Loading reports … Fetching the current index of long-form research.
Reports
Loading reports … Fetching the current index of long-form research.
China is setting the clean-power cost curve. America is setting the hydrocarbon supply curve.
China and the United States each built an energy production machine in the same decade, each larger than any third country has assembled, and each is now limited by the layer above production. The production advantages are durable and are not in dispute. The next decade of advantage is decided in the last mile: for China, whether the grid absorbs what it builds, measured by the share of available wind and solar it discards; for America, whether realized exports keep its liquefaction fleet full.
China is setting the clean-power cost curve, and America is setting the hydrocarbon supply curve. Both booms arrived in the same decade, and both are large enough to move world prices and supply chains. The useful question is what each can deliver, and where the path from production to delivered power meets the most resistance.
China added more than 430 gigawatts of wind and solar capacity in 2025, up 22 percent from the year before, taking the fleet to 1.84 billion kilowatts.1 Wind and solar reached 47.3 percent of China's power capacity and supplied 22 percent of its electricity.2 In the same year U.S. total energy output reached a record 107 quadrillion British thermal units, with dry natural gas at 39 trillion cubic feet and crude oil at 13.6 million barrels a day.3 The United States is the world's largest producer of oil and gas combined.4
The booms grew from different memories of scarcity. China industrialized at historic speed while importing rising volumes of oil and gas, burning immense quantities of coal, and managing air pollution as a political problem. Renewable power fit that agenda because Chinese firms could manufacture it, Chinese banks could finance it, and Chinese developers could install it, which made the decision strategic before it was environmental. The United States learned its lesson from import dependence and oil shocks, and for decades assumed that domestic production was in permanent decline. Shale reversed that assumption: in 2019 the country became a net exporter of total energy for the first time since 1952.5
Exhibit 1. Two scarcity memories built two production machines that both matter globally
Milestones · China and the United States · 2005–2025
| 2005 | 2010 | 2015 | 2020 | 2025 | |
|---|---|---|---|---|---|
| China | Renewable Energy Law | World’s largest wind market | World’s largest solar market | Dual-carbon pledge | Solar additions 315 GW; 430 GW wind and solar added |
| United States | Barnett shale gas takeoff | Tight oil boom (Bakken, Eagle Ford) | Crude export ban lifted (first LNG from Sabine Pass, Feb 2016) | Net exporter (since 2019) | Record 107 quads produced |
Source: China National Energy Administration; China State Council; U.S. Energy Information Administration; Cheniere Energy; International Energy Agency; Power Curve analysis. Each entry is dated to the event, and the Sabine Pass cargo is carried under 2015 with its own February 2016 date because it closes the export-ban decision. Data through December 2025.
The histories explain the durability. China's buildout is embedded in state utilities, equipment makers, grid companies, provincial governments, and the five-year plans. America's is embedded in private mineral rights, independent producers, oilfield-service firms, pipeline and LNG developers, and the politics of Texas, Pennsylvania, North Dakota, and the Gulf Coast. Each has constituencies that want it to continue, so an importer that plans on either boom slowing soon is planning on the weakest assumption available.
The two machines produce different things, and the fair comparison sets one annual flow against another. Exhibit 2 indexes China's annual wind and solar generation and America's annual oil and gas production, both measured in terawatt-hours, to 2015. Between 2015 and 2024 China's wind and solar output rose from 225 to 1,836 terawatt-hours, about eightfold, while U.S. oil and gas output rose from 13,993 to 20,307 terawatt-hours, or 45 percent.6 The Chinese line measures tempo and the American line measures scale. In 2024 U.S. oil and gas carried about eleven times the energy of China's wind and solar output, a gap that narrows, without closing, once each unit of electricity is credited with the fuel a thermal plant would burn to make it.
Capacity figures differ by compiler, and the difference is a matter of basis. IEA estimates put China's 2025 renewable additions near 500 gigawatts, with nearly 370 gigawatts of solar and 117 gigawatts of wind, on a broader methodology than the National Energy Administration's grid-connection series behind the 430 gigawatt figure.7 Every figure in this report is given on the basis its publisher states.
