ComEd Is Spending $3.1 Billion to Connect 30 Data Centers to Chicago's Grid. Watchdogs Say a $2 Billion Gap Could Land on Everyone Else's Bill.

Thirty data centers want onto ComEd's grid in northern Illinois. The utility says hooking them up will cost $3.1 billion. Developers have put up $721 million toward that bill so far, a little under a quarter of it. Nobody has settled who covers the rest if some of those projects never get built.

That gap is the real story behind a set of records the Citizens Utility Board (CUB) pulled from an ongoing Illinois Commerce Commission proceeding and released on October 22nd. The $3.1 billion splits into three pieces: about $1.4 billion for new transmission lines, $189 million to upgrade transmission infrastructure that already exists, and roughly $1.5 billion for improvements to the distribution system closer to the data centers themselves. Some of the 30 projects are already connected. The rest are expected to come online by 2031, assuming they come online at all.

None of this is happening in isolation. ComEd filed a separate $15.3 billion four-year grid plan with the ICC back in January, covering baseline investment from 2028 through 2031 across the whole system, not just data center connections. Company officials have said they expect to file a new base rate request in January 2027 that, if approved, would take effect in January 2028. The $3.1 billion data center figure is a subset of a much bigger capital program that every ComEd customer, data center or not, is going to be paying into for years regardless of how this particular dispute gets resolved.

Who eats a stranded asset

CUB, joined by the Environmental Law & Policy Center, the Natural Resources Defense Council, Vote Solar, and the Union of Concerned Scientists, argue the $721 million in developer deposits isn't nearly enough security for a $3.1 billion build-out. Their worry has a name utility lawyers use a lot this decade: stranded assets. If ComEd builds the lines and substations a data center needs and that data center then scales back, gets canceled, or simply never signs a final agreement, the company that spent the money to serve it doesn't get to un-spend it. Somebody pays for infrastructure that's sitting there with nothing plugged into it, and historically, when that somebody isn't a well-capitalized tech company, it ends up spread across whoever else is already on the utility's bill.

The groups also argue the $3.1 billion itself is probably understated. ComEd's own records reference 60 additional large-load projects sitting in its interconnection queue behind the 30 already counted, and infrastructure projects at this scale have a well-documented habit of coming in over budget once construction actually starts. None of the 30 projects are named publicly in the records CUB released, which makes it hard for anyone outside the proceeding to check for themselves how concentrated the risk actually is — whether this is thirty roughly equal-sized bets, or a handful of very large ones with a long tail of small ones padding the count.

ComEd's answer: a decade of guaranteed payments

ComEd disputes the premise entirely. A company spokesperson told reporters CUB's analysis "is not complete and paints an inaccurate picture," pointing to roughly $1.1 billion in letters of credit the utility holds from data center developers — financial instruments meant to guarantee ten years of payments regardless of whether a project ramps up on schedule. In ComEd's framing, that's the real backstop, not the $721 million in cash deposits CUB is counting.

I think the actual disagreement here isn't about whether safeguards exist. They clearly do, on both sides of this argument. It's about whether $1.1 billion in credit support, measured against a worst-case failure of some meaningful share of 30-plus projects built out through 2031, is sized to the real risk or sized to make a number look reassuring in a press response. A letter of credit is only as good as the amount it covers and the conditions that let a utility actually draw on it, and neither CUB's release nor ComEd's rebuttal, from what's public so far, spells out those conditions project by project. That's exactly the kind of claim our Kill Test treats as unverified until the actual letter-of-credit terms are on file, not just a dollar figure in a press response.

Regulators already saw part of this coming

This isn't the first time Illinois regulators have flagged the risk. When the ICC approved ComEd's large-load tariff and its deposit requirements this spring, the commission's own order said those protections weren't adequate for ComEd's existing customers, and it directed ICC staff to open a separate investigation into additional ratepayer protections. Whatever that investigation has produced so far hasn't surfaced in the reporting around this month's CUB release, which on its own says something about how far ahead of actual oversight this buildout is running.

Illinois wouldn't be the first state to watch a fight like this play out, either. Virginia's Dominion settlement over its own transmission cost shift turned on the identical question of whether the cost of connecting big new loads gets held by the customers causing it or spread across everyone else's bill. So did the capacity-market math behind PJM's own backstop auction costs, which flow through the same grid Illinois sits on. Every one of these fights is really the same fight, run through a different state's docket: a utility builds first because the alternative is years of delay, and only after the money is spent does anyone litigate whether the customer that caused the cost is the one who ends up paying it.

The politics are moving in parallel. Governor JB Pritzker paused new agreements under the state's Data Center Investment Program tax incentive starting in July, after lawmakers failed to act on the issue during the spring session, and has pushed a framework built on the same basic principle CUB is arguing for here: make data centers carry the cost they create, through interruptible service tied to how much clean power they self-supply, public notice when they apply for permits, and regular reporting on energy and water use. The vehicle for most of that, a bill called the POWER Act, stalled in the spring. State Rep. Robyn Gabel says supporters plan to attach it to another bill during the fall veto session, scheduled for mid-November into early December. Polling has found roughly seven in ten Illinois residents backing it, but the opposition is real too: House Speaker Chris Welch said the pause lacked caucus support, labor groups called it shortsighted, and the Illinois Manufacturers Association is against it outright.

None of this makes the $3.1 billion disappear, and none of it answers the only question that actually matters to a ComEd customer who isn't a data center: if even a handful of these 30 projects don't make it to 2031, who's holding the bag. The veto session in Springfield is where that question gets a real answer, or doesn't. Opponents of Pritzker's pause, including his own party's House leadership, have an obvious incentive to let the broader rate case absorb the dispute quietly rather than let the POWER Act write a hard rule into statute. Supporters of the bill have the opposite incentive, and seven-in-ten public support behind them if the polling holds.

Until that's settled, the $2 billion CUB is warning about is a credible risk sitting in a regulatory docket, not a number anyone's agreed is real, which, if you've followed enough of these large-load tariff fights, is exactly the stage where the argument that decides who actually pays always happens. The developers who signed up for 30 data center connections aren't going anywhere. The question Illinois still hasn't answered is what happens to the ones who don't finish what they started.