Fred Stafford · fred@publicpowerreview.org
Version 4.0 · July 29, 2026 · supersedes v3.0 of July 14, 2026
Companion essay, Breakthrough Journal:
The Public Power Anchor, Part I ·
Part II
This document assumes the reader knows the motivating argument — the trust problem in federal transmission, the stealth deregulation critique, the Pacific Intertie precedent — and presents the machinery.
The National Anchor Power Program builds new clean firm generation — initially Westinghouse AP1000s on pre-assessed, publicly-owned nuclear sites — on the federal balance sheet, and transfers each plant at commercial operation to a qualifying publicly-owned utility at a fixed acquisition price set at financial close. Interregional anchor lines deliver a contracted share of plant output to neighboring investor-owned utilities for resale to hyperscaler data center load. Every anchor corridor is a pair of parallel ±525 kV HVDC lines: a contracted delivery line running continuously to the destination utility, with tap converters where intervening utilities take firm offtake along the route, and a smaller bidirectional reliability line terminating at the first balancing-authority seam, statutorily reserved for emergency exchange. Transmission ownership follows the Pacific Intertie principle that ownership settles cost allocation, generalized as: territory assets to territory owners, interregional assets to a federal anchor authority, with operational dispatch of both lines delegated to the anchor public utility. A unified National Anchor Program Reserve Fund, funded by per-MWh contributions from every offtaker on the corridor, backs both construction cost overruns and the reliability line’s carrying cost through layered beneficiary-pays stacks in which the federal share is a derived residual rather than a fixed percentage. Federal cost-overrun protection exists only for the public plants in this program — the New Deal preference principle, updated from price to risk.
Five principles generate the architecture; every mechanism below applies one or more of them.
Public anchor. Every anchor line terminates, at its sending end, in a federally-built plant owned and operated by a public utility. There is no contested private beneficiary behind the transmission, and therefore no allocation dispute for a federal siting authority to adjudicate. This is what dissolves — rather than manages — the trust problem.
Ownership settles cost, for excludable benefits. Each owner finances, rate-bases, and recovers its own assets through its own mechanism. Nothing is pooled; no inter-utility cost allocation methodology is invoked; no FERC allocation docket is opened. Ownership tracks function: network assets that physically merge into a territory’s grid belong to that territory’s utility; each interregional HVDC link belongs to the entity whose function is the interregion itself; and on paired corridors the principle extends one level deeper, so that each function has its own line and every cost attaches to a discrete physical asset.
This principle has a domain, and the domain must be stated rather than assumed. Segmentation dissolves cost allocation only where the benefit is excludable. Contracted power goes to identified buyers, so a line can be drawn down the asset and each portion assigned to whoever receives what flows through it. The reliability line’s benefit — emergency transfer capability within a balancing authority — is non-rival and non-excludable. Owning thirty percent of an emergency tie does not confer thirty percent of the reliability; it confers all of it, as does every other share. No partition of a public good is possible, and segmentation therefore has nothing to grip. The reliability line requires a second and different mechanism, designated residual claimancy, set out below. The program dissolves cost allocation on the contracted vintage and renders it non-binding on the reliability vintage. It does not claim to dissolve it everywhere, and a claim that broad would not survive contact with a serious reviewer.
Market structures preserved. No participating state or utility changes its market structure. Non-RTO IOUs remain cost-of-service and vertically integrated; RTO members remain RTO members; the anchor public utility retains its dispatch autonomy. HVDC provides this preservation at the physics layer — no loop flow, no synchronous entanglement across seams — and federal ownership of the links provides it at the jurisdictional layer, since the links are not RTO transmission assets and sit outside the planning and tariff machinery that attaches to member-owned facilities.
Layered beneficiary pays. Both construction overrun risk and reliability line carrying costs are recovered through defined layers in which every constituency with a structural interest pays a share calibrated to its benefit. No party bears a full layer alone; no party is fully insulated. Where a benefit cannot be assigned to any reachable party, the federal government is the designated residual claimant — not as a discretionary judgment about national interest, but as the arithmetic consequence of assessing everyone who can be assessed.
Public preference, modernized. Only publicly-owned utilities may acquire federally-built plants, and only such plants receive federal cost-overrun absorption. IOUs participate as transmission co-owners, wholesale buyers, and retail intermediaries — meaningful commercial roles aligned with their interests — but cannot be acquirers and cannot directly receive federal construction-risk protection. Public utilities crossed by or adjacent to a corridor may participate more deeply than IOUs at every layer, because their tax-exempt financing carries no equity return and their governance answers to member communities rather than to a commission. Socialized construction risk is legitimate where the public that bears the downside owns the upside.
The architecture is frequently misread because three distinct questions are assumed to have the same answer. They do not, and forcing them to agree would break the design.
Ownership is set by jurisdictional necessity and cost of capital. A converter must be federally majority-owned inside an RTO footprint because any other structure hands the RTO a tariff hook; a line span must be federally owned rather than IOU-owned because state rate-basing places it inside a commission’s allocation discretion.
Cost recovery is set by benefit incidence. Who pays has nothing to do with who holds title.
Operations is set by contract. Operating agency has never been proportional to equity: LADWP operates the entire southern segment of the Pacific DC Intertie while holding roughly forty percent of it, under agreement rather than by right of ownership.
The practical consequence is that the federal anchor authority may own sixty to seventy percent of a converter station, pay twenty-five percent of a line’s carrying cost, and dispatch none of it. That divergence is the design working. Any published version of this architecture should state the separation explicitly, because a reader who assumes the columns should match will read the mismatch as an error.
Federal construction-phase ownership. New plants are built and owned during construction by a federal construction entity — DOE, DOC under the October 2025 $80 billion Westinghouse/Brookfield/Cameco partnership framework, a dedicated federal nuclear construction office, or a combination. Construction is financed through Treasury borrowing, appropriations, and the partnership’s existing authority, with Office of Energy Dominance Financing credit support. The federal entity contracts with Westinghouse and supporting contractors against defined specifications and bears construction risk as principal. The designated acquiring utility participates from the outset as future operator — operating-readiness review, workforce integration, procedural alignment — so operations transfer without delay at commercial operation.