Both countries made energy production repeatable and modular. In China, polysilicon, wafers, cells, modules, inverters, batteries, construction crews, grid companies, and local governments sit inside one expansion loop, reinforced by procurement targets and export competition that keep firms moving down cost curves. In America, drilling targets, lateral length, pressure pumping, sand and water handling, and basin infrastructure turned geology into a production system with measurable learning rates. Neither machine is easy to copy. Europe can build offshore wind and regulate demand, and it lacks China's supply chain and cost position. Importers can buy crude and LNG, and they lack U.S. mineral ownership, service depth, and pipeline network. Once production is repeatable, the binding constraint moves to absorption, grids, pipelines, storage, dispatch, and market design.
China's buildout is manufactured energy: the industrial system that built its export economy, applied to power generation. Beijing subsidized demand for wind and solar while building the supply chain around it, so its clean-energy complex deploys capacity with the tempo of a manufacturing sector. Renewable generation across all sources, hydro included, reached about 4 trillion kilowatt-hours in 2025, more than the European Union consumes.8
The buildout serves several aims at once: domestic energy security, lower import exposure at the margin, cleaner air, exportable industrial advantage, and demand for Chinese manufacturers. Because deployment is tied to manufacturing, Chinese overbuild lowers prices worldwide. When module capacity outruns demand, module prices fall everywhere, and when battery firms scale, storage costs fall. IRENA counted 692 gigawatts of renewable additions worldwide in 2025 and the IEA, on a broader basis, about 800; on either count China sits at the center.910
The scale gap facing competitors is large. China holds about 80 percent of world manufacturing capacity across the solar PV chain, and its 2025 module output of 574.5 gigawatts was 83 percent of the global total.11 The gap is widest at processing, the layer between mining and manufacturing: lithium, cobalt, graphite, and rare-earth refining sit mainly in China, which built the capacity when demand was small and kept it running through commodity cycles when other countries exited.12 A country that reshores module assembly without reshoring processing replaces one dependence with another, and the Inflation Reduction Act and Europe's Net-Zero Industry Act both face that sequencing problem.
The consequence is a changed default in every clean-power procurement. A government that wants to electrify quickly must justify rejecting the cheapest equipment. One that wants supply-chain security pays a premium in equipment prices, timelines, and subsidy budgets. One that wants both needs a third supplier with comparable cost and scale, and none exists at the level China commands. The cost curve can also become a policy instrument, and the watchlist carries any formal export restriction on modules, cells, or inverters.
The buildout has also slowed. First-quarter 2026 solar additions fell to 41.4 gigawatts from 59.7 a year earlier, the first first-quarter decline in five years, and installations over January to April fell by about half.13 Most of that is a policy artifact, since the June 1, 2025 move from fixed tariffs to auction pricing pulled a year of projects into the first half of 2025.14 It matters for this report because a buildout that slows while curtailment rises points to absorption as the binding constraint, which Part IV measures.
America's position rests on geology, private mineral rights, capital, technology, and a competitive service sector. U.S. total energy production reached 107 quadrillion Btu in 2025, the highest on record; marketed gas averaged 118.5 billion cubic feet a day, crude reached 13.6 million barrels a day, and natural gas plant liquids also set a record.15 On a common crude-plus-condensate basis the United States produced 13.6 million barrels a day in 2025, against 9.9 for Russia and 9.6 for Saudi Arabia.16
The Permian is the center of the system. EIA's March 2026 refinement put Permian shale and tight output at 6.0 million barrels a day of crude and 22.2 billion cubic feet a day of dry gas in December 2025.17 Counting conventional output, the basin supplied 23 percent of U.S. marketed gas in 2025 and about half the year's growth.18 Part of that step up is definitional, because the refinement added four plays to the basin's accounting and raised tight oil by about 0.2 million barrels a day and shale gas by about 0.8 billion cubic feet a day. At 6.6 million barrels a day of crude across 2025, the Permian alone would rank fourth on the world table, behind Russia, Saudi Arabia, and the rest of the United States.19
Shale must be reinvested continuously to stand still. A Midland Permian well peaks near 730 barrels a day and falls more than 70 percent by the end of its first year, and about 80 percent of a well's lifetime output arrives in its first two years.20 Capital discipline, service costs, pipeline limits, methane rules, and water constraints can all bend the production path. Output kept rising through the 2023 price retreat, which shows the system can grow through a falling price; it does not by itself separate a structural productivity gain from drilling committed during the 2022 spike.
