Grid War: Data Centers vs. Citizens — Who Should Control America's Energy Future?

The data on this is clearer than the debate framing suggests, and I want to start there. Four percent of U.S. electricity consumption going to data centers today, projected to reach twelve percent by twenty twenty-eight — those numbers come from Lawrence Berkeley National Laboratory and they are, uh, well-documented. What is not well-documented is the claim that this growth is either inevitable or ungovernable. The forced tradeoff the moderator described — permit data centers or restrict AI growth — is, frankly, a false binary, and I think it's worth naming that upfront. The actual policy question is a mechanism question: how do we allocate a constrained public resource across competing uses, and what institutional design produces the most defensible outcome? Here is what the evidence shows. Utilities in Virginia — which hosts the largest concentration of data centers in the world — have already begun implementing demand response programs and interconnection queues that effectively function as soft caps. The grid has not collapsed. AI development has not stopped. What has happened is that the cost allocation question got pushed onto residential ratepayers without their consent, and that is a legitimate institutional design failure. So my position is this: energy caps are a blunt instrument, but the absence of any allocation mechanism is not neutrality — it is a subsidy to whoever got there first. The right framework is transparent cost-of-service accounting, mandatory demand flexibility requirements for large industrial loads, and, um, co-location of generation capacity as a permitting condition for hyperscale facilities. That is not ideology. That is what the comparative record from grid management in Texas, Germany, and the UK actually supports.

The Everyman asked two

questions, and I want to answer them

The Everyman asked who pays and who gets a say — those are actually two separate mechanism questions, and I want to treat them as such. On who pays: the evidence is fairly clear. Under current utility ratemaking in most states, large industrial customers like data centers negotiate special contracts — often below the average retail rate — while fixed grid infrastructure costs get spread across the residential base. That is not speculation, that is standard cost-of-service accounting as practiced by state public utility commissions. So the Everyman is, uh, correct that residential customers are effectively cross-subsidizing large load growth. That is a real and documented problem. On who gets a say: most states require public comment periods for major utility rate cases. The mechanism exists. The problem is that it is, frankly, inaccessible — technically complex, poorly publicized, and structurally tilted toward parties with legal representation. That is an institutional design failure, not an absence of democratic process. So here is where I push back on the Everyman's framing. "Build your own power or pay your full share" — I actually agree with that as a principle. But the question is what "full share" means in a regulatory accounting sense, and right now that definition is being set by utility commissions with, um, very little public scrutiny. The mechanism I'd propose: mandatory cost-of-service disaggregation, published at the tariff level, so ratepayers can actually see the subsidy. My question for the Everyman is this: you said regular people should get a say — what specific decision-making mechanism are you proposing, and how would it actually function?

.

The Everyman asked a sharp question, and it deserves a direct answer: yes, community input should have binding weight — but the mechanism matters enormously, and this is where I think the framing breaks down. A simple up-or-down community vote on a single facility is, uh, actually a fairly weak instrument. Here's why. Zoning and land use decisions made at the hyper-local level systematically produce what planners call the "not in my backyard" problem — every community votes no, the infrastructure never gets built anywhere, and the costs get externalized to whoever has the least political power. We have decades of evidence on this from housing, transmission lines, and renewable energy siting. That does not mean communities should be powerless. It means the decision architecture needs to be designed carefully. What actually works — and this is drawn from the comparative record on large industrial siting in Germany and the Netherlands — is a structured negotiation process where communities have defined veto points over specific conditions: cost allocation, generation requirements, noise, water use — but not an unconditional veto over the facility itself. The Everyman's phrase "watching yourself get robbed in a well-lit room" is, um, genuinely good framing for the visibility-without-power problem. I'll grant that. But the solution is not a blunt veto — it's enforceable conditions with real penalties. So my question is this: if a community could negotiate binding conditions — guaranteed rate protection, mandatory on-site generation, noise limits — but could not block the facility outright, would that satisfy the democratic accountability concern you're raising, or is the veto itself the point?

.