Siting on pre-assessed public sites. Phase 1 plants are sited on never-completed nuclear sites in TVA territory. Two remain TVA-owned: Bellefonte (Alabama), with clean title confirmed by final judgment in 2021–22, and Hartsville (Tennessee), originally scoped for roughly 5,000 MW and under active TVA site-clearance preparation — together roughly 7 GW of TVA-owned site envelope with prior siting analysis, water rights, and standing or restorable interconnection. The third, Yellow Creek (Mississippi), passed from TVA to NASA in 1988 and now rests with Tishomingo County; its pilot proceeds through an explicit federal reacquisition — county-to-federal conveyance and disposition of the existing lease on the nuclear portions — returning the site to its original public purpose after two federal-project cancellations. Where the acquiring utility owns the site it retains ownership throughout, granting the federal entity construction access.
Fixed acquisition price. At commercial operation the plant transfers to the acquiring public utility at a price set contractually at financial close, reflecting the original cost estimate, with adjustment only for defined force-majeure and regulatory-change circumstances unrelated to construction execution. The acquiring utility finances the acquisition through its own bond authority; its rate base reflects the acquisition price regardless of actual construction cost. The acquisition price is a yardstick: it is what the plant should cost, priced federally in advance, and it is the number against which all overrun layers are measured.
Qualifying acquirers. TVA leads Phase 1. The framework’s requirement is publicly-owned operation at adequate scale and nuclear readiness, not TVA’s specific charter: NYPA, the Nebraska public power group (NPPD, OPPD, LES), LADWP, SRP, CPS Energy, and JEA are Phase 2 candidates. Eligibility is statutory and exclusive to public power.
Each anchor corridor performs two functions with distinct commercial characters and distinct cost recovery: contracted delivery and interregional reliability. Both are realized as ±525 kV HVDC, and the two functions are realized as two parallel physical lines sharing right-of-way, permitting, converter siting, ownership pattern, and operating authority.
All corridors are HVDC. Version 3 carried one AC corridor as a simplification, with its reliability function embedded as headroom on a single integrated line. That exception is withdrawn. Parallel AC lines share flow by network impedance rather than dispatcher assignment, so an AC corridor cannot separate its two functions physically and must instead carve notional capacity out of a shared asset — reintroducing precisely the allocation methodology the paired structure exists to eliminate, and leaving the reliability function without a discrete cost object for any commission to review. The AC exception also created a second inconsistency: it left one corridor without converter stations, and therefore without the jurisdictional mechanism on which the rest of the architecture depends. Uniform HVDC costs more in converter capital and buys uniformity of ownership pattern, cost recovery, jurisdictional treatment, and legislative drafting across every corridor and every phase.
The contracted line (Line A). A unidirectional HVDC link sized to total contracted capacity — roughly 1,000–1,900 MW per corridor depending on tap participation — carrying firm continuous delivery from the anchor plant to the destination IOU under a long-term (20–25 year) wholesale capacity contract, with its full capacity committed and no reserved headroom. Where intervening utilities take firm offtake, Line A becomes a multi-terminal scheme with tap converters sized to each intervening party’s contracted share.
The wholesale price recovers the acquiring utility’s plant acquisition cost, operating and fuel costs, and transmission charges, including a federal segment usage fee passed through the wholesale rate. The destination IOU resells the capacity at retail to hyperscaler data center load under state-approved PPA terms that additionally recover the IOU’s own transmission assets, delivery infrastructure, and regulated margin. The IOU keeps the retail relationship — the politically essential property — and the hyperscaler receives clean firm power at federally-financed cost without construction risk, never becoming a direct customer of the anchor utility across territory lines. The same structure applies at each tap: the intervening utility resells to its own contracted large-load class and keeps that retail relationship.
The reliability line (Line B). A smaller bidirectional HVDC link — 500 MW per corridor — carrying no contracted commercial flow, dedicated to emergency exchange between the anchor utility and the destination region during asymmetric stress events, in whichever direction the stress requires. This is capacity reallocation across geographic diversity, not new generation; its value is real but bounded, and the program prices it honestly rather than claiming fleet-scale interregional benefits. The reliability line is what no private merchant developer will ever build and hold: a standing asset maintained for someone else’s emergency.
Line B terminates at the first balancing-authority seam and does not continue to the load center. This is the second substantive change from version 3, which carried both lines coterminously to the destination. A balancing authority is a single reliability entity across its whole footprint, so a reliability line landing at the first point inside the destination BA has crossed exactly the same seam as one landing hundreds of miles deeper. The additional distance buys no additional balancing-authority crossing, no additional market interface, and no additional interconnection diversity. Line A is long because the load is far; Line B is short because the seam is near, and each line should be sized to its own purpose rather than to the corridor they happen to share.
Three further reasons make seam termination not merely economical but structurally necessary. A reliability line extended into an RTO’s interior would be sized to relieve internal transfer constraints — which is the RTO’s own planning function, allocated through its own tariff, and a federally-owned line performing it is the strongest available predicate for RTO jurisdictional assertion over the asset. Line B’s most robust defense against commercialization is not the statutory bar but operational reality: the anchor utility cannot enroll the line in any market because the line is not operationally separated from the anchor utility’s own grid. A segment running between two points inside a foreign RTO, with no connection to the anchor system at either end, would be a federally-owned facility internal to that RTO dispatched by a non-member — an anomaly the seam-crossing segment simply does not present. And three-terminal schemes on both lines of a corridor would double the multi-terminal fault-clearing problem for no reliability gain.
Line B takes no taps. Contracted energy requires a physical delivery point; reliability benefit propagates to other parties through the destination balancing authority’s own adequacy construct once the line reaches it. An intervening utility therefore receives a Line A tap and no Line B position, unless the seam itself falls within its territory, in which case it hosts the Line B terminal.
Scope and protection. Line B is statutorily reserved for emergency exchange only — identified as a protected asset by physical and ownership specification, not as a capacity category on a shared line — and barred from RTO markets, imbalance markets, and commercial bilateral dispatch outside defined emergency protocols. The anchor utility dispatches both lines as integrated components of one resource adequacy posture; stress-event support runs under defined operating protocols, not RTO dispatch authority. Commercialization pressure must therefore overcome statutory prohibition and operational reality together.
