Abandoned Alpha: Winning Portfolios Own Oil Products and Natural Gas

Hydrocarbons are discarded winning stakes not stranded assets

October 5, 2025

Executive summary

• Contrary to a common belief, petroleum and natural gas resources (“hydrocarbons”) are not stranded assets that investors can ignore. They have become abandoned alpha and are highly likely to be one of the most reliable sources for portfolio outperformance and human prosperity through the end of this decade and probably beyond (2025 Commodity Outlook: Abandoned Alpha, 8-Jan-2025).
  
• In 2025 to date, the consensus-busting outperformance of U.S. independent oil refiners’ price returns relative to the S&P 500 already validates our contrarian assessment from nine months ago. An equal-weighted basket of U.S. independent oil refiners has returned more than +50% YTD versus a +14% YTD return for the S&P 500.
  
• Moreover, the refiners’ outperformance has been consistent and broad: VLO has beaten the S&P 500 YTD on 91% of trading days in 2025, while MPC has beaten with 96% frequency. The members of a U.S. refiners’ basket have beaten the S&P 500 YTD price return with 95% frequency since May 16 and 99% since Aug 21. Individual refiners within the U.S. basket are beating 86% to 99% of the individual names within the S&P 500 YTD.
  
• Likewise, the price for an equal-weighted basket of U.S. independent natural gas producers surged far beyond market averages in 1H2025. By June 20 YTD, the gas basket had returned +24.7% versus +1.5% for the S&P 500.
  
• Following a sharp but brief midsummer downdraft driven by well understood fuel competition economics, the gas producers’ basket has advanced by +13.5% since Aug 19 to outpace the S&P 500’s +4.7% gain. Since Sep 22, the splits are even stronger: +8.5% for the gas basket and +0.3% for the S&P 500. Strengthening commodity economics underwrite this value creation, which we are happy to discuss in full detail in one-on-one meetings. The NYM 12-month (12M) natural gas strip has advanced by +14% from $3.45 (Aug 25) to $3.95 (Oct 3) and the 15M, 18M, and 24M strips are now all above $4.00 and rising.
  
• Meanwhile, hyperscalers and a wide array of other technology firms are collectively calling on U.S. gas-fired power plants and other U.S. electrical production capacity at sustained load growth rates not witnessed since the 1950s but likely to sustain into 2030.
  
• Abandoned Alpha means winning stakes have been discarded. Winning investors will pick them up. For investors who care about beating benchmarks for their clients while underwriting a prosperous, ecologically-sound future for human society, these stakes are available for extraordinarily cheap cost relative to own value and competing securities’ inflated valuations. The setup strongly implies a coming, rational, and large rotation to sell the expensively priced common stocks propelling index CAPE ratios to 40x and to buy cheaply valued common stocks that are essential but unloved.

Commentary

Hydrocarbons are essential feedstocks into every industry, and every investment theme geared toward inventing an abundant future for human civilization is entirely dependent on their responsible use.

Yet, the past decade promulgated dogmas that encouraged senseless jawboning over false choices, such as between supposed “clean” and “dirty” energy. Investors can no longer afford the distraction of manufactured fights over deliberately pejorative slurs.

Without access to hydrocarbons—as Silicon Valley is now increasingly, if belatedly, proving by its actions—there are no electric vehicles, no tertiary care hospitals, no data centers, and no artificial intelligence (AI) workbots. All planned solar PV arrays, wind turbines, and battery storage projects dissolve in a poof of irony. Every hyperscaler, IPP, and AI-equipment unicorn dies on the vine for lack of hydrocarbon nourishment and due respect for the periodic table of elements.

These facts of physical chemistry (“science”) conflict with investors’ indifference toward, if not outright divestment from, oil and gas assets. This conflict is not sustainable. The public and private securities of essential upstream producers, midstream distributors, and downstream refiners and processors of petroleum and natural gas commodities have become abandoned alpha.

Nine months ago, we made these observations the centerpiece of our 2025 Commodity Outlook (Abandoned Alpha, 8-Jan-2025) because we expected this truth would finally come home to roost in this year. Coming into January 2025, we assessed the truth-telling process would begin with consensus-busting advances in the shares of U.S. independent oil refiners and U.S. independent natural gas producers, as local demand started to surpass crucial thresholds for local supply. We continue to expect these local economics will evolve next into a wider and more global process that extends at least into 2028 and most likely beyond.

Moreover, this oil and gas theme, Abandoned Alpha, is part of a larger nexus within AI compute and innovations in nuclear power that we believe altogether is helping spur the most consequential transformation of the world political economy in the past 80 years. By some measures, ever. This long-cycle “storm” has been anticipated by George Friedman, William Strauss, Neil Howe, us, and others. But confused observers continue to misinterpret the present and coming transformations through a backward-focused lens that has scant hope of profitably navigating the future (Cyclical Transformations, Not Recession: Anyone? Anyone?, 31-Jul-2025).

With respect to the narrow question of oil refining and upstream gas, the market evidence through 2025 to date validates the assessment we published nine months ago. Investors who owned oil product spreads and natural gas strips were positioned correctly and have beaten their benchmarks by huge margins. Bearish investors positioned short were wrong and have lost.