U.S. LNG exports to Europe reached a record 10.3 billion cubic feet a day in 2025, up from 6.3 in 2024 and 68 percent of U.S. LNG volumes, and EIA expects net gas exports of 18.7 billion cubic feet a day in 2026 and 20.5 in 2027.21 Those cargoes move under private contracts, and they also back allied energy security as Europe cuts its Russian gas. The same structure ties U.S. gas prices more closely to world demand, and it creates the American last mile: the pipelines and liquefaction trains that must stay full as the fleet grows.
Placed on one map, the two leaderships are domain-specific. China dominates clean-power additions and the United States dominates crude. Russia and Saudi Arabia remain hydrocarbon powers, and India is the fast-growing demand story, adding about 57 gigawatts of renewables in 2025 against the European Union's 85, up almost 60 percent in a year.10 No country commands both domains.
China's 47.3 percent share of capacity against 22 percent of generation is often read as a sign of a lagging grid.2 The gap is arithmetic. Wind and solar have lower capacity factors than thermal plant, so a gap of about that size would appear in a system with a perfect grid and zero curtailment. Coal remains the anchor: the IEA estimates that coal-fired generation fell to 55 percent of China's electricity in 2025 from 70 percent a decade earlier.22 The measure of whether the grid is keeping up is curtailment, the share of available wind and solar energy the system discards.
Absorption has several layers, and each is technical and political at once. Transmission must carry power from the interior and the north to the coast across provincial lines. Storage must smooth daily and seasonal swings. Power markets must pay for flexibility, coal plants must move from baseload to balancing, and provinces must accept flows that cut the revenue of local generators. Both measurable legs are moving at scale. State Grid invested a record of about 650 billion yuan, near 93 billion dollars, in 2025, and west-to-east transmission capacity stood at 340 gigawatts against a 2030 target of 420.23 China added 66.43 gigawatts and 189.48 gigawatt-hours of new-type storage in 2025, up 52 and 73 percent, taking the fleet to 144.7 gigawatts.24 The curtailment series shows whether they are moving fast enough, and since 2025 it shows they are not.
National curtailment fell from above 10 percent in 2016 to about 3 percent by 2024. In the first half of 2025 it reversed: wind curtailment reached 6.6 percent against 3.9 a year earlier, and solar 5.7 against 3.0. Reuters, carrying the same release, assigns the two national rates the other way round; both exceed 5 percent on either assignment.25 On one analyst's full-year estimate, wind and solar generation would have been about 13 percent higher without utilization losses, a discard worth about 17 billion dollars.26
The regulator's own case deserves a hearing. In June 2024 the National Energy Administration allowed provinces with strong wind and solar resources to relax their utilization target from the long-standing 95 percent to as low as 90 percent, a tolerated curtailment of up to 10 percent.27 The principle behind it is sound: the economically optimal level of curtailment is above zero, because spilling marginal interior wind for a few hours a year can cost less than the wires or storage that would capture it.28 On that reading a 6.6 percent wind rate is a priced choice. This report keeps 5 percent as its threshold because the discard is concentrated in the provinces the transmission program was designed to connect, where the planned wires are already sanctioned. Tibet curtailed 30.2 percent of its wind and 33.9 percent of its solar in the same half year, far past the last few percent that optimization theory covers.25
A single half-year could have caught a deadline. The June 2025 move to auction pricing pulled installations into the first half of 2025, and a surge into a deadline raises discard for a time while interconnection catches up. The record since then rules that reading out. Curtailment rose again across January and February 2026, to 9.2 percent for solar against 6.1 a year earlier and 8.5 for wind against 6.2, on Bloomberg's report of the monitoring centre's data.29 The first-half 2026 release then put wind utilization at 90.9 percent and solar at 91.4, curtailment of 9.1 and 8.6 percent, above the 2025 readings and close to the 10 percent the regulator now tolerates in its best-resourced provinces.30 A deadline surge works off within a year, and this discard rose.