Why the pairing. The pairing resolves an operational bind a single shared line cannot: under asymmetric stress on the anchor side, a shared line forces a choice between curtailing contracted delivery — precisely when the destination’s data centers can least afford it — and forgoing the emergency import the reliability capacity exists to provide. Paired, Line A delivers uninterrupted while Line B moves emergency capacity independently. The pairing is also what allows the anchor utility’s emergency recourse to be reserved to Line B, which is what permits Line A to be genuinely non-recallable and therefore genuinely firm for PPA and capacity-accreditation purposes.
The two-function structure is likewise the program’s answer to the cost-socialization objection. The contracted line’s costs are paid entirely by its beneficiaries through the PPA stack. The reliability line’s value flows to both endpoints and beyond, and its costs are recovered through the layered stack below. Because the reliability function is a discrete asset with a discrete capital cost, every layer of that recovery attaches to something concrete: a commission reviewing a utility’s share reviews a specific line’s carrying cost per retail customer, not an allocation methodology carving notional capacity out of a shared asset.
Version 3 addressed the two endpoints of a corridor and treated everything between them as distance. It is not distance; it is territory belonging to third parties who receive nothing under an endpoint-only design and whose landowners bear the structures. Transit without benefit is the grievance structure that has killed interregional lines, and a design that hands the problem to backstop siting authority has relocated the trust problem rather than dissolved it. Corridor participation is the correction.
What an intervening utility is offered. Contracted offtake from the anchor plant, delivered through a tap converter on Line A sized to its share, plus a minority economic position in that tap converter. The tap converts a transit host into a beneficiary with a signed contract and a rate-base position, which is the only thing that dissolves rather than manages the objection.
What it is not offered. Segment ownership of Line A through its own territory, where the intervening utility is investor-owned. Moving cross-jurisdictional spans into a state rate base places the corridor’s largest single asset inside a commission’s allocation discretion, and household exclusion would then depend on that commission’s continuing willingness to maintain it in every future rate case — a promise the architecture otherwise never has to make. It also introduces adverse selection, since parties would elect where terrain is easy and local load is good and decline where neither holds, leaving the federal residual systematically worse than the corridor average.
The public-power exception. The cost-of-capital objection is specific to investor-owned utilities. A cooperative or municipal system crossed by the corridor finances with tax-exempt debt and earns no equity return, so its ownership of a line span costs the structure nothing. The jurisdictional objection is specific to converters, which remain federally majority-owned regardless of who sits downstream. Segment ownership optionality therefore extends to public intervening utilities and not to IOUs — the preference principle expressed one layer further up the corridor.
Eligibility: an identified large-load beneficiary is required. Reserve Fund liability attaches to offtake from federally-anchored plants, so whoever sits behind a tap enters the overrun stack. If an intervening utility took its share into general system supply, its households would be paying Fund contributions through their rates and the program’s household-exclusion property would break in the middle of the corridor. Exempting them instead would create a free rider on protection every other offtaker funds. A tap therefore requires a contracted large-load class of the same character as the terminal offtaker.
This condition bounds the mechanism, and the bound should be stated plainly. Corridor participation dissolves the intervening-party problem where a large load happens to sit and manages it everywhere else. A crossed cooperative with no large-load class has nothing to attach a contract to, and for those parties the program is back to backstop siting and compensation for easement. Claiming more would be the same overreach as claiming segmentation dissolves cost allocation generally.
Published, non-discretionary terms. Corridor route, total capacity, available tap capacity, tap siting, segment boundaries, and offtake price are published in advance. Any utility whose territory is crossed may elect on those terms within a defined window. There is no discretionary selection of who participates, and oversubscription is resolved by statutory rule rather than administrative judgment.
This sequencing is the Pacific Intertie’s own: the federal side designed and priced a complete system and published the yardstick to proposers before soliciting, rather than negotiating bilaterally with each party in turn. It is also what keeps renewed backstop siting authority from reintroducing the umpire problem. Condemnation across a party that was offered participation on published terms and declined is a categorically different posture — in litigation, before a court, and before Congress — than condemnation across a party never asked.
Elections close at financial close. Reserve Fund contributions begin during construction and are credited against metered charges after commercial operation, so a party electing later would ride free on construction-period risk. The election window closes at the moment the acquisition price is struck, and the two events are deliberately simultaneous.
Renegotiation is statutorily barred. Published terms are worth nothing without a commitment device. An N-party structure otherwise invites the last party to sign to hold out for improved terms, and the federal entity’s ability to refuse must be statutory rather than a matter of resolve.
The inducement is dilution, not charity. The Fund’s obligations are fixed against a plant of known cost and a reliability line of known cost. Adding an offtaker enlarges the contribution base without enlarging the obligation, so the per-MWh rate falls for every contributor. Since the contribution rate is the term hyperscaler procurement teams negotiate hardest, “each additional participant lowers everyone’s contribution” is a substantive argument that runs in the same direction as the political case for buying in transit parties.
Engineering consequences. A tapped Line A is a multi-terminal HVDC scheme, which is materially harder than adding a converter. A DC fault anywhere on a multi-terminal system blocks every converter unless the scheme can clear selectively, which requires either full-bridge MMC converters — roughly a third more semiconductors and a standing loss penalty at every station — or hybrid DC breakers, which exist in service but have thin operating history. Full-bridge is the base case; breakers are the alternative; the choice must be priced rather than assumed. The alternative topology of two cascaded two-terminal links with AC interposition at the tap uses entirely conventional equipment but costs two additional converter stations, adds conversion losses on all through-power, and puts destination-bound flow onto the intervening utility’s AC bus, which is the entanglement the architecture exists to avoid. It should be priced as the fallback and not adopted as the design.
The capacity budget binds. Total contracted capacity across the terminal offtaker and all taps cannot exceed the anchor plant’s net output less operating margin. A twin-AP1000 anchor at roughly 2,234 MW net supports on the order of 1,900 MW of total commitment. This, rather than any political consideration, is the ceiling on how many intervening parties one anchor can buy in, and it should be stated in corridor solicitations so that available tap capacity is a published number rather than a negotiation.
Anchor utility segments. The acquiring public utility owns the line segment within its statutory territory — the in-territory DC portion of both lines plus both sending converter stations — financed under its bond authority and recovered through its wholesale rate. No new authority required.
