Oil and gas stocks dominated market moves in March, yet only a subset of investors fully realized the upside. Those who did were following a systematic, institutional-grade stock selection process and extended their focus beyond the large integrated majors.
As energy prices climbed amid heightened tensions in the Middle East, the AI-driven strategy did not simply funnel capital into household names. Instead, it rotated into oil infrastructure and midstream companies that the models ranked higher on fundamentals and capital efficiency metrics - stocks that, in several cases, offered more upside than large integrated producers such as Chevron (NYSE:CVX) and Exxon Mobil (NYSE:XOM).
For less than $9 a month, premium members captured a number of these moves. Among the results cited by the strategy were a +46.64%% surge in Par Pacific Holdings Inc (NYSE: PARR), a +22.28% gain in Delek US Energy (NYSE:DK) and a +21.54% advance in Occidental Petroleum (NYSE:OXY).
Those examples are part of a broader set of March winners the AI models had highlighted prior to the escalation in the Middle East. The model's last rebalance - performed before the geopolitical shock - flagged several additional stocks that delivered substantial monthly returns:
- PBF Energy Inc (NYSE: PBF): +41.94% in March alone.
- Profrac Holding Corp (NASDAQ: ACDC): +36.16% in March alone.
- HF Sinclair Corp (NYSE: DINO): +28.18% in March alone.
- Marathon Petroleum Corp (NYSE: MPC): +25.91% in March alone.
The process behind these calls involved running more than 150 institutional-grade financial models across the investable universe. The machine learning ensemble sorted companies by medium-term upside potential and assigned those conclusions to a monthly strategy that equal-weights up to 20 selections.
According to the reporting on model performance, the groundwork laid by the AI meant that once macro conditions aligned with the models' expectations, the results followed. The AI's track record since its November 2023 launch includes a series of outperformance figures presented by the strategy team. Since launch, the collective list of AI picks returned +171.91%, representing a +116.95% outperformance versus the S&P 500 for the same period. For 2026 so far, the model's picks are reported at +9,04%, a +13.14% outperformance relative to the S&P 500 in the same timeframe.
Energy was not the only sector that produced notable winners this year. The strategy's earlier monthly selections captured large moves during the broad risk-on conditions in January and February. Examples provided include:
- Ultra Clean Holdings Inc (NASDAQ:UCTT): +139.56% in the first two months of the year - noted as later removed from the AI's list of picks.
- InnovAge Holding Corp (NASDAQ:INNV): +63%+ in February alone - noted as later removed from the AI's list of picks.
- Generac Holdings Inc (NYSE:GNRC): +65.26% in the first two months of the year - noted as later removed from the AI's list of picks.
The strategy emphasizes that some winners are dropped after they no longer meet the models' risk-reward thresholds. The monthly refresh retains, adds, or removes names based on how the ensemble reassesses medium-term prospects across more than 15 years of global financial data.
Looking ahead, the strategy team is preparing a new batch of AI-selected names for April. That list is scheduled for release on April 1 and will be distributed exclusively to premium members. The upcoming selections are described as a curated set of high-quality stocks that the models currently classify as oversold yet positioned for a meaningful recovery in April.
While conventional sentiment after March's energy rally still tilts toward the sector, the model's framework is also scanning areas that experienced sharp selloffs without a clear fundamental justification. Semiconductors, parts of technology, and consumer discretionary stocks have attracted attention from the model because these sectors often generate rapid reversals when downside has been priced in and momentum reasserts itself.
The critical criteria the AI applies include evidence that margins remain intact and revenue streams have stayed resilient despite recent share-price weakness. Subscribers receive the shortlist of up to 20 names each month rather than an exhaustive universe scan, allowing them to evaluate which selections best fit their risk profile and portfolio.
For existing premium members, the April slate will be distributed through the usual member channels. App users and web users are given subscription paths to access the full monthly list of picks.
How the models operate in practice is consistent across monthly cycles. Each strategy is refreshed at the start of the month, with up to 20 names drawn from a ranking process that combines more than 150 financial models. The ensemble is trained on over 15 years of data and uses equal weighting as a standard tracking benchmark. Portfolio managers or individual subscribers may deviate from equal weighting in implementation, but the approach provides a consistent yardstick for the strategy's identification of opportunities.
Stock selection remains probabilistic, and the framework's risk management is apparent in its willingness to remove holdings once they no longer meet the models' thresholds. Since the official model launch, that approach has coincided with several notable success stories.
Summary
The AI-driven stock selection process generated multiple double-digit winners in March, predominantly in energy and oil infrastructure names identified before the geopolitical shock. The model refreshes monthly, delivering up to 20 equal-weighted picks to premium subscribers; a new list for April will be released on April 1. Beyond energy, the models are monitoring semiconductors, technology and consumer discretionary sectors for oversold yet fundamentally resilient opportunities.
Key points
- Energy and oil-infrastructure stocks produced the largest March gains identified by the AI, including specific winners such as PARR, PBF and ACDC.
- The model ensemble runs over 150 financial models and uses more than 15 years of global data to rank medium-term upside potential, refreshing selections monthly with up to 20 equal-weighted names.
- Sectors under renewed scrutiny include semiconductors, technology and consumer discretionary, which the model views as potential asymmetric opportunities when downside has been priced in.
Risks and uncertainties
- Model performance depends on macro conditions aligning with the models' assumptions - if macro forces diverge, the expected outcomes may not materialize. This affects all sectors the model targets, notably energy and cyclical industries.
- Stocks flagged as winners can be removed from the list if they no longer meet the model's thresholds, introducing turnover risk for a strategy that equal-weights positions. This is relevant for subscribers implementing the recommended names.
- Concentration in sectors that experienced recent stress - for example, semiconductors and consumer discretionary - carries volatility risk if those sectors fail to produce the anticipated rapid reversals.