December 10, 2025
With year end approaching and debate as to whether there is an “AI bubble” all over financial media, we wanted to update the idea generation analysis we did using OpenBB’s Workspace back in August using 8/26/25 closing price based analytics (see youtube vid embedded below), and see if it generates any Mag7 or other core AI names (ex: the semi’s). Back in August 2025 Micron kept “bubbling up” as attractive as we quickly reviewed a host of scatter charts, and it has doubled since then.
First, we merge VecViz analytics with P/E and simple Momentum features on OpenBB.
Following the steps we used on August 26th, we bring the “VV Data Table (All Tickers)” widget (filtered to the 21d Horizon) and the OpenBB provided “Watchlist” widget1 into our workspace, and make them available to the OpenBB AI co-pilot. Both widgets are loaded with data reflecting the 10/9/2025 close.
Next we prompt the co-pilot to merge the tables and create a simple “Momentum” feature.
The charts we review (or skip ahead to filtered results)
After the co-pilot generated the table we replicated it and used the table settings to convert it into a series of scatter charts displaying the relationships between the features, to better understand the distribution and identify good and bad outliers.
We start with the simple Momentum feature on the x-axis and VecViz’s EUB (Expected Up Body)2 and EDB (Expected Down Body)3 features on the Y axis:
Then we look at the V-Score4 vs. the simple Momentum metric:
Then we look at EUB and EDB again, this time relative to the P/E Ratio:
We then zoom in on these EUB and EDB vs P/E charts to exclude extreme P/E outliers:
We then display the V-Score vs the P/E Ratio, again excluding the P/E upside outliers.
The filtered results:
You can view the video from August to see and hear the loose criteria I use when I looking at each scatter chart. Unfortunately there was no statistical study supporting the criteria used. I was just relying on experience with these metrics, trying to find tickers that seemed reasonable.
For the sake of brevity, rather than discuss each chart and how they relate to each other, I converted one of the scatter charts back to a table and and then filtered it via OpenBB’s column level table menus using the criteria listed below. The intent of the criteria is to be loose enough to avoid excluding good opportunities but tight enough to narrow the field materially.
- Momentum > -5
- P/E Ratio between 6 and 30
- EUB Ret > 4
- EDB Ret > -6
- V-Score >= 0
All the Mag7 tickers were among the 129 tickers included in this analysis. Of the 19 tickers that met the filter criteria, listed above, only one is a “Mag7”: META, which just made it under the cut on the 30 p/e cap. Furthermore, there were no semiconductor companies on the list, despite 8 (excluding Mag7 member NVDA) being among the 129. The list is fairly heavy in pharma and finance, along with metals, REITS, and some consumer / housing exposure.
Conclusion on the AI Bubble issue (from this perspective):
We had ~15 “core” AI names in our filter between the Mag7 and semi’s and one was chosen. Of the 114 other tickers considered ,18 were chosen. So, proportionately the filter favored tickers away from the core-AI neighborhood (18/114 > 1/15)5.
So, if there is an AI bubble, it is not so extreme that it precluded us from capturing any core-AI ticker in our screen. That said, this analysis indicates opportunity is primarily away from those core-AI tickers, that have led the market for so long.
The August 2025 video referenced above:
- Note that we expanded the tickers included in the watchlist to 129 (vs. the 45 we used in the 8/26/25 based analysis). ↩︎
- Expected Up Body (EUB) is the Vector Model probability weighted expected value between the closing price and the 95th price probability percentile to the upside. It tells you what price to expect if prices rise at a “upside base case” type of pace. ↩︎
- Expected Down Body (EDB) is the Vector Model probability weighted expected value between the closing price and the 95th price probability percentile to the downside. It tells you what price to expect if prices decline at a “downside base case” type of pace. ↩︎
- The V-Score is a ranking of forward price performance that is based on Vector Model inputs and outputs and calculated via a machine learning ensemble. ↩︎
- If the cap of 30 on P/E was relaxed to 100 P/E and all other criteria remained the same the only core-AI ticker that would have been added is MSFT (P/E of 34). If, in addition to relaxing the P/E criteria the EDB > -6% criteria was removed then core AI tickers MU, GOOGL, and AMAT would join MSFT in meeting the filter criteria. ↩︎