A separate measurement path points the same way. In the first quarter of 2026 wind capacity grew 23 percent while the average wind capacity factor fell from 27 to 22 percent, and solar capacity grew 33 percent while its capacity factor fell 11 percent. Weather explains about half the wind fall and a quarter of the solar fall, which leaves rising curtailment to account for the rest.31 That analysis shares an author with the full-year estimate, so it confirms the direction from a different data set by the same hand. The independent corroboration is the regulator's half-year releases, carried by separate wires in 2025 and 2026.
If China solves absorption at scale, it gains cheaper marginal power, more room to electrify, and clean-tech leadership in system design as well as hardware. If it does not, it keeps a vast renewable fleet, a coal anchor, and mounting pressure to build grids, storage, and market reform faster than its planning system has moved.
America's constraint sits in the same layer. Shale output depends on acreage, drilling, completions, processing, water, takeaway, and capital aligning at once, and associated gas from oil wells can outrun pipeline capacity when drilling grows fast. The political objectives also pull apart. Low gasoline and gas prices, high LNG exports, producer employment, lower emissions, and power for data centers can coexist in some years and collide in others, and every additional LNG cargo supports allies and producers while raising domestic price exposure.
Energy power now depends on system yield. For China the chain runs from installation through transmission, storage, dispatch, pricing, and the coal transition to delivered electricity. For the United States it runs from the wellhead through gathering, processing, interstate pipelines, liquefaction, shipping, and regasification. The link that fails to scale with the hardware binds.
Exhibit 6. For each machine the binding bottleneck decides the strategic payoff
Engine, bottleneck, leverage, failure mode · three machines
| Engine | Bottleneck | Leverage | Failure mode | |
|---|---|---|---|---|
| China renewables | Manufactured capacity at scale | Absorption: grids, storage, dispatch | Cost-curve power, electrification | Overcapacity, no absorption |
| U.S. oil | Shale drilling productivity | Decline rates, capital discipline | Market weight, price resilience | Price cycles, field decline |
| U.S. gas | Record output, LNG buildout | Pipelines, terminals, permits | Allied supply, LNG diplomacy | Congestion, price pass-through |
Source: EIA; IEA; China National Energy Administration; IRENA; selected policy documents; Power Curve analysis. Analytic framework, assembled as of the June 2026 cutoff; the bottleneck column is the binding constraint. The underlying series behind every cell are the ones cited in Parts II to IV. Data through December 2025.
The American test is realized LNG exports against the nominal liquefaction capacity in service, read quarterly, with two consecutive quarters below 90 percent counting as failure. No published series carries that ratio quarterly, so this report grades the annual average and leaves the quarterly reading to the watchlist. U.S. LNG exports averaged 15.0 billion cubic feet a day across 2025 against the 15.4 billion cubic feet a day of nominal capacity EIA recorded in October, about 97 percent.32 That pass came on a fleet that had barely grown. Five projects under construction add 10.0 billion cubic feet a day of nominal capacity, and Golden Pass shipped its first cargo in April 2026.33 Keeping two thirds again of today's fleet full is the American last mile.
The constraint already shows as a price. Daily prices at the Waha hub, where Permian associated gas meets its takeaway, settled below zero 39 times in 2025 and 34 times in the first ten weeks of 2026, against a record 49 in 2024, and the hub averaged negative 37 cents per million Btu in 2026 to mid-March against 1.15 dollars across 2025.34 A hub that pays producers to stop is pipeline congestion priced in public. The volume behind it has no national compilation: flared and vented gas is published by operator, by state, and by pipeline on incompatible bases, so China's discard can be totaled in dollars and America's cannot. Sizing it is the first item on this series' research agenda.
Importers face a cost-security-emissions trilemma, a framing borrowed from the World Energy Council and the IEA. What this report adds is the supply side: in 2026 the three axes are three counterparties. Chinese equipment buys the lowest cost and fastest deployment against concentration in a chain China holds at about 80 percent. Domestic manufacturing buys industrial capacity against slower deployment and a premium paid through equipment prices and public budgets. U.S. LNG buys supply security under long-term contracts against emissions and exposure to world gas prices. Only the PV concentration has a tier-one series; no tier-one series measures LNG contract tenor or the cost premium of industrial policy, so neither is sized here.