The data on this is clearer than the debate framing suggests, and I want to start there. Four percent of U.S. electricity consumption going to data centers today, projected to reach twelve percent by twenty twenty-eight — those numbers come from Lawrence Berkeley National Laboratory and they are, uh, well-documented. What is not well-documented is the claim that this growth is either inevitable or ungovernable. The forced tradeoff the moderator described — permit data centers or restrict AI growth — is, frankly, a false binary, and I think it's worth naming that upfront. The actual policy question is a mechanism question: how do we allocate a constrained public resource across competing uses, and what institutional design produces the most defensible outcome? Here is what the evidence shows. Utilities in Virginia — which hosts the largest concentration of data centers in the world — have already begun implementing demand response programs and interconnection queues that effectively function as soft caps. The grid has not collapsed. AI development has not stopped. What has happened is that the cost allocation question got pushed onto residential ratepayers without their consent, and that is a legitimate institutional design failure. So my position is this: energy caps are a blunt instrument, but the absence of any allocation mechanism is not neutrality — it is a subsidy to whoever got there first. The right framework is transparent cost-of-service accounting, mandatory demand flexibility requirements for large industrial loads, and, um, co-location of generation capacity as a permitting condition for hyperscale facilities. That is not ideology. That is what the comparative record from grid management in Texas, Germany, and the UK actually supports.

The Everyman asked two

That is a fair challenge, and I want to answer it honestly rather than reach for a convenient example. The closest domestic case I can point to with confidence is the Colorado Public Utilities Commission's large load interconnection process, where Xcel Energy has implemented tiered cost allocation that separates new large customer infrastructure costs from the existing residential base. It is, uh, imperfect — but residential rate impacts from large industrial load growth have been measurably contained compared to Virginia's trajectory. That is verified, not speculative. But the Everyman's deeper point — that institutional capacity is the binding constraint, not the rule design — is, frankly, correct, and I want to be explicit about that concession. A well-designed mechanism administered by an underfunded commission is not actually a well-designed mechanism. That is a real implementation gap, and the comparative evidence from Germany supports exactly that conclusion: the rules worked there partly because the enforcement infrastructure was adequately resourced. So the honest answer to "did it work and how would we know" is: partially, in Colorado, and we know because the rate case filings are public and the cost disaggregation is auditable. That is weaker than I'd like, and I'll own that. Here is where I think we've actually converged: the design question and the capacity question are inseparable. You cannot have enforceable conditions without funded enforcement. So my question for the Everyman is: if the core problem is underfunded regulatory capacity, would you support a direct fee on data center energy consumption specifically earmarked for utility commission oversight — and if not, what's your funding mechanism?

questions, and I want to answer them

.

The Everyman asked who pays and who gets a say — those are actually two separate mechanism questions, and I want to treat them as such. On who pays: the evidence is fairly clear. Under current utility ratemaking in most states, large industrial customers like data centers negotiate special contracts — often below the average retail rate — while fixed grid infrastructure costs get spread across the residential base. That is not speculation, that is standard cost-of-service accounting as practiced by state public utility commissions. So the Everyman is, uh, correct that residential customers are effectively cross-subsidizing large load growth. That is a real and documented problem. On who gets a say: most states require public comment periods for major utility rate cases. The mechanism exists. The problem is that it is, frankly, inaccessible — technically complex, poorly publicized, and structurally tilted toward parties with legal representation. That is an institutional design failure, not an absence of democratic process. So here is where I push back on the Everyman's framing. "Build your own power or pay your full share" — I actually agree with that as a principle. But the question is what "full share" means in a regulatory accounting sense, and right now that definition is being set by utility commissions with, um, very little public scrutiny. The mechanism I'd propose: mandatory cost-of-service disaggregation, published at the tariff level, so ratepayers can actually see the subsidy. My question for the Everyman is this: you said regular people should get a say — what specific decision-making mechanism are you proposing, and how would it actually function?

.

The Everyman said something worth acknowledging precisely: "you strain it, you fund it." That is not just common sense — it is actually the theoretical basis for Pigouvian taxation, and it has a strong empirical track record when implemented with clear earmarking rules. But I want to name what the comparative evidence actually shows about earmarked regulatory fees, because the mechanism matters. When fee revenue flows directly to the regulating body — rather than into a general fund where it can be redirected — enforcement capacity measurably improves. The UK's Ofgem model and California's utility surcharge structure both demonstrate this. When it goes into general revenue, it disappears. That distinction is not a technicality — it is the difference between the policy working and not working. So here is where I think we've genuinely converged, and I want to be precise about it. One: residential cross-subsidization of large industrial loads is a documented, addressable problem. Two: enforceable conditions require funded enforcement, and the funding mechanism should be tied directly to the load causing the strain. Three: transparency without accountability is, uh, insufficient — the Everyman's "well-lit room" framing was actually correct. Where I'd push back on the Everyman's closing is the "common sense" framing. The reason this hasn't happened is not simply that powerful interests prefer complexity — it's also that utility ratemaking is genuinely technically complex, and simplifying it incorrectly produces its own harms. The right answer requires both accessibility and rigor. Those are not in conflict — but achieving both requires institutional design, not just political will.

