Power to the People

The chapters and NEON's research

Strength of each link: ● strong ◐ medium ○ weak. Links are a first reading and have not yet been checked by the researchers.

Part 1: Abundance

1. 1. Ending a Million Years of Fire

   Humanity is ending a million-year era of burning nature for energy, because solar and wind tap renewable flows instead of finite stocks. In 1974, Dutch schoolchildren were shown a poster of the Earth as a burning candle and taught that civilization was doomed. Fuel was scarce for most of history, shaping zero-sum thinking. Rome, Haiti and England depleted their wood; England turned to coal out of necessity. Fossil fuels are ancient biomass, drawn down about a million times faster than nature replaces them. The 1972 Limits to Growth model underestimated feedback loops that slow depletion and, more fundamentally, feedback loops that make solutions grow exponentially. Using computer models to find such reinforcing loops has been the author's own work. Eco-anxiety and climate grief follow from beliefs about scarcity, and those beliefs are incomplete: the sun delivers more energy each hour than civilization uses in a year. Solar and wind work best in a system of small local units called holons, where energy is harder to hoard and distribution is fairer.

   No direct link with NEON research found.

   • ○ weak Work package: WP10 Integral models
     The chapter states the book's thesis that models with reinforcing feedback loops reveal fast-growing solutions. NEON's integral model planned such endogenous feedback loops; here the idea is stated, not worked out.
   • ○ weak Work package: WP9 Societal values
     The chapter asserts that decentralised solar and wind make energy harder to hoard and its distribution fairer. WP9 studied how the costs and benefits of the transition are distributed; the chapter does not argue the point.
2. 2. The Stone Age Will Not End For Lack of Stones

   The fossil fuel era is ending through a slow economic squeeze, not through depletion. Scarce, concentrated resources give power to whoever controls them: oil prompted the 1953 coup in Iran, sustains autocracies and funded Russia's war. That power is now waning. New Dutch coal plants worth over €6 billion, commissioned in 2015, lost most of their book value within a year. Britain closed its last coal plant in 2024 as wind replaced coal, Germany's auction for coal closures was oversubscribed, and in India solar tariffs fall below the fuel cost of existing coal plants. Shale gas and shale oil producers destroyed capital, and between 2010 and 2019 the five oil supermajors paid shareholders $216 billion more than their free cash flow, covering the gap with debt and asset sales. Falling energy return on investment raises extraction costs, while for renewables it is no longer a constraint. The industry behaves like a slow liquidation, and its reserves are becoming stranded assets.

   No direct link with NEON research found.

   • ○ weak Work package: WP10 Integral models
     Shared topic only: the cost of renewables against fossil fuels, which the integral-model scenarios compare. NEON did not study fossil company finances, stranded assets or energy return on investment.
3. 3. The Nuclear Mirage

   Nuclear power is disqualified as a solution, although it is low-carbon, impressive and safer than fossil fuel. Four problems decide the verdict. Annual nuclear growth peaked in 1985 and has been close to zero in recent decades, while solar and wind keep growing. New nuclear costs around three times as much per kilowatt-hour as wind and utility-scale solar, projects typically overrun their budgets by a factor of two to three, and newer plants cost more than older ones, a negative learning curve. Civilian programmes raise the risk of proliferation: replacing all fossil energy would take about 20,000 reactors, and reprocessing and thorium make weapons material easier to obtain. Finally, nuclear must run at high output all year, which fits poorly with a grid of rooftop solar, batteries, electric vehicles and local energy communities that balance themselves at house, neighbourhood and region level, called holons. Every euro spent on nuclear is a euro not spent on the technologies that are winning.

