Perfect World Project Energy Ledger
5.69M
EVs · 1.94% of fleet
As of Aug 2026
National energy & transportation tracker

Does the energy transition pencil out?

Mandates, credits, programs, rollbacks, wins, and failures — all of it checked against the only referee that matters: the grid's actual meters. Real registration counts, real generation data, real demand projections. No cheerleading in either direction.

The grid, 2024
4,308.6 TWh generated · 4,110 TWh consumed EIA final data. Gas 43% · nuclear 18% · coal 15% · wind 10.5% · hydro 5.6% · solar 5.1%
The fleet, 2025
5.69M battery EVs of 293.8M light-duty vehicles 1.94% of the fleet. Plus 1.77M plug-in hybrids and 13.9M conventional hybrids (AFDC registration counts)
The forecast
+224 GW summer peak demand by 2035 — a 24% jump NERC's Jan 2026 assessment, up 69% from what it projected just one year earlier. Mostly data centers — not EVs
The policy field
Federal credits dead · CA mandate in court · NEVI rebooting The $7,500 credit ended Sep 30, 2025. Quarterly EV sales fell 46% in one quarter, then began recovering
Why this exists

Both camps sell you a story. One says the transition is inevitable and nearly done; the other says it has already failed. The meters say neither. This ledger tracks every mandate, subsidy, rollback, and real-world result in one place — and maps the claims against what the grid can actually deliver.

Your state

Where do you land?

The national numbers hide 51 very different grids. Pull up yours.

Location is resolved in your browser and never stored — same privacy rules as the rest of the hub.

Visualizer · supply vs demand

The country's electricity ledger, 2014 → 2035

A decade of nearly flat demand ended in 2021. EIA's final 2024 number — 4,110 TWh of end use — was an all-time record, and the projection band below is what 1.0% to 2.4% annual growth does to it by 2035.

Band: 1.0%/yr (pre-2020 trend) to 2.4%/yr (a NERC-era trajectory — NERC's 10-year peak forecast implies ~2.2%/yr). Dashed reference: 2024 net generation, 4,308.6 TWh. Generation has to grow with demand; the gap between the lines is today's entire operating margin plus net imports and losses.

Table view
Visualizer · what actually moves the needle

Demand drivers, in the same units

The surprise in the data: today's EV fleet is a rounding error on the grid. The demand shock actually arriving this decade is data centers. The demand shock that would arrive if every mandate succeeded is the full-fleet number.

TWh per year. EV rows derived from registration counts × 11,500 mi/yr × 0.30 kWh/mi wall-to-wheels; data center rows from LBNL's December 2024 report (176 TWh in 2023, 325–580 TWh projected by 2028). 2024 US generation: 4,308.6 TWh.

Table view
Visualizer · where the 4,309 TWh comes from

What the grid runs on, 2024

Any new demand — EVs, data centers, heat pumps — gets served by this mix at the margin. Gas is the swing fuel in most regions; every state's own mix is in its state dossier.

EIA 2024 final, utility-scale net generation. "Other" bundles biomass, geothermal, petroleum, other gases, and pumped-storage losses.

Table view
The short version

What the numbers say, plainly

  • EVs are not straining the grid today. 5.69M EVs draw roughly 20 TWh a year — about 0.5% of national demand. Anyone blaming current grid stress on EVs is pointing at the wrong load: data centers used ~176 TWh in 2023 and are the main driver of NERC's record forecast.
  • But the stated goals don't pencil without building generation. Electrifying all 293.8M light-duty vehicles needs ~1,014 TWh/yr — a 25% increase over 2024 demand, equivalent to ~129 new 1-GW plants running at 90% capacity factor. That was always a multi-decade build, and it now competes with data centers for every megawatt.
  • The policy engine that was supposed to drive adoption has been dismantled. Federal purchase credits died Sep 30, 2025; the California mandate architecture (12 states + DC) was revoked by Congress and is in court; NEVI built ~384 charging ports from a $5B program before its 2026 reboot.
  • The market kept moving anyway — slower. Post-credit sales crashed 46% then recovered 14% by Q2 2026. The fleet keeps growing every quarter. What changed is the slope, and the slope is what every 2030 goal was priced on.
The back of the envelope, made honest

Pencil it out yourself

Pick a level of fleet electrification and the driving assumptions. The calculator converts it to terawatt-hours and holds it up against the 2024 grid. Defaults are the national averages.