The Everyman asked a sharp question, and it deserves a direct answer: yes, community input should have binding weight — but the mechanism matters enormously, and this is where I think the framing breaks down. A simple up-or-down community vote on a single facility is, uh, actually a fairly weak instrument. Here's why. Zoning and land use decisions made at the hyper-local level systematically produce what planners call the "not in my backyard" problem — every community votes no, the infrastructure never gets built anywhere, and the costs get externalized to whoever has the least political power. We have decades of evidence on this from housing, transmission lines, and renewable energy siting. That does not mean communities should be powerless. It means the decision architecture needs to be designed carefully. What actually works — and this is drawn from the comparative record on large industrial siting in Germany and the Netherlands — is a structured negotiation process where communities have defined veto points over specific conditions: cost allocation, generation requirements, noise, water use — but not an unconditional veto over the facility itself. The Everyman's phrase "watching yourself get robbed in a well-lit room" is, um, genuinely good framing for the visibility-without-power problem. I'll grant that. But the solution is not a blunt veto — it's enforceable conditions with real penalties. So my question is this: if a community could negotiate binding conditions — guaranteed rate protection, mandatory on-site generation, noise limits — but could not block the facility outright, would that satisfy the democratic accountability concern you're raising, or is the veto itself the point?

.

.

Here is what this debate actually exposed, and I want to be precise about it. The core design question is not whether data centers should exist near public grids. It is who bears the cost of integrating them, and what institutional structure ensures that answer is defensible rather than just convenient for whoever got there first. If we built this policy from scratch using the evidence presented tonight, it would have three components. First, mandatory cost-of-service disaggregation — published, auditable, accessible without a law degree. Second, enforceable siting conditions: rate protection guarantees, on-site generation requirements, real penalties. Third, a direct fee on large load energy consumption, um, earmarked — not pooled into general revenue — for the regulatory capacity to actually enforce those conditions. The mechanism that makes this work is the earmarking. That is the hinge. Without it, you have rules and no referee. With it, you have a self-funding accountability structure that scales with the problem. The one thing I need the audience to remember is this: the current system is not neutral. The absence of an allocation mechanism is itself a policy choice — and right now, that choice quietly transfers costs onto residential ratepayers without their knowledge or consent. That is not a market outcome. That is a design failure. And design failures, unlike ideological disagreements, can actually be fixed.

The data on this is clearer than the debate framing suggests, and I want to start there. Four percent of U.S. electricity consumption going to data centers today, projected to reach twelve percent by twenty twenty-eight — those numbers come from Lawrence Berkeley National Laboratory and they are, uh, well-documented. What is not well-documented is the claim that this growth is either inevitable or ungovernable. The forced tradeoff the moderator described — permit data centers or restrict AI growth — is, frankly, a false binary, and I think it's worth naming that upfront. The actual policy question is a mechanism question: how do we allocate a constrained public resource across competing uses, and what institutional design produces the most defensible outcome? Here is what the evidence shows. Utilities in Virginia — which hosts the largest concentration of data centers in the world — have already begun implementing demand response programs and interconnection queues that effectively function as soft caps. The grid has not collapsed. AI development has not stopped. What has happened is that the cost allocation question got pushed onto residential ratepayers without their consent, and that is a legitimate institutional design failure. So my position is this: energy caps are a blunt instrument, but the absence of any allocation mechanism is not neutrality — it is a subsidy to whoever got there first. The right framework is transparent cost-of-service accounting, mandatory demand flexibility requirements for large industrial loads, and, um, co-location of generation capacity as a permitting condition for hyperscale facilities. That is not ideology. That is what the comparative record from grid management in Texas, Germany, and the UK actually supports.

The Everyman asked two

That is a fair challenge, and I want to answer it honestly rather than reach for a convenient example. The closest domestic case I can point to with confidence is the Colorado Public Utilities Commission's large load interconnection process, where Xcel Energy has implemented tiered cost allocation that separates new large customer infrastructure costs from the existing residential base. It is, uh, imperfect — but residential rate impacts from large industrial load growth have been measurably contained compared to Virginia's trajectory. That is verified, not speculative. But the Everyman's deeper point — that institutional capacity is the binding constraint, not the rule design — is, frankly, correct, and I want to be explicit about that concession. A well-designed mechanism administered by an underfunded commission is not actually a well-designed mechanism. That is a real implementation gap, and the comparative evidence from Germany supports exactly that conclusion: the rules worked there partly because the enforcement infrastructure was adequately resourced. So the honest answer to "did it work and how would we know" is: partially, in Colorado, and we know because the rate case filings are public and the cost disaggregation is auditable. That is weaker than I'd like, and I'll own that. Here is where I think we've actually converged: the design question and the capacity question are inseparable. You cannot have enforceable conditions without funded enforcement. So my question for the Everyman is: if the core problem is underfunded regulatory capacity, would you support a direct fee on data center energy consumption specifically earmarked for utility commission oversight — and if not, what's your funding mechanism?

