# The Great Inversion

## Executive summary of a series on rebuilding America's electricity institutions

**Problem.** Electricity has never been cheaper to produce, but American electric bills are rising faster than inflation and becoming a top-tier political issue.

**Cause.** Institutions built for scarce, central generation now govern abundant, distributed generation. Connection is still governed by substantial project-specific study and discretionary cost allocation, even as FERC has begun reforming the bulk interconnection process. Utilities also remain primarily compensated through capital-based ratemaking.

**Solution.** Five principles that finish the unbundling America began in the 1990s and stopped at the substation: a utility should earn on outcomes delivered, not capital deployed; the access decision should sit apart from the wires owner’s commercial interest; where and when a resource operates should affect what it earns; access should follow published rules, not case-by-case permission; and incremental system costs should sit with those who create them, socialized transparently where attribution fails. Each principle carries a menu of mechanisms with a stated, deliberately contestable preference, and every preferred mechanism already operates somewhere.

**Cost to the public.** Direct public expenditure is limited, but implementation is not costless. Transition costs include regulatory staff, data systems, metering, and IT and should be judged against the capital and operating spending the program avoids. Scenario aggregation puts that comparison near ten to one.

**Expected outcomes.** Faster and more predictable connection; better use of existing network capacity; lower long-run system cost; stronger cost attribution for incremental large loads; and potentially substantial avoided generation investment where flexible load can use existing headroom.

**First actions.** Governors call for a performance-regulation docket and for published interconnection metrics, by direction where commissions are appointed and by public case where they are elected. Commissioners open that docket and require machine-readable grid data. Legislators authorize performance regulation, a right to connect, and a standard value-stack tariff. FERC makes connect-and-manage the interconnection default where technically feasible and sets Order 2222 deadlines with consequences.

---

America has more generating capacity waiting in interconnection queues than it currently operates: 2,061 gigawatts against roughly 1,400 gigawatts in service, much of it financed, engineered, and sited. The typical project waits more than five years, and historically about 13 percent of that capacity ever connects. At the same time, data-center campuses that would pay a premium for power cannot buy it, gas turbine order books stretch toward five years, PJM's capacity charges have climbed from $2.2 billion to $16.4 billion a year, and retail electricity prices are rising fast enough to decide governors' races.

The grid no longer suffers from scarce electricity. It suffers from scarce permission.

**The inversion.** Every assumption the grid's institutions were built on has flipped. Generation was scarce and central; it is now cheap and everywhere. Storage was impossible; it is a manufactured product on a falling cost curve. Demand was passive; it is growing, flexible, and increasingly brings its own supply. Connection, once trivial, is the binding constraint on everything from AI capacity to reshored factories. The institutions absorbed every reform: process changed five times since 1978, and the payoff and the gate did not. Paper I traces how they formed (1907 compact, 1935 federal-state seam, 1990s wholesale unbundling that stopped at the substation). Paper II documents five institutional failures: ratemaking that pays utilities for capital rather than outcomes, access administered by the party being competed with, durable rules denied to the fastest-connecting resources, connection rationed by discretionary process, and a capacity auction that has hit administrative price caps in four consecutive auctions while attracting almost no new entry.

**The thesis.** The monopoly has narrowed. Wires remain a natural monopoly. Generation, storage, demand flexibility, interconnection, and grid data do not, yet the rules governing all of them still assume they do. The remedy completes the unbundling the United States started in the 1990s: separate the monopoly platform from everything contestable, this time at the distribution level, and make access to the network a matter of published rules at every voltage level.

### Unbundling unfinished: five principles to complete the job

Each principle is stated here by its preferred mechanism. Paper III carries the full menu of options under each, with a stance and an evidence tier for every mechanism evaluated, including the ones we set aside and the ones we reject.

- **Pay for performance.** Total expenditure (Totex) revenue caps and yardstick benchmarking on the British and Hawaiian models replace cost-of-service ratemaking. This reform comes first because it changes the utility's economic payoff before the program asks it to relinquish control of access, data, or procurement.
- **Separate the referee.** Interconnection, grid data, and flexibility markets move to an independent or ringfenced Distribution System Operator. The utility keeps the poles and the storm response and loses the gatekeeper's chair.
- **Pay resources for what they do for the grid.** Default dynamic retail rates with hedged options, posted locational offers wherever a resource can defer an upgrade, and Australian-style real-time operating envelopes in place of static interconnection caps.
- **Connect by rule.** A standardized pathway for small resources meeting published operating limits, a connect-and-manage pathway required of every RTO for bulk generation and large loads, reconductor-first requirements for transmission expansion, and fast-track access for hyperscalers in exchange for verified peak-hour curtailability.
- **Make new load fund reliability.** Forward obligations phase in while the centralized auction becomes a residual backstop. Data centers carry their own capacity costs, and generators finance against long-tenor contracts instead of an auction that swung tenfold in one cycle.

**The prize.** On the order of 100 GW of headroom from large-load flexibility on the existing system, which is roughly $100-220 billion of generation capex that need not be built, spanning EIA reference costs and observed 2025-26 market costs; a doubling of throughput on reconductored transmission corridors without new permitting; and months-scale access to the healthiest fraction of the 2,061 GW queue. The first tranche of this capacity program is already built.

**The consumer case.** The cheapest electricity ever generated is arriving alongside bills rising faster than inflation, in the fourth such episode in American history and the first without a cost story behind it, because institutions stand between cheap supply and the meter. Each principle aims at a line item: connection by rule lets $40-98/MWh generation compete into the supply line; obligations move data-center capacity costs off household bills and onto their drivers; and performance regulation, reconductor-first showings, and shared-savings deferrals discipline the delivery line, now the fastest-growing part of many bills. Hedged flat options, bill protection, and low-income rates are built in, and the commitment is measurable: real bill growth below matched peer states.

**The coalition.** Everyone outside the queue: hyperscalers, DER developers, independent generators (once Principle 5 is heard as "more PPAs"), industrial customers, and the institutional capital that finances all of them. The opposition concentrates where current rules produce earnings, which is why the sequencing leads with the reform that changes what earnings reward.

**The path, an illustrative sequence.** No act of Congress required. 2027: states open performance-based regulation dockets; FERC rulemakings on connect-and-manage and Order 2222 deadlines. 2028-29: totex orders take effect, automated interconnection goes live, large-load flexibility tariffs standardize. 2030-31: separated access administration where the escalation criteria are met, first hybrid obligation delivery year. 2032-35: the centralized auction recedes to a residual backstop where obligation coverage passes the reliability-and-cost test, and value-stack tariffs end the net-metering fights. Success is measurable annually: interconnection in days for small resources, offers inside 18 months for large ones, queue conversion above half for mature cohorts, and real bill growth below matched peer states.

The grid's first two settlements arrived only after visible, expensive failure, over incumbent opposition, and were later defended by their former opponents. Every precondition for the third is now met.

The end state: in 2035, a factory's 20 MW request returns in minutes with headroom, prices, a flexible alternative, and a firm date, and no one waits years for permission. The companion document What Success Looks Like describes that year in full, from the developer's morning to the regulator's dashboard.

*Three core papers: I. "How We Built Institutions for the Wrong Grid." II. "The Great Inversion: Cheap Supply, Scarce Connection." III. "Completing the Market." Two companion documents: "What Success Looks Like" and "Examples in the Evidence Base."*