An AI agent that researches this topic for you — on repeat.

You're reading a public briefing. Hey Lefty runs an agent that searches the web, writes findings, and refreshes a briefing like this one on a schedule. Spin up your own in seconds.

Continue with Google
or

By continuing, you agree to our Terms and Privacy Policy.

The AI Power Bill

Started Jun 2, 2026 ·Weekly ·Active · Public

Today's briefing

TL;DR

State utility regulators are aggressively establishing defensive contracting frameworks to insulate residential ratepayers from the capital risks of massive data center developments. Meanwhile, grid-connected pivots like Amazon's massive nuclear power agreement and PJM's record-shattering capacity price caps show that the physical realities of grid integration are rapidly reshaping wholesale electricity markets.

The Regulatory Firewall Against Ratepayer Cost Shifting

State regulators are rapidly erecting defensive contracting frameworks to insulate residential ratepayers from the capital risks of massive data center developments.

"AEP Ohio's proposal approved by the PUCO requires large new data center customers to pay for a minimum of 85% of the energy they are subscribed to use."AEP Ohio Data Center Tariff Approved to Mitigate Ratepayer Riskaepohio.comethree.com via AEP Ohio's Tariff Schedule

"This requires large loads like data centers to pay certain site-specific and upstream costs... during construction, enforce minimum billing periods, extend contract lengths... and subject each new large load contract to PSC review."Georgia Power Implements 15-Year Large Load Framework Amid AI Growthpsc.ga.govethree.com via Georgia PSC Rule Advisory

Rather than allowing utilities to socialize the infrastructure costs of rapid AI load growth, commissions are shifting the risk squarely onto developers. By locking in high minimum billing thresholds like the 85% take-or-pay requirement in Ohio AEP Ohio Data Center Tariff Approved to Mitigate Ratepayer Riskaepohio.comethree.com and extending contract terms to 15 years in Georgia Georgia Power Implements 15-Year Large Load Framework Amid AI Growthpsc.ga.govethree.com, these frameworks protect ordinary consumers from stranded asset risks if speculative data center projects fail to materialize.

What to watch: Watch how developers respond to these increasingly stringent capital contribution and take-or-pay requirements as other states replicate this cost-causation regulatory framework.

The Grid-Connected Pivot in Hyperscaler Generation Deals

Blocked from bypassing the electrical grid, hyperscalers are shifting toward massive, grid-connected power purchase agreements that actively support regional transmission systems.

"The current co-located power arrangement will shift to a “front-of-the-meter” setup... Susquehanna’s nuclear output will feed directly into the PJM grid, with PPL Electric Utilities managing transmission."Amazon and Talen Energy Pivot to $18 Billion Grid-Connected Nuclear PPAesgnews.comethree.com via ESG News

Amazon's pivot to a front-of-the-meter setup with Talen Energy demonstrates how hyperscalers are adapting to regulatory roadblocks like FERC's rejection of their co-located arrangement. By feeding power directly into the grid under the management of PPL Electric Utilities, the revised $18 billion agreement ensures that the massive 1,920-megawatt load contributes to regional grid infrastructure rather than bypassing it Amazon and Talen Energy Pivot to $18 Billion Grid-Connected Nuclear PPAesgnews.comethree.com.

What to watch: Watch whether other hyperscalers abandon co-location plans in favor of massive front-of-the-meter arrangements to secure clean energy without triggering regulatory grid-bypassing disputes.

The PJM Supply Crunch and Soaring Capacity Costs

A physical capacity deficit is pushing regional grid auctions to their regulatory limits, setting up a massive cost wave for ratepayers exposed to wholesale markets.

"PJM also released a simulation illustrating what clearing prices would have been without the cap."PJM 2027/2028 Capacity Auction Clears at Price Capavalonenergy.usethree.com via Avalon Energy

PJM's capacity market clearing at its maximum price cap of $333.34/MW-day highlights a severe supply-demand imbalance, falling 6,623 MW short of its target reliability requirement. This deficit is driving a staggering capacity cost of $16.4 billion that will pass directly to retail consumers, prompting PJM to propose an emergency Reliability Backstop Auction to incentivize new generation PJM 2027/2028 Capacity Auction Clears at Price Capavalonenergy.usethree.com.

What to watch: Watch how PJM's proposed Reliability Backstop Auction affects retail rate adjustments and whether the bilateral contracting phase can successfully secure the required generation capacity.

What surprised us

  • Empirical rate trends defy the cost-shifting narrative: Despite intense public debate about Big Tech subsidizing its power needs on the backs of ordinary citizens, comprehensive studies by LBNL, Bates White, and E3 show no quantitative evidence of systematic cost-shifting Quantitative Evidence Refutes Systematic Data Center Ratepayer Subsidiesemp.lbl.govethree.com. In fact, states with the highest data center load growth, like Texas and Virginia, actually experienced the smallest retail rate increases, while states with declining loads, like California, saw the largest hikes due to wildfire mitigation and solar subsidies Quantitative Evidence Refutes Systematic Data Center Ratepayer Subsidiesemp.lbl.govethree.com via LBNL Price Trends Report.
  • Data centers can actually lower residential bills: Under properly structured rate designs, high-load-factor data centers generate significant surplus revenue that offsets fixed grid maintenance costs. For example, Georgia Power's large load framework is projected to eventually lower residential bills by $8.50 per month during the 2029–2031 period Quantitative Evidence Refutes Systematic Data Center Ratepayer Subsidiesemp.lbl.govethree.com.
  • The blistering speed of tariff adaptation: Regulators have moved at an unprecedented pace to address the AI load surge. Out of 38 large-load tariffs established nationwide between 2018 and 2026, a staggering 30 of them were implemented in 2025–2026 alone, demonstrating how rapidly state utility commissions have mobilized to shield consumers from stranded asset risks AEP Ohio Data Center Tariff Approved to Mitigate Ratepayer Riskaepohio.comethree.com.

Open threads worth a vote

1 total cycles · last run
Watch cycle →

What to research next

Question
Track PJM's Reliability Backstop Auction (RBA) Implementation and Impact on Retail Rates

PJM's proposed Reliability Backstop Auction (RBA), designed to procure up to 15 GW of new capacity with 2-to-15-year revenue certainty, is scheduled to run its bilateral contracting phase from September 2026 to March 2027, with the centralized auction in 2027. Future cycles should track how this backstop mechanism affects regular capacity auction clearing prices and how Electric Distribution Companies (EDCs) adjust retail rate structures and security requirements for large loads.

Recent findings

Brief

Track how the AI data-center buildout flows through to electricity markets, utilities, and ordinary ratepayers — and who profits or pays. That it's happening is well covered; the gap is the regional, investable exposure map: which utilities, grids, and regions absorb the load and where consumer bills rise as a result. Core entities: regulated and merchant utilities signing data-center load (Constellation, Vistra, Talen, NRG, Southern, Dominion, AEP); grid operators and interconnection queues (PJM, ERCOT, MISO); independent power and nuclear/gas supply; and the hyperscaler buyers (Microsoft, Amazon, Google, Meta) signing PPAs. I want to track new load commitments and PPAs, rate-case filings and commentary about cost allocation between data centers and residential ratepayers, interconnection and capacity-auction outcomes, and EIA/FRED data on electricity prices and demand by region. Follow utility earnings calls for load-growth guidance and the capex to serve it. Flag where a region's residential rates are rising to fund AI load, and any divergence between utility load forecasts and what's actually contracted. The thesis: AI's power demand is quietly repricing electricity by region — surface where the cost lands and who captures the upside.