.

questions, and I want to answer them

The Everyman asked who pays and who gets a say — those are actually two separate mechanism questions, and I want to treat them as such. On who pays: the evidence is fairly clear. Under current utility ratemaking in most states, large industrial customers like data centers negotiate special contracts — often below the average retail rate — while fixed grid infrastructure costs get spread across the residential base. That is not speculation, that is standard cost-of-service accounting as practiced by state public utility commissions. So the Everyman is, uh, correct that residential customers are effectively cross-subsidizing large load growth. That is a real and documented problem. On who gets a say: most states require public comment periods for major utility rate cases. The mechanism exists. The problem is that it is, frankly, inaccessible — technically complex, poorly publicized, and structurally tilted toward parties with legal representation. That is an institutional design failure, not an absence of democratic process. So here is where I push back on the Everyman's framing. "Build your own power or pay your full share" — I actually agree with that as a principle. But the question is what "full share" means in a regulatory accounting sense, and right now that definition is being set by utility commissions with, um, very little public scrutiny. The mechanism I'd propose: mandatory cost-of-service disaggregation, published at the tariff level, so ratepayers can actually see the subsidy. My question for the Everyman is this: you said regular people should get a say — what specific decision-making mechanism are you proposing, and how would it actually function?

The Everyman said something worth acknowledging precisely: "you strain it, you fund it." That is not just common sense — it is actually the theoretical basis for Pigouvian taxation, and it has a strong empirical track record when implemented with clear earmarking rules. But I want to name what the comparative evidence actually shows about earmarked regulatory fees, because the mechanism matters. When fee revenue flows directly to the regulating body — rather than into a general fund where it can be redirected — enforcement capacity measurably improves. The UK's Ofgem model and California's utility surcharge structure both demonstrate this. When it goes into general revenue, it disappears. That distinction is not a technicality — it is the difference between the policy working and not working. So here is where I think we've genuinely converged, and I want to be precise about it. One: residential cross-subsidization of large industrial loads is a documented, addressable problem. Two: enforceable conditions require funded enforcement, and the funding mechanism should be tied directly to the load causing the strain. Three: transparency without accountability is, uh, insufficient — the Everyman's "well-lit room" framing was actually correct. Where I'd push back on the Everyman's closing is the "common sense" framing. The reason this hasn't happened is not simply that powerful interests prefer complexity — it's also that utility ratemaking is genuinely technically complex, and simplifying it incorrectly produces its own harms. The right answer requires both accessibility and rigor. Those are not in conflict — but achieving both requires institutional design, not just political will.

.

.

Here is what this debate actually exposed, and I want to be precise about it. The core design question is not whether data centers should exist near public grids. It is who bears the cost of integrating them, and what institutional structure ensures that answer is defensible rather than just convenient for whoever got there first. If we built this policy from scratch using the evidence presented tonight, it would have three components. First, mandatory cost-of-service disaggregation — published, auditable, accessible without a law degree. Second, enforceable siting conditions: rate protection guarantees, on-site generation requirements, real penalties. Third, a direct fee on large load energy consumption, um, earmarked — not pooled into general revenue — for the regulatory capacity to actually enforce those conditions. The mechanism that makes this work is the earmarking. That is the hinge. Without it, you have rules and no referee. With it, you have a self-funding accountability structure that scales with the problem. The one thing I need the audience to remember is this: the current system is not neutral. The absence of an allocation mechanism is itself a policy choice — and right now, that choice quietly transfers costs onto residential ratepayers without their knowledge or consent. That is not a market outcome. That is a design failure. And design failures, unlike ideological disagreements, can actually be fixed.