   • ◐ medium Work package: WP10 Integral models
     One cost argument against nuclear assumes that storage adds little to the cost of a well-designed renewable system; WP10 set out to compute system costs including storage. NEON did not study nuclear power, and its plan cites work (Jenkins et al. 2018) finding nuclear or CCS almost necessary when seasonal storage is modelled poorly.
   • ◐ medium Work package: WP3 Demand, storage and markets
     The fourth argument against nuclear rests on houses and neighbourhoods balancing supply and demand locally, the subject of WP3's work on local market models, demand-side management and storage.
   • ○ weak Work package: WP2 DC electricity grids
     Shared topic: grid congestion under distributed supply, which WP2 quantified in scenarios.
4. 4. More Energy Than We’ll Ever Need

   Energy for human civilization is abundant: the sun delivers about 173,000 terawatts to the Earth while humanity uses about 18, roughly ten thousand times less. Wind carries about a hundred times human demand. Best-in-class panels convert over a quarter of sunlight into electricity, tandem cells in labs exceed 34%, and the theoretical limit for multi-junction designs lies above 50%. Fossil energy reaches us by a long detour through ancient photosynthesis, extraction, transport, refining and combustion, and power plants lose 60 to 70 percent of it as waste heat. A solar panel turns photons into electrons directly, needs no fuel and produces no waste stream, and this simplicity gives it a steep learning curve. The author's fully covered roof in the Netherlands yields twice his household's annual need, including heat pump and electric car. Solar and wind are flows: one household's panels leave a neighbour's sunlight untouched, so energy stops being zero-sum. Building the infrastructure still takes materials, labour and capital and will be contested.

   No direct link with NEON research found.

   • ○ weak Work package: WP1b Building integrated PV
     The opening figure's solar potential assumes panels on all roofs or the whole built environment, the surfaces building-integrated PV targets. The chapter's argument holds by orders of magnitude without it.
   • ○ weak Work package: WP3 Demand, storage and markets
     The author's roof yields twice his household's annual need; WP3 studied the hourly and seasonal mismatch that an annual figure leaves out.
5. 5. The Rising Sun

   International Energy Agency forecasts have underestimated solar power for two decades, missing its steady exponential growth. After fifteen years in internet consultancy, the author saw that solar was small but growing exponentially at a predictable rate; people at Shell and the IEA were not convinced. Reverse-engineering the yearly panel production implied by the World Energy Outlook scenarios from 2002 to 2025 produced a graph in which actual annual sales rise steeply while each new IEA projection stays almost flat. An early hypothesis was that the IEA model treats solar panels like coal, as something consumed in the year it is produced, and cannot handle a product that yields energy for 30 years. The graph spread on Twitter, in PV Magazine and The Economist, and in a TED talk by Al Gore. A sidebar gives the arithmetic: solar grew about 36% a year, from 0.02% of global electricity in 2005 to nearly 9% today, with an exponential fit of R² 97%.

   • ◐ medium Work package: WP10 Integral models
     The chapter blames forecasts that cannot represent the exponential growth of a long-lived technology. WP10 planned to model adoption and learning as endogenous feedback loops and to challenge incorrect narratives.
   • ○ weak Work package: WP12b NEON annual outlook
     NEON's annual outlook was meant as an alternative to existing forecasts; the chapter's critique of the IEA outlook shares that purpose but does not mention it.
6. 6. The Learning Curve Crashing the Fossil Fuel Party

   Energy forecasts should always use learning curves: every doubling of cumulative production cuts cost by a fixed percentage. Solar modules lose about 20% of their cost per doubling, from about $106 per watt in 1975 to about $0.20 now, and solar electricity fell from $80 to about $0.04 per kilowatt-hour. Wind learns at about 12%, batteries at about 28%, and power converters follow the semiconductor curve; the four reinforce each other. Fossil fuels barely learn because they are mature and material-intensive, and nuclear learns negatively because each plant is a megaproject. Solar uses over 100 times less material per unit of energy than fossil fuels and is made in small repeated units, closer to chip manufacturing than to mining. The IEA missed this through paradigm blindness, linear intuition about exponential growth and institutional inertia. Policy such as the German Energiewende helped early scale-up, but the learning itself comes from the physics of the technologies.