Destination utility segments. The destination utility owns the AC infrastructure from the receiving converter station to load: the delivery line and the network reinforcement — substations, reconductoring, VAR compensation, protection upgrades — required to integrate the injection. An IOU rate-bases these through state retail ratemaking, recovered through the hyperscaler PPA. Standard utility transmission ownership; no novel authority.
Intervening utility segments. An intervening utility owns the AC infrastructure from its tap converter to its own large-load customers, recovered through its own retail ratemaking against that contracted class. A public intervening utility may additionally own the Line A span through its territory, financed with tax-exempt debt and recovered through the same channel.
Federal anchor authority segments. The federal anchor authority owns the cross-jurisdictional portions of both lines — for Line A, the full route from the anchor utility’s boundary to the receiving converter, in as many discrete spans as the corridor’s tap structure requires; for Line B, the shorter run from the anchor utility’s boundary to the seam terminal — plus majority stakes in every receiving and tap converter station.
Two properties of this ownership are load-bearing and neither is optional. Federal capital is Treasury-cost and carries no equity return, so nothing enters any state rate base and nothing is socialized across territories; the one federally-owned segment is also the one paid for entirely by its beneficiaries. And federal majority ownership at converters is the jurisdictional mechanism by which the links stay outside RTO planning and tariff machinery. The authority additionally concentrates build-phase capability — siting, permitting, eminent domain under renewed backstop authority, construction financing, asset stewardship — in one institution whose competence compounds across projects and across every Phase 2 acquirer, sparing each public utility a territorial-expansion fight it would otherwise face alone.
Note that the federal segment is not a single span. On a tapped corridor it is a set of spans separated by converter stations in which the federal authority holds majority but not sole ownership. The authority’s charter must therefore describe segment ownership as a set rather than a continuous route.
The receiving and tap converter stations. These are the structurally novel assets and the control points where HVDC’s autonomy properties would be surrendered if ownership were misassigned. A utility that owned and dispatched a converter would hold de facto operational authority over the anchor utility’s exports; a converter inside an RTO footprint owned by an RTO member becomes, through that RTO’s planning and tariff machinery, the legal hook by which the RTO reaches the entire link. The design therefore separates economic from operational ownership: federal majority (60–70%) with full operational authority, local utility minority (30–40%) as a purely economic, rate-based position with return on equity but no control rights. Every tap converter takes the same treatment as every receiving converter — the rule is a property of the asset class, not of its position on the corridor.
The minority stake serves two purposes. It makes the destination or intervening utility a co-investor rather than merely a customer, giving it a rate-base position commensurate with the retail relationship it holds and aligning its interest with the asset’s long-term performance. And it equalizes positions across the pilot coalition, so that no partner is structurally disadvantaged relative to another.
That second purpose now requires program-level calibration rather than the corridor-level comparison version 3 used. Version 3 set the minority percentages so that HVDC-pilot IOUs held rate-base positions comparable to the AC pilot’s longer owned segment; with the AC pilot withdrawn, that reference point no longer exists, and with tap converters added, some parties now hold positions in two stations while others hold one. Minority percentages should be set by the Anchor Program Office at program level against a defined parity metric, published with the corridor terms, and not inherited from any single corridor’s arithmetic.
Public power at the converter. Where the destination or intervening utility is publicly owned and outside an RTO, the jurisdictional justification for federal majority largely disappears — there is no RTO tariff, no regional stakeholder process, and no capacity market reaching in. In that case public-power majority ownership of the receiving converter is available and preferable: tax-exempt financing carries no equity return and no cost-of-capital penalty against Treasury, and a public authority answers to member communities rather than to a commission, removing the prudency proceeding from the path entirely.
The Pacific Intertie’s Sylmar Converter Station is the direct precedent — the receiving terminal of a federally-originated HVDC line, owned by a consortium of municipal utilities and one IOU, with a municipal utility as managing and operating agent. Where a public utility is an RTO member, the same result requires the interest to be held outside RTO functional control, as much PDCI and COTP capacity is held today under transmission ownership rights and existing transmission contracts. That is achievable but is an additional negotiation rather than a default.
Operational delegation. The federal authority delegates dispatch of its segments to the anchor public utility under a long-term operating agreement (initial term on the order of 40 years, FERC-filed, binding on successor administrations, with due-process limits on withdrawal). The anchor utility dispatches each full link — sending converter to receiving converter, through every tap, Line A and Line B alike — as integrated assets under one control room, with no operational handoff at any ownership boundary. The federal role in operations is stewardship and oversight (NERC compliance, security, financial accountability), not control. The anchor utility’s institutional reach extends through operations rather than ownership, avoiding any expansion of its statutory territory.
The control point is specified, not assumed. On a multi-terminal DC scheme one terminal regulates DC voltage while the others operate in power-control mode, and that terminal holds the operational whip. Conventional engineering practice assigns voltage regulation to the largest inverter, which on these corridors would be the destination terminal — quietly transferring the control point to the destination utility inside its own RTO and hollowing out the delegation. The operating agreement therefore specifies DC voltage regulation at the sending terminal, all receiving and tap terminals in power-control mode, and droop sharing defined for contingencies. This specification is the technical content of operational delegation; without it the delegation is asserted rather than effected.
Maintenance. The same operating agreement governs physical O&M of the federal segments. The anchor utility performs line and conventional substation maintenance — crews, vegetation management, storm response, outage scheduling — as a reimbursed service to the federal owner, at cost, with converter internals under long-term OEM service agreements administered through the same framework. This matches capability to role and keeps maintenance and dispatch, which are operationally entangled, inside one institution. In NERC registration terms the federal authority is Transmission Owner and the anchor utility Transmission Operator — a standard division. Reimbursed O&M flows into the federal segments’ revenue requirement, recovered through the same channels as their capital, so no appropriation funds routine maintenance and no ratepayer class carries it. The arrangement is the PDCI’s own pattern carried forward: LADWP operated and maintained southern-system equipment whose ownership it shared with SCE and municipal co-owners, under agreement, for decades.
Ownership, capacity, and scheduling rights are separable. The Intertie agreements distinguish line ownership from capacity ownership from scheduling rights, and the cities of Burbank and Glendale hold bidirectional scheduling rights over segments they do not proportionally own. That separation is the instrument for bringing public utilities into the capital stack without fragmenting control: defined emergency scheduling rights proportional to share are grantable without touching the anchor utility’s operating agency, and give a cooperative or municipal board something concrete to weigh against a capital contribution.
