On October 29, 2026, the Information Technology Industry Council (ITI) issued a set of industry-led commitments for responsible data-center growth, titled “ITI Issues Industry Commitments for Responsible Data Center Growth.” The announcement, made in Washington, is the trade group’s most direct attempt yet to answer the question that now follows nearly every large data-center proposal into a hearing room: who pays when a data center arrives?
The question is not rhetorical. A hyperscale campus can draw power on the scale of a small city, and the substations, transmission lines, and generation built to serve it cost real money. If regulators assign those costs carelessly, they land on household bills. If they assign them too aggressively, developers argue, they slow the infrastructure build-out the country says it wants. ITI’s roadmap tries to settle the argument before a commission does — by putting the industry on record that developers will carry the costs themselves.
The five commitments
The roadmap names five promises. First, the industry will build, bring, or buy the new generation resources and electricity needed to meet data-center demand. Second, it will pay for all new power-delivery infrastructure upgrades required to serve data centers. Third, developers will voluntarily negotiate separate rate structures with utilities and relevant state governments. Fourth, the industry will invest in local communities through local hiring and workforce training. Fifth, it will coordinate backup-generation resources with communities and grid operators during periods of scarcity.
Each verb matters, and so does each omission. “Build, bring, or buy” is deliberately flexible: a developer may fund a new plant, arrive with supply it has already contracted, or purchase power from existing or planned sources. What the language locks in is the principle that incremental demand should arrive with incremental supply, rather than leaning on a grid that was planned before anyone imagined a 200-megawatt customer. The announcement material available so far does not specify megawatt thresholds, deadlines, or how “new” generation is measured against existing capacity, and it does not say who audits compliance. Those are the details utilities and regulators will ask about first, and until ITI or its members publish them, they should be treated as unconfirmed.
The second and third commitments are the heart of the document, because they address the wires and the bill. Paying for “all new power-delivery infrastructure upgrades” is a broad promise — substations, line extensions, and the deeper network upgrades a campus triggers. The traditional principle here is cost-causer-pays, but as this desk has explained before, the boundary is blurry: direct connection facilities are clearly the customer’s responsibility, while deeper upgrades that also benefit other customers or relieve existing congestion raise harder questions about sharing. ITI’s language does not resolve those questions. It stakes out an opening position: the campus pays.
Why “voluntarily” is the most important word
The third commitment deserves the closest reading, because it contains the document’s most important qualifier. Developers will “voluntarily negotiate” separate rate structures with utilities and state governments. That is a promise to negotiate, not a promise of a specific tariff, and it is not a substitute for a commission order. State commissions still decide what gets approved, and a separate rate class for large loads only protects other customers if the commission writes it that way and enforces it.
That distinction is the whole story of the last year in large-load rate design. Binding rules, where they exist, look very different from pledges. In Oregon, regulators approved Portland General Electric’s large-load tariff framework in May 2026 — Schedule 96 — which establishes a dedicated rate class for loads above 20 megawatts, requires customers to cover 100 percent of distribution network upgrades, sets minimum demand charges at 90 percent of contracted capacity, and stretches contract terms from 10 years to 30 years for the largest loads, with a surcharge for projects above 100 megawatts. That is a tariff: enforceable, measurable, and tied to actual bills.
Elsewhere the fight is still open. In North Carolina, Duke Energy reached an October 7 agreement with Google, Microsoft, Amazon, and other stakeholders on large-load contract terms, but the settlement uses the existing rate structure with new rules rather than creating a separate tariff class, and critics have noted that some transmission upgrade costs would initially be spread across all customers and repaid over the contract term — which raises rates in the short term and leaves residual risk if a customer exits early. California moved in September to create a separate rate class for high-consumption facilities with upfront upgrade payments and long-term contract obligations, and Virginia’s governor announced a Data Center Accountability Framework in September pairing an executive order with legislation proposed for 2027. At the federal level, a House bill that would have asked states to consider requiring loads of 100 megawatts or more to pay the full incremental cost of the facilities built to serve them passed 417–3 in September, then stalled in the Senate on a 57–43 cloture vote, three votes short of the 60 needed.
ITI’s roadmap lands in the middle of that patchwork. It is best read as a bridge: it sets expectations now, in every state at once, while binding rules arrive state by state and take years to finalize. The roadmap’s reference to negotiating with state governments suggests ITI knows where the real authority sits. A commission weighing a large-load tariff can now point to the industry’s own commitment to negotiate a separate rate structure — and can ask a developer to put that commitment in writing, in a tariff, where breaking it has consequences.
Bring your own power: Vantage and Liberty Energy
The roadmap points to company examples, and the largest is the partnership between Vantage Data Centers and Liberty Energy. Announced in early January 2026, the agreement has Liberty Power Innovations, a Liberty Energy subsidiary, develop, own, and operate utility-scale power solutions for Vantage’s North American campuses, with the partners planning to deliver up to 1 gigawatt of power capacity over five years, including a reserved 400 megawatts of generation capacity scheduled for 2027, with potential to expand beyond 1 gigawatt.