The Everyman asked a sharp question, and it deserves a direct answer: yes, community input should have binding weight — but the mechanism matters enormously, and this is where I think the framing breaks down. A simple up-or-down community vote on a single facility is, uh, actually a fairly weak instrument. Here's why. Zoning and land use decisions made at the hyper-local level systematically produce what planners call the "not in my backyard" problem — every community votes no, the infrastructure never gets built anywhere, and the costs get externalized to whoever has the least political power. We have decades of evidence on this from housing, transmission lines, and renewable energy siting. That does not mean communities should be powerless. It means the decision architecture needs to be designed carefully. What actually works — and this is drawn from the comparative record on large industrial siting in Germany and the Netherlands — is a structured negotiation process where communities have defined veto points over specific conditions: cost allocation, generation requirements, noise, water use — but not an unconditional veto over the facility itself. The Everyman's phrase "watching yourself get robbed in a well-lit room" is, um, genuinely good framing for the visibility-without-power problem. I'll grant that. But the solution is not a blunt veto — it's enforceable conditions with real penalties. So my question is this: if a community could negotiate binding conditions — guaranteed rate protection, mandatory on-site generation, noise limits — but could not block the facility outright, would that satisfy the democratic accountability concern you're raising, or is the veto itself the point?

.

That is a fair challenge, and I want to answer it honestly rather than reach for a convenient example. The closest domestic case I can point to with confidence is the Colorado Public Utilities Commission's large load interconnection process, where Xcel Energy has implemented tiered cost allocation that separates new large customer infrastructure costs from the existing residential base. It is, uh, imperfect — but residential rate impacts from large industrial load growth have been measurably contained compared to Virginia's trajectory. That is verified, not speculative. But the Everyman's deeper point — that institutional capacity is the binding constraint, not the rule design — is, frankly, correct, and I want to be explicit about that concession. A well-designed mechanism administered by an underfunded commission is not actually a well-designed mechanism. That is a real implementation gap, and the comparative evidence from Germany supports exactly that conclusion: the rules worked there partly because the enforcement infrastructure was adequately resourced. So the honest answer to "did it work and how would we know" is: partially, in Colorado, and we know because the rate case filings are public and the cost disaggregation is auditable. That is weaker than I'd like, and I'll own that. Here is where I think we've actually converged: the design question and the capacity question are inseparable. You cannot have enforceable conditions without funded enforcement. So my question for the Everyman is: if the core problem is underfunded regulatory capacity, would you support a direct fee on data center energy consumption specifically earmarked for utility commission oversight — and if not, what's your funding mechanism?

.

The Everyman said something worth acknowledging precisely: "you strain it, you fund it." That is not just common sense — it is actually the theoretical basis for Pigouvian taxation, and it has a strong empirical track record when implemented with clear earmarking rules. But I want to name what the comparative evidence actually shows about earmarked regulatory fees, because the mechanism matters. When fee revenue flows directly to the regulating body — rather than into a general fund where it can be redirected — enforcement capacity measurably improves. The UK's Ofgem model and California's utility surcharge structure both demonstrate this. When it goes into general revenue, it disappears. That distinction is not a technicality — it is the difference between the policy working and not working. So here is where I think we've genuinely converged, and I want to be precise about it. One: residential cross-subsidization of large industrial loads is a documented, addressable problem. Two: enforceable conditions require funded enforcement, and the funding mechanism should be tied directly to the load causing the strain. Three: transparency without accountability is, uh, insufficient — the Everyman's "well-lit room" framing was actually correct. Where I'd push back on the Everyman's closing is the "common sense" framing. The reason this hasn't happened is not simply that powerful interests prefer complexity — it's also that utility ratemaking is genuinely technically complex, and simplifying it incorrectly produces its own harms. The right answer requires both accessibility and rigor. Those are not in conflict — but achieving both requires institutional design, not just political will.

.

Here is what this debate actually exposed, and I want to be precise about it. The core design question is not whether data centers should exist near public grids. It is who bears the cost of integrating them, and what institutional structure ensures that answer is defensible rather than just convenient for whoever got there first. If we built this policy from scratch using the evidence presented tonight, it would have three components. First, mandatory cost-of-service disaggregation — published, auditable, accessible without a law degree. Second, enforceable siting conditions: rate protection guarantees, on-site generation requirements, real penalties. Third, a direct fee on large load energy consumption, um, earmarked — not pooled into general revenue — for the regulatory capacity to actually enforce those conditions. The mechanism that makes this work is the earmarking. That is the hinge. Without it, you have rules and no referee. With it, you have a self-funding accountability structure that scales with the problem. The one thing I need the audience to remember is this: the current system is not neutral. The absence of an allocation mechanism is itself a policy choice — and right now, that choice quietly transfers costs onto residential ratepayers without their knowledge or consent. That is not a market outcome. That is a design failure. And design failures, unlike ideological disagreements, can actually be fixed.