Why the Majors Want the Electron Business

For a century the business model of an oil-and-gas major was to find hydrocarbons, produce them, move them, and sell them as molecules — crude oil, refined products, and natural gas. Selling electricity was somebody else's business, the province of utilities and independent power producers. The behind-the-fence data-center boom is dissolving that boundary, and to understand why, you have to see the situation from the major's side of the table.

Begin with demand. The energy transition has cast a long shadow over the long-term demand outlook for oil, and the majors are actively hunting for durable new outlets for natural gas — a fuel they hold in abundance and which burns cleaner than coal or oil. AI data centers represent exactly that: a fast-growing, credit-worthy, long-duration source of gas demand that is, crucially, indifferent to the retail energy politics that complicate other markets. A hyperscaler signing a decade-long agreement for on-site gas-fired power is, from the major's perspective, a premium long-term gas customer who also happens to want the power delivered as electrons. If you already own the gas, moving one step downstream to sell the electricity captures more of the value chain from the same molecule.

Now consider capability, because this is where the majors' entry is more than opportunistic. Behind-the-fence gas generation is, at bottom, a gas logistics and large-project execution challenge. You need reliable gas supply, which may mean pipeline access, gas processing, and in some cases building the pipeline lateral to reach a remote site. You need to manage fuel price risk across years. You need to execute a large, capital-intensive construction project on a tight schedule and operate complex equipment for decades. These are precisely the muscles an integrated major has built over a century: securing supply, managing commodity risk, and delivering multi-billion-dollar projects. A frontier-model lab does not know how to build a pipeline lateral or hedge a gas position; Exxon does in its sleep. That capability asymmetry is the majors' entry ticket.

ExxonMobil has publicly described ambitions to power data centers directly, pairing its natural gas with generation and pointing to carbon capture as the mechanism to manage the emissions. The strategic pitch is integrated: our gas, our generation, our capture technology, delivered as firm power to your campus, off the grid and on your timeline. Chevron has moved in a parallel direction, outlining plans to build gas-fired generation aimed at data-center customers, again leaning on its gas resources and project-execution capability. The specific structures differ and continue to evolve, but the shared thesis is unmistakable: the majors see behind-the-fence power as a large, gas-intensive market where their core competencies are decisive advantages rather than liabilities.

Faraday's read is that this entry is the clearest possible signal of the market's maturity. Stopgaps do not attract Exxon. When the largest, most capital-disciplined energy companies on earth reorganize strategy around selling power to data centers, they are betting that the demand is structural and durable — that AI's appetite for firm, self-supplied power is not a two-year spike but a decade-scale build. Their involvement also changes the competitive landscape for everyone else. A hyperscaler weighing whether to build its own generation team now has the option to instead contract with a major that will supply gas and power as an integrated service, the way it might contract for any other utility. That could accelerate the whole behind-the-fence movement, because it lowers the operational barrier: you no longer have to become a power company yourself if a power company will come to you.

There is a second-order consequence worth naming. If the majors succeed in becoming the power suppliers to AI, they bind the future of gas demand to the future of compute demand — two of the largest capital stories of the decade fused together. That fusion cuts both ways. It gives gas a growth narrative in an era when its long-term demand was in question, and it gives AI a supply of firm power that sidesteps the grid. But it also means the emissions question, which we take up in the next lesson and the next module, moves to the center of the story, because tying compute to gas ties compute to carbon — and that is a liability the majors are proposing to manage rather than avoid. Whether they can, and at what cost, is the question on which much of this market's license to operate will turn.