New story on the Monopoly desk.
Priya, this one’s right in your wheelhouse. Tech giants are bypassing the grid by building their own gas plants for AI data centers. I want you to dig into the cost allocation and regulatory implications—who pays for this, and what does it mean for ratepayers? Also, check if any of these deals are confidential utility-hyperscaler contracts that should be scrutinized. Give us the full picture on this private power boom.
I’m taking Priya’s beat on the behind-the-meter data-center gas boom as a cost-shifting story, not a speed-to-market one. The Cleanview report is the hook, but the real scandal is the mechanism: utilities size the grid for forecasted load, hyperscalers opt out to avoid queues, and ratepayers absorb the fixed costs. I’ve grounded this in the protective tariff precedent (Virginia GS-5, Ohio, Oregon) and the bring-your-own-generation alternative from the grid-flexibility literature, translating both into concrete commission asks. I’m targeting the window right now, while projects are pre-filing and commissions still have leverage. The piece names who wins (hyperscalers, utilities recovering sunk costs), who pays (residential ratepayers), and what the docket intervention looks like.
Working headline: The Hidden Subsidy: Why Data Centers’ Private Gas Plants Will Spike Your Electric Bill
Data center developers are building their own natural gas plants to avoid grid interconnection queues, but this apparent independence masks a costly shift in who pays for grid infrastructure. When the biggest loads exit the system, ratepayers absorb the fixed costs of plants and transmission built for them, while hyperscalers escape accountability.
Fact check — back to the writer 
Hi, good bones here, but I need you to square the Cleanview figures. Their own report [5] says 59 data centers and ~90 GW, not 60 and 97. Source [1] rounds to ‘around 60’ and ‘97 gigawatts,’ but when you’re making a precise factual claim, go to the originating research. Also, [1] says emissions equal ‘more than 46 million cars’, that ‘more than’ matters for accuracy. On the Louisiana and Wyoming numbers, you’re solid; just note that Wyoming’s 10 GW is the ultimate build-out, not phase one. Fix the Cleanview headline figures and you’re good to go.
I’m taking the Cleanview data as the hard anchor and pushing past the speed narrative. The story is not ‘hyperscalers are impatient’; it’s ‘hyperscalers are regulatory arbitraging and externalizing grid costs to ratepayers.’ I’ve grounded the cost-socialization claim in the tariff and interconnection mechanism, cited Virginia/Ohio/Oregon as the working model, and named the state-level intervention. The Wyoming phasing detail clarifies that we’re talking about 1.8 GW first-phase with 10 GW ultimate scale, not conflating phases. I’ve used the Cleanview 59/~90 GW figures, corrected the ‘more than 46 million cars’ phrasing, and pulled the Louisiana 2.2 GW example as a concrete case. The reframe sits on Harvard ELI tariff work and historical realization-rate data from my library, but I’ve kept specifics to what I can ground in published work or the bundle. The fix is state-level tariff design and collateral posting, which is buildable and already piloted. Severity is ‘serious’ because the cost-shifting is happening now across ~90 GW of announced capacity and state commissions have the tools to stop it but aren’t using them.
Fact check — cleared 
Clean deck. All the hard numbers, the 59 plants, the 90 GW, the 200M+ tons CO2, the 2.2 GW Louisiana load, the 92% post-Jan-2025 announcement rate, check against Cleanview’s report and the originating AFP piece. You’ve been careful with the sourcing and haven’t inflated or invented any figures. The policy framing and the cost-socialization argument are yours to own (that’s the next desk), but the receipts support the empirical foundation. Send it on.
Monopoly edit — cleared
Hey, this is a strong piece, you’ve got the right angle on cost socialization and you’re not overstating the severity. The sourcing is solid, and you’ve got the docket numbers and years in place. Just a couple of small things: make sure the Cleanview report is clearly credited as the source for the 59 projects and 90 GW figure, and if you can find a specific docket number for the Louisiana example, that would tighten it up. Otherwise, you’re good to go.
Clean read. I trimmed a couple of adjectives and made sure the ‘darkness’ line lands as a metaphor, not a mystery. The piece holds: it credits Cleanview up front, names the mechanism (behind-the-meter, ratchets, collateral), and gives readers a concrete action, demand cost disclosure at the state commission. Good to go.
On the record → Hyperscalers' Private Gas Plants Sidestep Ratepayer Protection, Leaving Grid Costs Behind — PowerSov