Both risk pools are recovered through layered stacks administered by a single National Anchor Program Reserve Fund.
Cost overrun stack. Measured against the acquisition price (AP):
The stack preserves the program’s two non-negotiables simultaneously: the acquiring utility’s ratepayers never face unbounded Vogtle-style pass-through, their exposure being capped at the defined first-loss band and priced into the wholesale rate from the start; and no plant is abandoned mid-construction for want of overrun financing — the V.C. Summer failure mode — because coverage above the first-loss band is pre-funded and statutory.
Tap offtakers sit in the Fund layer alongside the terminal offtaker. They are not acquirers and take no first-loss position.
Line A transmission capital carries no layered structure. Transmission construction risk is far less volatile than nuclear construction risk, and Line A capital recovers conventionally, pro rata by contracted capacity across the terminal offtaker and all taps. A 400 MW tap on a 1,900 MW line carries roughly twenty-one percent of Line A’s transmission revenue requirement.
Reliability line stack. The reliability line produces no commercial revenue by design, so its carrying cost is recovered through annual cost-of-service capacity payments. Seam termination materially reduces that cost relative to version 3: the saving is conductor, structures and towers rather than siting, since the corridor is cleared for Line A regardless, and the converter count is unchanged — though co-siting the Line B terminal with a Line A tap where the seam and the tap coincide shares civils, AC connection, spares and staffing, which discounts the second station meaningfully.
The layers are:
Anchor utility, ~15%, through its wholesale rate. Two justifications, and the second is the stronger. The benefit is genuinely contained: a vertically integrated public system captures emergency-import value across its own footprint and nowhere else. And the party that dispatches an asset must have capital at risk in it, or the program has created an operator with control and no exposure. This is also the cheapest layer to obtain, since a federal power agency’s self-set wholesale rates involve no commission and no prudency review. If anything it is understated, since marginal reliability value scales inversely with system size and the anchor system is the smaller party on every Phase 1 corridor.
Destination-side layer, variable. Set by destination structure, treated in the next section. Recovered through general retail rates rather than through the hyperscaler PPA, because the resource adequacy benefit accrues across a service area rather than to contracted load. A locational premium is earned only where the paying party sits inside a genuinely constrained zone; a party in unconstrained territory captures no more benefit than any other participant in its balancing authority, and charging it a locational share over-recovers against actual benefit.
Reserve Fund, ~50%, reflecting hyperscalers’ position as the principal drivers of the load growth that occasions the buildout, and the direct value of stress-event support to their operations.
Federal, the residual. Not a fixed percentage. What remains after every reachable beneficiary is assessed.
Designated residual claimancy. The federal layer is the mechanism that replaces segmentation on the reliability line, and it deserves to be named as such rather than treated as a funding gap.
On a conventional interregional line, cost allocation failure is fatal. Parties cannot agree on shares, no party is obliged to close the gap, and the line is not built — which is the documented failure mode of nearly every interregional project of the last two decades. Under this program, allocation failure raises the federal cost, because the federal layer is defined as the remainder. The negotiation cannot deadlock, because deadlock has a defined outcome and a party obliged to absorb it.
Line A dissolves the allocation battle by partition. Line B renders it non-binding by residual claimancy. Both remove the possibility of a corridor dying in a cost-allocation fight, which is the failure the program exists to prevent — but they do so by different mechanisms, and conflating them would misstate what the architecture can carry.
Capacity payments use cost-of-service methodology — documented infrastructure cost, normal regulatory oversight — rather than waiting on accreditation reforms that do not yet exist. Where such reforms arrive, the step-down described below applies.
The federal residual is not a policy choice. It is a function of how many beneficiaries the destination’s institutional structure makes reachable, and it varies substantially by corridor. This is a feature worth surfacing rather than smoothing: the number is derivable, which is exactly the property one wants when it is presented to appropriators.
Three regimes, in order of increasing federal exposure.
A single vertically integrated destination contains the benefit entirely within one utility’s footprint. One payer, one commission, and the destination-side layer is straightforward.
A holding company with multiple operating companies contains the benefit within the corporate family. An intercompany cost-sharing contract of the type filed at FERC by integrated Southeastern systems already apportions system benefits and savings among the operating companies against a worked allocation basis, typically load-ratio share. The argument writes itself: the contract already apportions the benefit, so apportion the cost by the same formula. Nothing is invented and no new negotiation is opened. The cost is multiple state commissions rather than one — the FERC-approved methodology is not relitigated, but prudency is, once per operating company, and operating companies receiving no anchor offtake will contest a cost with no offsetting revenue.
Joint transmission arrangements at the destination place additional beneficiaries inside the electrical system who are party to no intercompany contract — cooperatives and municipal systems that jointly own and plan a statewide transmission system with the destination IOU. These are not a cost-assessment problem. They are qualifying public entities under the program’s own eligibility rules, and the correct instrument is buy-in rather than assessment: equity participation reduces federal capital rather than merely federal carrying cost, at tax-exempt cost of capital, with no commission in the path and with institutional interests aligned against commercialization. Such arrangements also establish that the destination has decades of working practice at mixed IOU, cooperative and municipal joint ownership of transmission — evidence of feasibility, not merely of willingness.
An RTO destination is the hardest case, and the federal residual is correspondingly largest. RTO membership is numerous, non-affiliated, and has no contractual relationship to any payer. Only the directly-connected utility can be assessed, and even its locational share is earned only if it sits in a constrained zone — a fact that must be checked in the RTO’s own studies rather than assumed, since administrative price collars can suppress the market signal for years while the underlying constraint persists in the RTO’s unconstrained counterfactual.
Capacity market accreditation is closed to Line B on two independent grounds, and the second is closed by design rather than by rule.
RTO accreditation frameworks enumerate resource types, and a transmission line is not among them; the underlying effective-load-carrying-capability method compares hourly resource output against load shapes and presupposes a thing that generates. This is not a policy preference an RTO could reverse by choosing to — it is a structural absence of category.
More decisively, external capacity resources must be non-recallable, and Line B is the anchor utility’s recall mechanism. Reserving emergency recourse to the reliability vintage is what permits Line A to be firm; Line B cannot satisfy non-recallability without ceasing to be the thing that makes the pairing work.