Europe shows the trilemma most clearly. It wants clean power at competitive cost, industrial resilience, cheaper energy after the 2022 price shock, and less Russian exposure, and it is buying through all three channels at once, which is the rational response and an expensive one. Japan and Korea manage the security side with LNG portfolios and depend on global manufacturers for hardware. Southeast Asia deploys solar mostly on Chinese equipment and builds LNG import capacity as a reliability backstop.
The strongest objection is that this report has durability backwards. Transmission lines, batteries, pipelines, and liquefaction trains are capital projects with published costs and schedules of a few years, while a clean-tech manufacturing complex assembled over two decades, and a shale resource base with two decades of delineated drilling inventory, cannot be bought quickly at any price. On that reading the production layer is the durable advantage, the last mile is a lag, and the standings at the end of the decade look like the standings now.
Two points narrow the objection. The binding part of China's last mile is the part capital does not buy: interprovincial dispatch priority, provincial tax bases that depend on local coal plants, and a pricing rule that pays generators for energy and rewards no one for flexibility are political settlements, and they move on a political clock. And a flow constraint costs output every year it goes unsolved, so the stock advantage compounds only once the flow clears. The objection would have prevailed had curtailment fallen back under 5 percent as the 2025 grid and storage additions came on. The first-half 2026 release showed the reverse. The claim this report defends is therefore the narrower one: the production advantages on both sides are durable, and the marginal decade is decided in the last mile, where both systems still lose energy they have already paid for.
China must turn 1.84 billion kilowatts of renewable capacity into reliable electricity that displaces coal at scale, and America must turn record output into durable allied leverage without hollowing out the production base. On the record through the first half of 2026, America holds the delivery layer and China does not. China's test is failed on a discard that rose in 2025 and again in 2026, and that discard is the only one of the two losses anyone has totaled. America's test is passed on a liquefaction fleet that had barely grown, and its constraint already prices itself at Waha. The American lead is provisional. The next reading is EIA's export series against the capacity in service as the Golden Pass and Plaquemines trains come up, and two consecutive quarters below 90 percent would reverse it.
| Variable | What a move would mean | Reading | Falsifying trigger | Horizon |
|---|---|---|---|---|
| China wind and solar generation share | The absorption problem is being solved on the output side | 22% (2025) | Above 25% of national generation | 2027 NEA release |
| China wind curtailment | Capacity is built and then discarded; the constraint binds | 9.1% (H1 2026); 6.6% (H1 2025; 5.7% on Reuters) | Back under 5% across a full half-year | Full-year 2026 and H1 2027 NEA data |
| China solar curtailment | The same test on the faster-growing fleet | 8.6% (H1 2026); 5.7% (H1 2025; 6.6% on Reuters) | Back under 5% across a full half-year | Full-year 2026 and H1 2027 NEA data |
| China quarterly solar additions | The buildout slows before absorption catches up | 41.4 GW (Q1 2026, from 59.7) | A second consecutive quarter below the year before | Q2 2026 NEA data |
| U.S. LNG exports against nominal liquefaction capacity | America’s stated last-mile test; the trains are kept full | 15.0 of 15.4 Bcf/d, about 97% (2025) | Two consecutive quarters below 90% of nominal capacity in service | EIA monthly exports and liquefaction capacity |
| Permian crude output | The production treadmill stops outrunning decline | 6.0 mb/d shale and tight, Dec 2025 | Two consecutive quarters flat or falling | EIA monthly, 2026 |
| China coal share of generation | The coal transition is real, or the fleet stays the anchor | 55% (2025), from 70% in 2015 | Below 50% of national generation | IEA review, 2027 |
| China equipment export policy | The global clean-power cost curve is a Chinese policy instrument | Rare-earth magnet controls, 2025 | Formal restriction extended to modules, cells, or inverters | MOFCOM notices |
Each falsifying trigger is set in the direction of the move that would change the view. Readings are as of the June 2026 cutoff, except the two curtailment rows, which carry the first-half 2026 release.
Source: China National Energy Administration; Bloomberg; Reuters; U.S. Energy Information Administration; China Ministry of Commerce; Mercom. Data through June 2026.