Assumptions
Fleet electrified100%
293.8M of 293.8M light-duty vehicles
Miles per vehicle per year11,500
FHWA national average ≈ 11,500
Efficiency (wall-to-wheels)30 kWh/100mi
Includes ~10% charging losses. Sedans ~26, crossovers ~30, pickups ~40+
TWh per year of new electricity demand
Increase over 2024 US demand (4,110 TWh)
1-GW power plants at 90% capacity factor
Billion gallons of gasoline displaced per year
Visualizer · scenario vs the 2024 grid

Where your scenario lands

The price at the plug

Cost per mile: your state's electrons vs gasoline

The macro numbers are one question; your driveway is another. This uses your state's actual 2024 average power price.

Gasoline price$3.20/gal
Comparison car's fuel economy28 mpg
New-car fleet average ≈ 28; a Prius ≈ 52
EV cost per mile
Gas cost per mile
Per year at 11,500 miles (negative = EV costs more)
Break-even gas price — below this, gas wins on fuel cost

Reading the result

Three things to keep in frame

  • Efficiency is the transition's real argument. Full electrification adds ~1,000 TWh of electricity but retires ~4,500 TWh of gasoline heat — EVs turn roughly 3 units of fuel into 1 unit of needed electricity. The physics pencils; the question is the build-out.
  • Energy is the easy half; power is the hard half. A TWh spread over overnight charging is cheap. The same TWh at 6pm on a July peak is a new power plant. Managed charging is the difference, and almost no state mandates it.
  • The queue is the bottleneck. ~2,600 GW of proposed generation sits in interconnection queues — more than double the existing grid — waiting years for studies. The constraint isn't ambition; it's throughput.
Mandates · credits · programs

The policy field, tracked

Everything that was supposed to drive the transition, with its current legal status. Updated as of August 2026.

The big one

The mandate architecture is in court

Every state EV mandate in America runs through California's Clean Air Act waivers. In June 2025 Congress revoked them by CRA resolution — the first time the Congressional Review Act was ever used on a state waiver. California and ten states sued, arguing waivers aren't "rules" the CRA can touch. In March 2026 DOJ counter-sued CARB, which is enforcing anyway under emergency authority. In June 2026 EPA sent four more waivers to Congress for the same treatment. Until the courts rule, every "adopted" badge below carries an asterisk.

Federal timeline

What Washington built, then unbuilt

    State mandates

    Advanced Clean Cars II — 12 states + DC

    ACC II requires rising ZEV sales shares from model year 2027, reaching 100% in 2035 (some states adopted only through 2032). Current posture of each:

    Sales-share requirements apply to manufacturers, not buyers. "Paused" = governor or agency suspended enforcement. All are clouded by the waiver litigation above.

    Programs & credits

    The incentive machinery

      The watchlist

      Decisions coming to a docket near you

      Dates that will move this ledger. Check back as they land.

        Results, not intentions

        Scoreboard

        What the transition has actually delivered, what it has actually broken, and what the market did when the subsidies stopped.

        Visualizer · the subsidy cliff

        Quarterly US EV sales, 2024 → 2026

        The $7,500 credit died September 30, 2025. Buyers rushed the deadline, then the floor dropped: −46% in one quarter. Two quarters later the market found a new, lower baseline and began climbing again.

        Battery-electric only, thousands of vehicles. 2024–Q2 2025 are Cox Automotive estimates; Q3 2025 – Q2 2026 from post-expiry sales reporting. Q2 2026 was +14.2% vs Q1 but still −20.5% vs Q2 2025.

        Table view
        Visualizer · the long slope

        Annual US EV sales, 2016 → 2025

        Ten years of the adoption curve every 2030 target was priced on. 2025 is the first year the line didn't go up — the Q3 buying panic wasn't enough to offset the post-credit collapse.

        Battery-electric only, thousands. Argonne National Laboratory / Cox Automotive estimates; 2025 is the sum of quarterly reporting above.