The available route is the capacity benefit of ties. Every RTO reduces its reserve margin requirement by an allowance for emergency assistance from neighbors, and that allowance enters the requirement as a straight subtraction: more assumed tie assistance, lower procurement target, less capacity purchased by RTO load. The benefit is RTO-wide, quantifiable, and lands on exactly the population that receives the reliability spillover.
It is also, at present, free. RTOs assume tie assistance, reduce their requirements accordingly, and compensate nobody. That is defensible where ties exist for other reasons and were built anyway. It is not defensible for an asset built to order for that purpose.
The reform ask is therefore narrower than creating an accreditation category for transmission. The methodology already exists in RTO practice: add an increment to the modeled portfolio, measure the resulting improvement in expected unserved energy, and express it as a fraction of what a perfect resource would deliver. Nothing in that procedure requires the increment to be a generator — it requires a thing that improves expected unserved energy, which a purpose-built tie with defined transfer capability and a characterized counterparty availability distribution demonstrably is. The ask is: run the existing calculation on a purpose-built emergency tie, and let the result enter the tie allowance as a resource-specific increment rather than being absorbed into a pool-wide assumption reviewed in committee. That is a manual revision, not a market redesign.
Price in the capacity market, collect through transmission cost allocation. These are separate steps and separating them matters. Valuing the contribution at the capacity clearing price gives price discovery. Collecting through a transmission tariff — formula-based charges to zones, on the joint-operating-agreement model that governs seam facilities between RTOs — avoids two hazards. It avoids capacity performance obligations, which would expose the line to non-performance penalties in precisely the correlated stress events where the anchor system is itself short, converting a reliability asset into a financial liability borne by the anchor utility. And it avoids the statutory bar entirely, since having costs allocated through a transmission tariff is regulated cost recovery rather than market participation, requiring no amendment to the reliability line reservation.
The step-down is statutory and contingent. The federal share falls when the mechanism exists, not when it is promised. Legislation should specify that the federal reliability-line share reduces by a defined increment upon an RTO’s adoption of a resource-specific tie-benefit methodology and an associated cost allocation, with the released share picked up by RTO load. This converts a permanent federal obligation into a bridge, and it gives the program’s engagement with transmission economists a concrete ask rather than a general one.
Two conditions bound the route and both must be checked before it is relied upon. The tie-benefit allowance is capped by a maximum-assistance parameter; if an RTO’s allowance already sits at that ceiling, an incremental tie moves nothing. And the tie’s derated contribution must be materially nonzero, which depends on stress correlation between the anchor system and the destination RTO — a question no one has answered for any specific corridor.
There is no capacity market, so the tie-benefit route has no analogue. The functional equivalent is integrated resource planning: an asset that lowers a system’s required reserve margin appears as avoided capacity in the resource plan and is recognized at the state commission. The same logic — the system buys less because this exists — in a different institutional wrapper. Easier in that no tariff redesign is required; harder in that it may be several resource plans before several commissions rather than one filing.
Where the destination-side layer nonetheless proves hard to place, the residual-claimancy property means the corridor still proceeds. That is the point of the mechanism.
Corridors terminating in non-RTO territory with dense institutional structure carry materially smaller federal shares than corridors terminating in RTOs. This should be stated in the program’s own presentation rather than discovered by a reviewer, because it converts an apparent inconsistency across pilots into a derived result — and because it argues for sequencing at least one institutionally dense corridor early, where the federal ask is smallest and the demonstration cheapest.
The Fund is a dedicated federal trust fund unifying both coverage layers under one contribution structure.
Contributions. Per-MWh contributions on offtake from federally-anchored plants fund both layers, levied at a uniform rate on every offtaker — terminal and tap alike. Combined rates land in the $21–32/MWh range (roughly $10–15 for overrun coverage, $11–17 for reliability line payments), well below single-payer alternatives because the other layers carry meaningful shares, and low enough to keep anchor PPAs competitive with alternative clean firm procurement. Corridor participation reduces this rate, since Fund obligations are fixed against plants and lines of known cost while each additional offtaker enlarges the contribution base. Seam termination of Line B reduces it further.
Uniformity is not a simplification. A differentiated rate would require someone to exercise judgment about who pays what, which reintroduces exactly the discretionary determination the published-terms structure exists to eliminate.
Advance contributions. PPA contracts require contributions beginning during construction, credited against operational contributions after delivery begins, calibrated to accumulate meaningful reserves by first commercial operation so the coverage band is funded rather than merely nominal for the earliest pilots. The mechanism is CWIP inverted: advance capital raised during construction, but from the offtakers whose demand occasions the program rather than from households, held as protected trust reserves, entered voluntarily through PPA commitment. Contributions on a cancelled pilot remain pooled rather than refunded — a bounded concession consistent with the Fund’s risk-pooling logic. Because advance contributions begin at financial close, tap elections close at the same moment.
Uniform rate-setting. Contribution rates are set uniformly fleet-wide within statutory five-year intervals, adjusted only at interval boundaries on accumulated actuarial experience. Pure actuarial pricing would charge early PPAs more than later ones, and sophisticated procurement teams would rationally wait — a first-mover penalty that would prevent the Fund from ever accumulating. Uniform intervals eliminate the penalty; the modest additional federal tail exposure this transfers is bounded and appropriately federal. Phase 2 pilots join the same Fund under the same intervals. The methodology, intervals, and adjustment protocols are statutory, because administrative discretion would invite exactly the strategic timing the structure exists to preclude.
Capitalizing the federal reliability share. Where the federal residual is large, the preferred vehicle is a capitalized corpus within the Fund rather than annual appropriations. Commercialization pressure on the reliability line accumulates at the appropriations cycle, where the argument that an idle asset should be permitted to earn revenue and reduce the federal ask has its natural hearing. A capitalized corpus removes the annual proceeding to which that argument would attach. Where a corpus is not achievable, the fallback is mandatory rather than discretionary spending paired with a reliability service agreement matching the operating agreement’s term, so that withdrawal requires affirmative legislation rather than a failure to act.
Governance. Federal oversight of rate calibration and disbursement; an offtaker advisory mechanism without veto authority; state commission engagement on destination-side layers through normal rate proceedings. Because the Fund would hold federal capital alongside offtaker contributions against two distinct obligations, the priority of claims between overrun coverage and reliability payments in a year when both are called must be fixed in statute rather than left to Fund governance.