Every figure traces to a named, dated source. Any reading that carries a verdict is cross-checked against a second independent source before use, with one exception: the January to February 2026 curtailment reading rests on one wire, and the first-half 2026 release now carries the 2026 leg of the China verdict. Installed capacity and delivered generation are reported separately and never summed. Exhibit 2 compares two annual flows in terawatt-hours from one compilation, Our World in Data's energy dataset; the U.S. leg is primary energy produced and the Chinese leg is electricity generated, so the two levels differ in end-use value and the exhibit is read for growth, with the level comparison in the text stated as raw energy content. Renewable-addition totals differ by compiler, so each is reported on the basis its publisher states; the same rule governs the Permian gas readings, where 19 percent is shale and tight formations and 23 percent is the whole basin. Curtailment readings are National Energy Administration operating figures, first-half for 2024 to 2026 and January to February for 2026, the only series published on a consistent national basis; the full-year utilization-loss figure is on a different basis and is labeled as such. Limitations: the provincial negative-price series are not comparable across the Chinese spot pilots and no number is carried for them; the full-year 2025 utilization-loss estimate is a single analyst's calculation and the capacity-factor analysis that corroborates it shares an author, so the independent corroboration rests on the NEA releases and the wires that carried them; the 97 percent liquefaction-utilization reading is this report's own ratio of a full-year export average to a point-in-time capacity, which flatters the layer; American flared and vented volumes are published on incompatible bases, so the American last mile is shown as a price and is not totaled; and the triggers in the watchlist are the author's own thresholds.
| Dataset | Source | Frequency | Coverage | Retrieved | Used in |
|---|---|---|---|---|---|
| Wind and solar capacity and generation shares | China State Council / National Energy Administration | Annual | 2025 | Feb 12, 2026 | At a glance, Parts I and IV |
| Renewable capacity additions by region | IEA Global Energy Review 2026 | Annual | 2025 | Apr 2026 | Exhibits 3 and 4, Table A2 |
| Renewable capacity additions, global | IRENA Renewable Capacity Statistics 2026 | Annual | 2025 | Apr 2026 | Part II |
| Wind and solar curtailment rates | National Energy Administration operating data, via Bloomberg and Reuters | Half-year and two-month | H1 2024 to H1 2026 | Oct 15, 2026 | Exhibit 5, Table A4 |
| Utilization losses and capacity factors | CREA (Myllyvirta and Qin), Jun 1, 2026; Carbon Brief (Myllyvirta), Jun 3, 2026 | One-off and quarterly | 2025 to Q1 2026 | Jun 3, 2026 | At a glance, Part IV, Table A4 |
| China transmission investment and storage additions | State Grid via China Daily; CNESA via pv magazine | Annual | 2025 | Feb 2, 2026 | Part IV |
| LNG exports, liquefaction capacity, Waha prices | U.S. EIA, Today in Energy; Reuters on daily Waha settlements | Annual and daily | 2019 to Apr 2026 | Apr 23, 2026 | Part V, watchlist |
| China electricity generation mix | IEA Global Energy Review 2026; Ember Global Electricity Review 2026 | Annual | 2015 to 2025 | Apr 2026 | Part IV |
| Wind and solar generation; oil and gas production | Our World in Data energy dataset (Ember; Energy Institute Statistical Review 2025) | Annual | 2015 to 2024 | Nov 3, 2026 | Exhibit 2, Table A1, Part I |
| U.S. total energy, dry gas, crude production | U.S. EIA, Today in Energy and Monthly Energy Review | Annual | 2025 | May 11, 2026 | Part III |
| U.S. marketed gas and Permian shares | U.S. EIA, Today in Energy | Annual | 2025 | Mar 13, 2026 | Part III |
| Permian tight oil and shale gas | U.S. EIA, Today in Energy | Monthly | December 2025 | Mar 17, 2026 | Part III, watchlist |
| Crude output by country | U.S. EIA International Energy Statistics | Annual | 2025 | 2026 | Exhibit 4, Table A3 |
| U.S. LNG exports and export forecast | U.S. EIA, Today in Energy | Annual | 2024 to 2027 | Apr 16, 2026 | Part III, watchlist |
| Solar PV supply-chain shares | IEA Solar PV Global Supply Chains; China PV Industry Association | One-off | 2025 | 2026 | Part II |
| Q1 2026 solar additions | National Energy Administration, via Mercom and PV Tech | Quarterly | Q1 2026 | Apr 24, 2026 | Part II, watchlist |
China is wind plus solar electricity generation. The U.S. series is oil plus gas production in primary energy terms. The 2024 levels are 997 TWh of wind and 839 TWh of solar for China, and 9,977 TWh of oil and 10,330 TWh of gas for the United States. Values are rounded to the nearest terawatt-hour; indexes are computed from the unrounded series.