        Table view
        Delivered

        Accomplishments on the board

        Broken or busted

        Failures on the board

        All 50 states + DC

        State by state

        EV adoption against each state's own grid: registrations (AFDC 2025), net generation and retail sales (EIA 2024), average power price, and mandate status. "Headroom" = generation minus in-state sales; negative means the state imports power.

        Head to head

        Compare two states

        StateEVsSharePortsEVs/portGen TWhUse TWhHeadroom¢/kWh

        Tap any row for the full state dossier — generation mix, charging network, and the state's own electrification math. Share = battery EVs ÷ all registered light-duty vehicles. Ports = public charging ports (AFDC, Aug 2026). DC's headroom is what an all-imports grid looks like.

        Show your work

        Sources & methods

        Every number on this page traces to one of these. Retrieved August 12, 2026 unless noted.

        Primary data
        Vehicle registrations by state (2025) — DOE Alternative Fuels Data Center, light-duty registration counts (Experian snapshot).
        afdc.energy.gov/vehicle-registration
        Net generation by state, 1990–2024 — EIA state historical tables, released Sept 2025 (2024 final).
        eia.gov/electricity/data/state/annual_generation_state.xls
        Retail sales, end use & prices, 2024 — EIA Electric Power Annual, tables 2.2, 2.8, 2.10.
        eia.gov/electricity/annual
        Demand forecast — NERC 2025 Long-Term Reliability Assessment (published Jan 2026): +224 GW summer / +246 GW winter peak growth by 2035; five regions at high reliability risk by 2030 (MISO, PJM, ERCOT, WECC-Basin, WECC-Northwest).
        nerc.com → Long-Term Reliability Assessments
        Data center load — LBNL, "United States Data Center Energy Usage Report" (Dec 2024): 176 TWh in 2023 (4.4% of US electricity), 325–580 TWh projected by 2028.
        eta.lbl.gov
        Charging network by state (Aug 2026) — DOE Alternative Fuels Data Center station counts: 255,073 public EV charging ports at 81,621 locations, 74,349 DC fast, national.
        afdc.energy.gov/stations/states
        Quarterly EV sales — Cox Automotive / Kelley Blue Book EV sales reports (2024–2025 estimates); post-expiry quarters from industry sales reporting: Q3'25 437,487 · Q4'25 234,171 · Q1'26 216,399 · Q2'26 247,226.
        Policy status — CRA resolutions signed June 2025; California et al. v. United States (waiver litigation, 2025–); US v. CARB (March 2026); EPA referral of four additional waivers to Congress (June 12, 2026); OBBBA credit sunsets (30D/25E/45W ended Sep 30 2025; 30C ended June 30 2026); FHWA NEVI revised guidance (Aug 2025) and FY26 apportionment ($885M, Oct 2025).
        Methods & caveats
        • EV energy math: vehicles × miles/yr × kWh/mi (wall-to-wheels, charging losses included). Defaults 11,500 mi and 0.30 kWh/mi. Today's fleet: 5.689M EVs → ≈20 TWh/yr (~0.5% of demand). PHEVs excluded from the electricity math.
        • Projection band is illustrative: 1.0%/yr ≈ the pre-2020 flat era's upper edge; 2.4%/yr ≈ the growth NERC's peak forecast implies. NERC forecasts peak power (GW); we show annual energy (TWh). The two move together but are not the same thing.
        • Headroom caveat: generation minus retail sales is a rough surplus measure. It ignores imports/exports timing, line losses, and — critically — capacity at peak. A state can be energy-positive and still capacity-short on a July evening.
        • Registration counts are a snapshot, not sales; they lag scrappage and migration. Mandate statuses are as reported through Aug 2026 and change monthly — treat badges as "last known posture," not legal advice.
        • 1-GW-plant equivalent = 8.76 TWh/yr × 90% capacity factor = 7.88 TWh/yr per plant. It's a sizing intuition, not a siting plan.
        • Cost-per-mile tool uses each state's EIA 2024 all-sector average price as a home-rate proxy (residential rates usually run a bit higher) and a user-set gasoline price. Public DC fast charging typically costs 3–4× home rates, so heavy road-trippers should discount the savings figure. Annual sales history: Argonne National Laboratory LDV sales series + Cox Automotive.
        Related modules

        Data Centers — the loads actually driving the forecast · Superfund — the last industrial build-out's cleanup bill · Polygraph — claims vs record · Hub home