In grid-constrained markets, that arrangement lets a campus move ahead without waiting years for transmission upgrades. It also illustrates why the “bring your own power” idea favors scale: financing and operating a gigawatt of generation is a different business from building a data center, and it pushes developers toward energy-company partnerships rather than going it alone.
There are limits to what the example proves. The partnership’s site locations, generation technology mix, and financing terms have not been disclosed in the public record, and a gigawatt of planned capacity is not a gigawatt of energized capacity. What it demonstrates is a business model taking shape — the developer teams with an energy company so the power arrives with the campus — and that model fits ITI’s first commitment directly. Expect more pairings in this style, because a developer without its own power plan will have a weaker case before a local board than one that arrives with generation in hand.
Backup power as a grid resource: Iron Mountain
The second company example points in a different direction: not generation, but storage. Iron Mountain Data Centers is building a 12-megawatt, 23-megawatt-hour battery energy storage system at its NJE-1 facility in Edison, New Jersey, developed with Calibrant Energy and scheduled for service by the end of 2026, alongside an existing 18,000-panel rooftop solar array rated at 7.2 megawatts.
Storage solves a different problem than generation. It shifts when a facility draws power, soaks up surplus supply, and lets a campus ride through short grid constraints without firing a generator. When the grid is under stress, the battery can discharge to reduce the facility’s demand; when demand eases, it recharges from off-peak grid power or on-site solar. That connects directly to ITI’s fifth commitment — coordinating backup-generation resources with communities and grid operators during scarcity — because a battery that can be dispatched without touching critical IT load turns a passive consumer into something closer to a grid partner.
Any plan to run backup generators during scarcity still carries air-permit and noise implications for neighbors — presumably why ITI frames coordination around communities as well as grid operators. The Iron Mountain project is a single site, not an industry-wide transformation, but it shows what the scarcity commitment looks like when it is hardware rather than language.
The workforce promise, and why it is the easiest to dilute
The fourth commitment — invest in local communities through hiring and workforce training — addresses the other half of the siting problem. A 200-megawatt hyperscale facility might employ 100 to 300 permanent staff once operational, well under one job per megawatt, while a manufacturing plant of comparable power demand might employ thousands. Communities that grant tax incentives expecting manufacturing-scale job creation can feel shortchanged, and that fairness question has become a regular feature of zoning hearings and moratorium debates.
Hiring and training pledges try to close that gap, and the industry needs electricians, technicians, and operators. Training programs aimed at local residents could serve both goals, giving host communities a stake in the facility and feeding a pipeline the industry needs. But ITI’s announcement gives no hiring targets, funding amounts, or partner institutions. Without numbers, this is the easiest commitment to dilute — and the one most likely to become a siting condition anyway, as local boards ask developers to attach figures when they file applications.
What would make this pledge credible
The honest assessment is that ITI’s roadmap is simultaneously meaningful and insufficient, and both halves matter. It is meaningful because an industry trade group has now put its members on record, in public, for the core ratepayer-protection principles — bring new supply, pay for delivery upgrades, accept a separate rate structure — that consumer advocates have been demanding in rate cases. Regulators can cite it, and local officials can quote it back to a developer.
It is insufficient because a voluntary pledge has no enforcement mechanism, no published signatory list, and no audit. The public record does not establish a full signatory list, attached dollar amounts or sites, or consequences for non-compliance. A pledge without a roster invites the obvious question — who, exactly, promised? — and a pledge without verification invites pressure from consumer advocates and local officials to supply one.
There are three tests to watch. First, whether state commissions cite the roadmap in large-load tariff orders and convert its language into binding terms — minimum bills, collateral, long contract terms, and customer-paid upgrades of the kind Oregon has already adopted. Second, whether individual developers attach numbers to the workforce and generation commitments when they file siting applications, so that a national pledge becomes a project-level obligation. Third, whether the bring-your-own-power model scales beyond the largest developers, or whether it becomes another advantage of balance-sheet scale, leaving smaller developers to negotiate interconnection queues the old way.
Pledges do not replace tariffs — but they change the negotiation
None of this makes the tariff fights go away, and ITI presumably does not expect it to. A voluntary commitment cannot set a rate, cannot order a transmission line, and cannot bind a future commission. What it can do is change the opening position of every negotiation that follows. A developer that has publicly promised to pay for delivery upgrades and negotiate a separate rate structure has a harder time arguing, in a rate case, that those costs should be socialized. A utility that can point to an industry-wide commitment has an easier time asking regulators to approve the infrastructure a campus needs. A community that can hold a developer to its own trade group’s words has more leverage than one negotiating from scratch.
That is the real significance of October 29. The data-center industry spent years arguing that it needed more power, faster. It is now arguing that it will bring the power, pay for the wires, and accept its own rate class — and it has said so in a single document, on the record, in Washington. The question that follows is the one regulators have always asked, and that no pledge can answer by itself: when the campus is built and the bills arrive, who actually pays? The tariff cases now working through commissions in Oregon, North Carolina, California, Virginia, and elsewhere will decide that question project by project, state by state. ITI’s commitments give those proceedings a starting point. They do not give them an ending.