| Year | China wind + solar, TWh | Index | U.S. oil + gas, TWh | Index |
|---|---|---|---|---|
| 2015 | 225 | 100 | 13,993 | 100 |
| 2016 | 307 | 137 | 13,587 | 97 |
| 2017 | 422 | 188 | 14,150 | 101 |
| 2018 | 543 | 241 | 16,202 | 116 |
| 2019 | 629 | 280 | 18,017 | 129 |
| 2020 | 728 | 323 | 17,547 | 125 |
| 2021 | 983 | 437 | 17,794 | 127 |
| 2022 | 1,190 | 529 | 18,817 | 134 |
| 2023 | 1,470 | 653 | 19,978 | 143 |
| 2024 | 1,836 | 816 | 20,307 | 145 |
Source: Our World in Data energy dataset, retrieved November 3, 2026, from Ember and the Energy Institute, Statistical Review of World Energy 2025. Data through December 2024.
Rest of world is computed as the residual: 800 less the four named regions. The IRENA series, which counts on a narrower basis, puts global additions at 692 GW.
| Region | 2025 additions, GW | Note |
|---|---|---|
| China | 500 | Nearly 370 GW solar PV and 117 GW wind |
| European Union | 85 | A record, about 10 percent above 2024 |
| India | 57 | Almost 50 GW solar; wind additions doubled above 6 GW |
| United States | 49 | Down about 10 percent on 2024 |
| Rest of world | 109 | Residual against the IEA global total |
| Global total | 800 | Up 16 percent on 2024 |
Source: IEA, Global Energy Review 2026, "Solar PV and wind"; IRENA, Renewable Capacity Statistics 2026. Data through December 2025.
Crude oil including lease condensate on one basis across the board. The Permian Basin alone produced 6.6 million barrels a day in 2025, which would place it fourth on a global crude table, behind Russia, Saudi Arabia and the rest of the United States.
| Country | Crude output, mb/d |
|---|---|
| United States | 13.6 |
| Russia | 9.9 |
| Saudi Arabia | 9.6 |
| Canada | 4.9 |
Source: U.S. EIA International Energy Statistics, 2025. Data through December 2025.
The first four rows compare H1 2024 with H1 2025; the next two compare January and February 2025 with the same months of 2026; the last two compare H1 2025, as reported on the wires, with the NEA's H1 2026 release, published as utilization rates of 90.9 percent for wind and 91.4 percent for solar. The national series ran above 10 percent at its 2016 peak and near 3 percent through 2024. Across full-year 2025, generation would have been about 13 percent higher without utilization losses, a discard valued at roughly $17 billion, on a basis not comparable with these rates and on a single analyst's calculation.
| Series | Year earlier | Latest |
|---|---|---|
| Wind, national, H1 | 3.9 | 6.6 |
| Solar, national, H1 | 3.0 | 5.7 |
| Wind, Tibet, H1 | 2.3 | 30.2 |
| Solar, Tibet, H1 | 5.1 | 33.9 |
| Wind, national, Jan–Feb | 6.2 | 8.5 |
| Solar, national, Jan–Feb | 6.1 | 9.2 |
| Wind, national, H1 2026 | 6.6 | 9.1 |
| Solar, national, H1 2026 | 5.7 | 8.6 |
Source: National Energy Administration operating data, via Bloomberg, August 5, 2025 and April 8, 2026, and on the Reuters wire, August 5, 2025; 36Kr and the Transition Economics Institute, October 15, 2026, for the first-half 2026 release; CREA, June 1, 2026, for the full-year figures. Data through June 2026.
Hart, Brendan. 2026. “The Two Energy Superpowers.” The Power Curve, October. https://brendanhart.com/reports/the-two-energy-superpowers.