Specifically, an equal-weighted basket of refiners (VLO, MPC, PSX, CVI, DK, DINO) has returned +50.8% YTD, with each member company beating the S&P 500’s YTD price return (+14.2%) and the basket beating the benchmark by 3,660 bp. Moreover, outperformance has been consistent: VLO has beaten the S&P 500 YTD on 91% of trading days in 2025, while MPC has beaten with 96% frequency. The basket’s frequency of trading-day beats against the benchmark’s YTD price return is 95% since May 16 and 99% since Aug 21. Individual refiners within the basket are beating 86% to 99% of the individual names within the S&P 500 YTD.

Likewise, the price for an equal-weighted basket of gas producers (EQT, EXE, RRC, AR) surged far beyond market averages in 1H2025. By June 20 YTD, the gas basket had returned +24.7% versus +1.5% for the S&P 500. The underlying power-demand economics were so strong they spurred summertime price-driven fuel switching from gas toward coal for electricity generation. This seasonal dynamic drove a drawdown in gas shares (sector-typical, if larger-than-normal), as the Trump Administration facilitated the fuel switching with retirement deferrals on the J.H. Campbell 1.42 GW coal plant in Michigan and other coal assets slated to be closed in 2025. The drawdown was sharp but complete by Aug 19, with the gas basket then returning –5.1% YTD against a +9.0% YTD return for the S&P 500. Since Aug 19, the gas basket has advanced by +13.5% to outpace the S&P 500’s +4.7% gain. Since Sep 22, the splits are even stronger: +8.5% for gas basket and +0.3% for S&P 500. The NYM 12-month (12M) natural gas strip has advanced by +14% from $3.45 (Aug 25) to $3.95 (Oct 3) and the 15M, 18M, and 24M strips are now all above $4.00 and rising.

Behavioral evidence from the tech community also supports the assessment we made nine months ago and continue to hold today. It is now clear that 2025 will go down in the history books as the year Silicon Valley’s visionary mechanical engineers and computer scientists finally saw the U.S. electricity grid as it is rather than as “dirty energy” aspersions dream it to be. We assess the scientists in the software tech community now see the challenges and opportunities already long understood by the physical scientists in the energy community. Now standing on common ground as partners, they are together constructing the logical and necessary engineering solutions through natural gas, as well as nuclear and other energy solutions, to supply 99.999% reliable electricity to data centers for AI applications that will, hopefully, turbocharge human knowledge and universal progress.

As strong AI-driven power demand growth calls on energy feedstocks both on-grid and behind the meter, the hyperscalers are also responsibly committed to preserve America’s high standard of electrical reliability (99.97%) for residential and commercial consumers in the hottest summers and coldest winters, even if it means order of precedence will rank order the security of a vulnerable elderly person freezing in South Bend over a hedge fund analyst drinking a mochaccino and calling on Gemini in New York City. This is good. The combination of social responsibility and world-leading innovation is yet another example of why America wins against all world competition.

Here are just five examples at the hyperscale level: (1) April 2025: xAI extends installation to 35 onsite gas turbines at its Memphis, Tennessee Colossus cluster to power >100,000 GPUs, (2) August 2025: Meta’s $10 billion hyperscale campus development in Louisiana will rely on dedicated power from three new gas-fired power plants that Entergy will build by 2028 at a cost of $3 billion. (3) June 2025: Microsoft announces its 20-year purchase power agreement (PPA) with Constellation Energy will include restart of Unit 1 at the Three Mile Island nuclear power plant in Pennsylvania for dedicated power to multiple data centers, (4) June 2025: Crusoe Energy orders 19 GE Vernova LM2500XPRESS gas turbines to serve OpenAI’s Stargate project in Abilene, Texas; and (5) February 2025: EdgeConneX holds first public hearings on its plan (now moving forward after July 2025 filings) to develop two data centers in New Albany, Ohio using onsite gas-fired power to support 1.2 million square feet of computing infrastructure.

One crucial aspect to understand about Abandoned Alpha is its recognition of the surprisingly investor-discarded values in oil and gas assets does not mean Abandoned Alpha opposes solar, wind, hydro, geothermal, ocean wave, nuclear fission, nuclear fusion, or any other competitive and sensible solution to humanity’s fuel and non-fuel problems. To the precise contrary, Abandoned Alpha expressly rejects the nonsensical assertion of binary and fatal conflict, including what we expect to be the ultimate game-changing technology for power markets in the 2040s and afterward, nuclear fusion.

More practically today and through 2030, the inevitable rotation away from expensive CAPE 40x-valued common stocks will benefit unloved, too-cheaply-valued common stocks in energy, materials, and industrials. The central task now for any competitive analyst and/or PM, human or AI, is to identify these common stocks, to build a long position in them, and to own them through 2028+.

Nuclear fusion energy, once it is developed at scale, will forever turn over this apple cart, by definition (A Star Is Born). We monitor every fusion experimental result reported publicly by every laboratory in the world, to the best of our ability to track progress. But fusion is not here yet and even once it inevitably arrives, human economies will still need hydrocarbons for use as materials, not least of which in wires and cables.

At that point, a valid research question will be whether humanity shifts over time to source hydrocarbons for materials use from air, water, and/or land (plants), rather than harvesting subterranean basins, in some combination of natural and synthetic processes. That real and coming risk is beyond the scope of present investment realities and therefore is put to the side for now.