Three TVA-anchored corridors demonstrate the architecture across the full range of structural conditions. All are paired ±525 kV HVDC.
Bellefonte → Metro Atlanta (~250 miles; Line A ~1,500 MW, Line B 500 MW to the TVA/Southern seam). Twin AP1000s (~2,300 MW) at Bellefonte — the prior NRC reference site for a two-unit AP1000 combined license application (withdrawn 2016; the reusable asset is the docketed review record, not a live license), with a switchyard built for ~2,400 MW of original capacity; wholesale offtake by Georgia Power for metro Atlanta data center load.
Under version 3 this was the simplest pilot: an AC line, a two-owner split at the state line, no converters and no federal middle. That framing does not survive the HVDC conversion, and the more accurate description is nearly its inverse. Bellefonte is the institutionally densest corridor in Phase 1 and therefore the one with the smallest federal residual. It crosses a non-RTO destination whose operating companies apportion system benefits under a FERC-filed intercompany contract; it terminates in a state whose transmission is jointly owned and planned by the destination IOU together with a generation and transmission cooperative, a municipal joint action agency, and a municipal utility; and every one of those public participants is a qualifying entity under the program’s own eligibility rules and a candidate for converter equity. It is jurisdictionally simple and institutionally rich, which is precisely the combination that minimizes what the federal government must carry.
Hartsville → Northern Virginia (~500 miles; Line A ~1,500 MW plus tap capacity, Line B 500 MW to the TVA/PJM seam). The structurally consequential pilot: from a TVA-owned site originally permitted for four reactors, crossing from non-RTO TVA into PJM to serve the largest data center market in the world. Introduces the full multi-party paired-corridor architecture — TVA sending segments and converters, federal cross-jurisdictional spans, at least one intervening PJM utility taking firm offtake through a tap converter, and destination IOU downstream AC — plus converter ownership separation and operational delegation. It establishes that federal anchor generation can serve RTO-coordinated load through a non-RTO public operator, with the anchor utility dispatching both lines end to end and the RTO scheduling AC integration at its own buses, neither institution’s market structure altered. It carries the largest federal residual in Phase 1 and is therefore the corridor on which the tie-benefit step-down matters most.
Yellow Creek → Central Mississippi (~300 miles; Line A ~1,000 MW, Line B 500 MW to the TVA/MISO seam). The westward seam: TVA into MISO South, serving the central-Mississippi hyperscaler cluster. Proves the cross-RTO architecture portable across a second, differently-structured seam — and, through the federal reacquisition of the site, demonstrates the program’s mechanism for returning surrendered public sites to public service. An alternative western pilot originating at a TVA-owned coal site, which would eliminate the reacquisition dependency and offer a differently-structured MISO seam, is under evaluation and would substitute for this corridor rather than supplement it.
Aggregate Phase 1: roughly 7 GW of new clean firm generation, 1.5 GW of dedicated reliability capacity across three corridors, and roughly 1,050 Line A corridor-miles. Line B mileage is route-dependent and materially shorter than Line A on every corridor — plausibly 250–400 miles in total against the 1,050 that coterminous lines would have required, which is the principal capital saving from seam termination. Phase 2 extends the framework to NYPA (upstate generation, downstate IOU offtake within NYISO), Nebraska public power (MISO/SPP data center clusters), and LADWP (Western Interconnection, coordinated with existing Intertie infrastructure), demonstrating portability across acquirers, market structures, and regions. Phase 2 does not wait on Phase 1 completion.
New legislation.
Federal Anchor Generation Authority: federal construction-phase plant ownership, transfer mechanics, acquisition-price methodology, and statutory acquirer eligibility limited to publicly-owned utilities.
National Anchor Program Reserve Fund: the trust fund, both coverage layers, contribution and advance-contribution rules, uniform rate-setting intervals, disbursement criteria, priority of claims between the two obligations, authority to hold a capitalized federal reliability corpus, and governance.
Federal tail absorption authority with defined thresholds and reporting.
TVA Act amendments: a narrow fence carve-out authorizing wholesale sales from anchor plants to neighboring utilities for defined large-load classes — drafted to reach intervening utilities taking tap offtake as well as terminal destination utilities — and debt cap relief (to roughly $50–60 billion, or a dedicated acquisition-financing exception).
Renewed federal backstop siting authority scoped exclusively to transmission anchored to program generation — narrower than any general Section 216 revival, with correspondingly stronger constitutional footing. With the AC exception withdrawn, all anchor lines run at ±525 kV DC and the authority need not be drafted across voltage classes. The authority should be conditioned on prior publication of corridor participation terms and expiry of the election window, so that it applies only to territory whose utility was offered participation and declined or did not qualify.
Federal anchor authority charter: cross-jurisdictional segment ownership across both lines of each corridor, described as a set of spans rather than a continuous route; majority ownership of every receiving and tap converter station with local-utility minority economic participation; program-level calibration and publication of minority percentages; and operational delegation authority including the control-point specification.
Reliability line reservation: statutory dedication of each Line B — identified by physical and ownership specification — to emergency exchange, with express prohibition on participation in RTO or imbalance markets and on commercial bilateral dispatch outside defined emergency protocols. Drafted to permit cost recovery through transmission cost allocation without permitting market participation, since the two are distinct and only the latter is the hazard.
Corridor participation rules: publication requirements for corridor terms, the election window and its closure at financial close, the statutory bar on renegotiation, oversubscription resolution by rule, and the tap eligibility condition requiring an identified contracted large-load class.
Reliability share step-down: reduction of the federal reliability-line share by a defined increment upon an RTO’s adoption of a resource-specific tie-benefit methodology and associated cost allocation, contingent on the mechanism existing rather than promised against its arrival.
Existing authority. Federal capital-program and Treasury financing; the DOC/Westinghouse partnership; Office of Energy Dominance Financing credit support; IRA nuclear credits (45U, 48E — modified to flow to federal construction entities); FERC Order 679 incentives.
Administrative. A federal Anchor Program Office administering project evaluation, construction oversight, transfers, the Fund, corridor solicitations and participation elections, and operating agreements; a standardized joint-project documentation template drawing on the Intertie agreements, including the separation of line ownership from capacity ownership from scheduling rights; interagency coordination across DOE, DOC, FERC, TVA, NRC, and Treasury.
