These dividend-paying data center power plays are on Goldman’s latest top 10 list

The real incremental winner is not the “story stock” itself but the tollbooth on the enabling capex: Baker Hughes should monetize turbine/generator demand with better backlog visibility and a higher-quality service mix, while Kodiak gets the growth narrative but also takes on more execution and balance-sheet sensitivity as it moves from pure compression into quasi-power development. That makes KGS more levered to project conversion than to headline demand, so the stock can keep working only if these orders turn into repeatable contract wins, not one-offs.

The second-order effect is a partial reallocation of power demand away from regulated utility capex and into private on-site infrastructure. That is constructive for midstream names with gas optionality, but it can be a headwind for utilities and merchant power names if investors conclude data-center load will be served off-grid rather than through rate base expansion. In the next 1-3 months, the key market variable is whether these deals show up in backlog and forward guidance; over 6-18 months, the thesis depends on whether hyperscalers scale this model broadly or treat it as a stopgap.

The contrarian risk is that the market is paying today for a 2030 EBITDA path that may require many more project wins than have been validated. KGS has already rerated materially, so the easy multiple expansion may be behind it unless management shows accelerating contracted returns; BKR likely has better asymmetry because equipment demand can compound without the same project-concentration risk. COP and MPC look more like passive beneficiaries of energy-sector rotation than direct data-center winners, so I would not chase them on this theme alone.