   • ◐ medium Work package: WP10 Integral models
     Endogenous learning curves for solar, wind, storage and electric vehicles are written into the WP10 plan; the chapter makes learning curves the basis of every forecast.
   • ○ weak Work package: WP1b Building integrated PV
     The NEON plan used the same material-per-kWh argument to motivate thin-film PV; WP1b studied device materials, not learning rates.
   • ○ weak Work package: WP5 Medium voltage fast charging
     Power converters are one of the chapter's four learning technologies; WP5 designed medium-voltage converters for chargers but did not measure learning rates.
7. 7. Our Sunny Future

   Solar and wind will push fossil fuels out of the energy system in the coming decades, driven by a cycle in which growth lowers costs and lower costs drive further growth. The projection to 2100 rests on S-curves fitted by hand to 20 years of data. For solar, a curve with a maximum of 68% of energy, a midpoint in June 2039 and a steepness of 0.2 fits with an R² of 99.8% and an NRMSE of 2.0%. Wind fits almost as well, with a maximum of 24%, and remains the cheapest backup to solar in windy countries with dark winters such as the Netherlands. Hydro loses market share, nuclear declines for economic and political reasons, and biomass is phased out as a fuel. Aviation and shipping keep oil longest and switch to eFuels and electricity after 2050. A sidebar explains the six-step fitting method and why NRMSE takes priority over R². The direction is presented as reliable, the exact percentages as uncertain.

   • ◐ medium Work package: WP10 Integral models
     The chapter's projection to 2100 is a hand-fitted global S-curve. WP10 was meant to produce bottom-up scenarios of solar and wind adoption that could support or contradict it; the chapter uses no NEON model output.
   • ○ weak Work package: WP12b NEON annual outlook
     The chapter is in effect an energy outlook to 2100; NEON's annual outlook covered the Dutch transition.
8. 8. Doing More With Less

   Renewable energy needs orders of magnitude less raw material than fossil fuel, and the minerals it uses stay in the system instead of being burned. One kilowatt-hour from coal burns about 500 grams of coal; a complete solar system with battery and converter needs about 2 grams of material over its lifetime, and 0.1 grams of newly mined material with 95% recycling. Degrowth through austerity is rejected as politically unworkable; dematerialisation is the reduction that works. Identified lithium resources would last almost 500 years at current production, and sodium-ion batteries offer an alternative. LFP batteries avoid cobalt, reluctance motors avoid rare earths, and aluminium replaces copper, which remains the tightest constraint in the 2030s. Lead-acid batteries reach 99% recycling because the material is valuable, products are uniform, metals can be remelted and collection exists; the first three conditions hold for panels and lithium batteries. A global renewable system needs 6 to 7 billion tonnes of material in total, against 15 billion tonnes of fossil fuel each year.

   No direct link with NEON research found.

   • ○ weak Work package: WP1a Airborne wind energy
     The NEON plan motivated airborne wind with the same grams-per-kWh argument (coal 300 g, wind 10 g, airborne wind 2 g). The chapter gives a conventional solar-plus-battery system the same 2 g/kWh and does not mention airborne wind.
   • ○ weak Work package: WP1b Building integrated PV
     The plan expected thin-film building-integrated PV to use a fraction of a gram of material per kWh; the chapter's figures are for standard silicon modules.
   • ○ weak Work package: WP4 Electric cars and trucks
     Shared topic: rare-earth-free motors and lighter vehicle wiring. WP4 designed electric powertrains but did not study material scarcity.
9. 9. No New Petrostates

   Solar and wind cannot produce new petrostates, because the conditions that let fossil fuel exporters coerce others do not apply to them. Fossil infrastructure, with its few fields, refineries and pipelines, is easy to control and to use as a weapon. Oil money has financed coups, autocracies and Russia's war in Ukraine. Sunlight falls on every country, installed panels and batteries keep working without a continuous flow of supplies, and panel margins of 10 to 15% leave nothing for a cartel to capture. China makes about 80% of panels and 77% of battery cells, yet factories can move while oil deposits cannot, and no country has a geological monopoly on silicon, aluminium, lithium, iron or phosphorus. Forced labour in Xinjiang calls for traceability and import rules and is a separate issue from supply security. Under a Chinese export ban, Europe's 338 gigawatts of solar would keep generating and China would lose most. Rooftop solar with storage makes households hard to coerce.

   No direct link with NEON research found.