The acquiring public utility gets fleet expansion at fixed acquisition prices with bounded first-loss exposure and full dispatch autonomy — including over every asset it operates but does not own — plus emergency import capability its own customers pay a calibrated share for and its own operators control.
Destination utilities get firm clean wholesale capacity without construction risk, rate-base growth on their assets, retained retail relationships with the largest new customers in their territories, and no change to their market structure.
Intervening utilities get the same on a smaller scale: contracted offtake for their own large-load customers, a rate-base position in the tap converter serving them, and a retail relationship they would not otherwise have — instead of a transmission corridor through their territory serving someone else.
Public utilities crossed by or adjacent to a corridor get an equity route into interregional infrastructure at tax-exempt cost of capital, with defined emergency scheduling rights, and — where the destination sits outside an RTO — the possibility of majority ownership at the receiving converter.
Hyperscalers get long-term clean firm PPAs at federally-financed cost, uninterrupted contracted delivery even during anchor-side stress, and a bounded, statutorily predictable contribution rate that falls as corridor participation broadens.
General ratepayers on every segment of every corridor bear no construction overrun risk and pay only calibrated shares of a reliability capability that serves them — nothing at all on the contracted line.
The federal government gets clean firm deployment at fleet scale, industrial-base support for the AP1000 supply chain, and interregional infrastructure — with its exposure bounded to two tails it has chosen to stand behind as builder and as residual claimant, both of which shrink as reachable beneficiaries are reached.
Every ownership choice in the program, down to the title on a converter station, is made to preserve market structure against capture. That is the design’s answer to stealth deregulation: a buildout architected so that nothing about who builds, who owns, who pays, or who dispatches requires any participant to become something other than what it is.
Uniform HVDC. The AC corridor exception is withdrawn. All anchor lines are paired ±525 kV HVDC. This resolves an inconsistency in version 3 — the Bellefonte corridor was described as 500 kV AC while the reliability architecture assumed HVDC pairing — and removes the embedded-headroom variant entirely. Downstream corrections: the pairing rationale is no longer HVDC-specific; anchor utility segments are DC rather than “AC or DC”; the border-substation variant for destination segments is deleted; Phase 1 aggregate reliability capacity rises from 1.0 GW plus embedded headroom to 1.5 GW dedicated; and the backstop siting authority no longer requires drafting across voltage classes.
Corridor participation. New section. Intervening utilities are offered contracted offtake through Line A tap converters plus minority economic positions in those converters, on published non-discretionary terms within an election window closing at financial close, conditioned on an identified contracted large-load class. Segment ownership optionality extends to public intervening utilities only.
Line B terminates at the first balancing-authority seam rather than continuing to the load center, and takes no taps.
The federal reliability share is a derived residual rather than a fixed twenty percent, and varies by destination structure. Designated residual claimancy is named as the Line B analogue of segmentation.
The excludability limit is published. Segmentation dissolves cost allocation for excludable benefits only; the reliability line’s benefit is non-excludable and requires the residual mechanism instead.
Destination structure section. New. Sorts destinations into vertically integrated, holding-company, joint-transmission, and RTO regimes, with monetization routes for each: the capacity-benefit-of-ties route for RTOs, priced in the capacity market and collected through transmission cost allocation; integrated resource planning recognition for non-RTO destinations. Capacity market accreditation is identified as closed to Line B on resource-category and non-recallability grounds.
The minority converter stake rationale is rebuilt. Version 3 calibrated minority percentages against the AC pilot’s longer owned segment; with that pilot withdrawn and tap converters added, calibration moves to program level against a published parity metric.
The control point is specified. DC voltage regulation at the sending terminal, receiving and tap terminals in power-control mode.
Public power at the converter. Majority public ownership of receiving converters is available where the destination is publicly owned and outside an RTO, on the Sylmar precedent, with the ownership/capacity/scheduling-rights separation as the participation instrument.
Bellefonte is reframed from the simplest pilot to the institutionally densest one, carrying the smallest federal residual.
Three-column separation of ownership, cost recovery, and operations is stated explicitly as a new short section.
Legislative package expands from seven items to nine, adding corridor participation rules and the reliability share step-down, and amends items 4, 5, 6 and 7.
These are unresolved and should not be treated as settled in any external-facing version.
Tie-benefit ceiling. Whether any target RTO’s maximum-assistance parameter already binds, which would close the monetization route entirely. Requires the RTO’s own reserve requirement study rather than its manuals.
Tie derating. The derated contribution of a purpose-built emergency tie has not been calculated for any corridor, and depends on stress correlation between the anchor system and the destination region. Adjacent systems under common weather may show less diversity benefit than the interregional literature’s headline figures imply; resource-mix diversity may matter more than weather diversity and has not been separately assessed.
Intercompany contract scope. Whether an intercompany cost-sharing agreement of the Southeastern type can carry a transmission-type carrying cost, as distinct from the energy and capacity transactions it was written for, is a question for the rate schedule text rather than for summaries of it.
Coalition parity metric. The parity standard against which minority converter percentages are calibrated is undefined now that positions differ in number as well as size across corridors.
Converter topology pricing. Full-bridge MMC versus hybrid DC breakers on tapped corridors is unpriced, and is the principal cost uncertainty introduced by corridor participation.
Line B carrying cost after seam termination. The version 3 figure of $200–300 million annualized per corridor assumed coterminous lines. The reduction is conductor and structures only, with converter count unchanged and co-siting savings where the seam and a tap coincide, but no revised figure has been derived.
Local absorption at seam terminals. A 500 MW emergency injection into a smaller intervening utility’s network may require AC reinforcement that a large destination system would not, which is unstudied and could offset the seam-termination saving.
Corridor routing. Intervening territories on any Phase 1 corridor are inferred from service-territory geography rather than from routing analysis. No intervening utility should be named in external material before a corridor study.
Competitive development. RTO-sanctioned bilateral contracting between large loads and new generation now exists as a tariff channel for substantially the transaction this program proposes. The program’s answer — such channels match load to whatever gets built and do nothing about who bears construction risk on clean firm plant, which is the thing that has not been built — is available but must be argued rather than assumed in any external version.