Percentile exceedance rates were a weaker, particularly for the Sigma based metrics, but some were still strong enough to be noteworthy:

So, there is some modest 3m forward predictive power in the ratio of analyst price target exceedances of NVDA price probability percentiles, even after at least partially adjusting for mean reversion. See below for what the current level of the % of Price Targets in excess of 63d Price Probability %tiles suggests the expected returns are, with a broad confidence interval around each, to the upside and downside:

Conclusion

Like the 0.8 average Z-score of analyst target exceedances of upside quant price probability percentiles. the expected value for NVDA’s forward 3 month return of +8.7% is not at all indicative of an AI “bubble”⁴. That said, given that NVDA’s average 3m forward rolling return for the period studied, (which dates back to June 2023) has been 20.5%⁵, the AI trade for the next 3 months, as represented by NVDA⁶, will likely underwhelm.

1. In August, we used OpenBB to evaluate the distribution of analyst price targets for AAPL relative to the 3 month (63d) forward price probability percentiles generated by VecViz’s Vector Model and the bell curve based “Sigma” model. ↩︎
2. Excel is still a thing, right? ↩︎
3. As an aside, note that NVDA’s analyst price target exceed the Vector Model 95% tile 5.6% of the time on average and the Vector Model 99%tile 0.99% of the time. It is trained to reflect actual price movement, but it also reflects analyst opinion! For reference Sigma’s corresponding percentiles ar exceeded 21.5% and 9% of the time by analyst price targets on average, respectively. ↩︎
4. Relative to the prior 2.5 year study period ↩︎
5. The arithmetic average rolling 63d (i.e. 3 month) price return has been ~20% during this time frame. The compounded 63d (i.e., 3 month) growth rate for NVDA during the same time frame has been ~11%. ↩︎
6. The core of the 2025 AI trade, which in our subjective opinion relates to the datacenter build-out and the chips that go in the datacenter. ↩︎