   • ○ weak Work package: WP3 Demand, storage and markets
     Shared topic only: household and community generation with storage, which WP3 studied as prosumers and local storage, not as protection against coercion.
10. 10. Everything Must Go Somewhere

    Burning fuel kills seven to nine million people a year through air pollution, and replacing combustion with solar and wind removes most of that harm and the 76 percent of greenhouse gas emissions that come from energy. The title refers to Barry Commoner's second law of ecology: everything must go somewhere. About half of the deaths come from outdoor pollution by coal plants and diesel traffic, half from indoor cooking smoke; gas cooking alone kills an estimated 19,000 people a year in the US and 40,000 in Europe. Solar emits 20 to 50 times less CO₂ per kilowatt-hour than coal over its life cycle and pays back its manufacturing emissions in one to two years. Coal causes 24.6 deaths per terawatt-hour, solar 0.02. Once energy is clean, agriculture remains the main polluter through nitrogen and phosphorus runoff, nitrous oxide and pesticides. Meat replacements, precision fertilisation, pixel cropping and weeding robots can cut that load. The Montreal Protocol shows that countries can agree to stop crossing a planetary boundary.

    No direct link with NEON research found.

    • ○ weak Work package: WP4 Electric cars and trucks
      Shared topic only: diesel trucks and buses as a source of urban air pollution. WP4 designed electric trucks and buses but did not study health effects.
11. 11. Land: The Ultimate Scarce Resource

    Solar panels need little land: supplying every person on Earth with the average American's energy use, about 75 kilowatt-hours a day, takes about 75 square metres of panels per person. That equals the roof area per person in developed countries, 10% of the built environment and 0.4% of Earth's land surface. The Netherlands has little sun and dense, energy-intensive industry, so wind leads there: about 70% of energy from turbines spread over roughly 10% of the Dutch North Sea, with foundations covering 0.01% of that area. For 90% of the world's population, rooftop solar with modest storage costs least. Vaclav Smil's energy-density argument is contradicted by four decades of markets choosing solar and wind over nuclear. Agriculture occupies 60% of habitable land, three quarters of it for livestock, and two beef burgers a year take more land than the panels for a person's entire energy use. Replacing animal protein with plant-based products and precision fermentation could return about two thirds of agricultural land to nature.

    • ◐ medium Work package: WP1b Building integrated PV
      The land arithmetic relies on solar panels fitting on roofs and other built surfaces. WP1b developed building-integrated PV for roofs, facades and windows; it did not measure the available surface.
    • ◐ medium Work package: WP10 Integral models
      The claims that a wind-led mix is cheapest for the Netherlands and rooftop solar with modest storage for most of the world are cost-optimisation questions the NEON plan put to its integral model.
    • ○ weak Work package: WP1a Airborne wind energy
      The chapter sees no land-use problem for wind in the North Sea; the NEON plan argued for airborne wind partly because it needs half the area and draws fewer objections to sight and noise.
    • ○ weak Work package: WP7 The human factor
      The chapter treats objections to solar farms as factual errors; WP7 studied acceptance and objection as behaviour to be researched.
12. 12. A Mote of Dust Suspended in a Sunbeam

    Abundant energy for every person is compatible with all nine planetary boundaries, provided solar, wind and batteries are deployed at the rates their learning curves allow and agriculture is reformed. Jay Forrester's system dynamics and the 1972 report Limits to Growth gave a formal way to model human pressure on the planet. Kate Raworth's doughnut adds a social floor to the ecological ceiling, and the planetary boundaries of Rockström and Richardson give that ceiling coordinates. In the bathtub metaphor, most boundaries concern what flows into the Earth's ecosystem. Combustion causes about 75% of the drivers of climate change, ocean acidification and aerosol loading, which solar and wind address. Agriculture drives the rest, plus the nitrogen and phosphorus cycles, novel entities, land-system change, freshwater use and biosphere integrity. Meat replacements would cut these impacts by about two thirds, and precision agriculture and weeding robots reduce runoff and pesticides further. Five technologies suffice for machines: solar panels, wind turbines, batteries, digital power converters, and hydrogen or synthetic fuels for hard cases.

    No direct link with NEON research found.

    • ○ weak Work package: WP10 Integral models
      Shared topic: system models of the transition. The chapter uses system dynamics and the bathtub metaphor; WP10 chose agent-based simulation and did not cover planetary boundaries.
    • ○ weak Work package: WP3 Demand, storage and markets
      Batteries and hydrogen are among the chapter's five technologies; WP3 compared storage options but did not test whether five technologies suffice.

Part 2: Resilience

1. 13. The Prize

   A clean, quiet, electric world that serves the people at the bottom is possible, and in 2026 it does not yet exist. A daydream set in Amsterdam in 2050 shows clean air, clear canals, quiet electric vehicles, and a Kenyan farmer whose eight square metres of solar panels run robots, a fridge and a heat-pump grain dryer. At an imagined conference at LUT University, veteran analysts report that emissions peaked in the late 2020s, warming heads for about 1.3 degrees, solar dominates energy supply, hydrogen stayed in niches, and batteries took over road transport, most shipping and much of aviation. In his own talk, the narrator says his main contribution may be the idea of organising the grid as nested, semi-autonomous cells, or holons. Back in 2026, 770 million people lack electricity, cooking smoke kills about 3.2 million people a year and outdoor air pollution 4.2 million, while the grid remains centralised, fragile and increasingly expensive. The rest of the book is about closing that gap.

   No direct link with NEON research found.

   • ○ weak Work package: WP4 Electric cars and trucks
     The 2050 vision has electric cars, trucks and ships replacing combustion engines; WP4 designed such powertrains. The chapter describes outcomes, not research.
   • ○ weak Work package: WP10 Integral models
     The thread that mainstream projections underestimate solar, electric vehicles and batteries matches WP10's aim; the chapter gives no model result.
   • ○ weak Work package: WP3 Demand, storage and markets
     Neighbourhoods balancing their own supply and demand, and vehicle-to-grid, appear in passing; WP3 studied both.
   • ○ weak Work package: WP2 DC electricity grids
     Battery-electric ships charged at offshore wind farms touch WP2's work on electric ship power systems with Damen and Royal IHC.
   • ○ weak Researcher: Maarten Steinbuch
     Named as the author's friend and fellow advocate of electric cars and trucks since 2009, without reference to NEON, of which he was the main applicant.
   • ○ weak Partner: lut-university
     The imagined 2050 conference takes place at LUT University, with Christian Breyer, whom the NEON plan lists as an international modelling partner of the integral model.
2. 14. Storing solar and wind

   Batteries will handle most storage for solar and wind, with e-fuels for seasonal storage and long flights and supercapacitors for short bursts. Converting electricity to hydrogen and back loses 65% of the energy, against 10% for a lithium iron phosphate battery, so fuels suit storage that is used only once or twice a year. Pack prices fell 93% between 2010 and 2025, to 108 dollars per kWh; a paper co-written with Floor Alkemade puts learning at about 28% per doubling, which points to below 25 dollars by 2050. Over the same fifteen years transformer prices roughly quadrupled while stationary batteries became 15 times cheaper, yet grid operators treat batteries as a nuisance. A study by the author's company for the Dutch government found that smart charging of electric vehicles plus a 5 kWh battery per house could roughly halve the peak houses cause on a neighbourhood transformer. Distributed storage also makes the grid more resilient, because each part can keep running when others fail.

   • ◐ medium Work package: WP3 Demand, storage and markets
     The claim that smart charging plus a 5 kWh battery per house roughly halves the peak on a neighbourhood transformer is the mechanism WP3 studied: demand-side management, smart charging and neighbourhood storage.
   • ◐ medium Work package: WP10 Integral models
     The 2050 battery price rests on a learning rate of about 28% per doubling; endogenous learning curves for storage are part of the WP10 plan.
   • ◐ medium Researcher: Floor Alkemade
     The chapter takes its battery learning rate from a paper by the author and Floor Alkemade (TU/e), who leads the integral-model work package in the NEON plan.
   • ◐ medium Partner: Zenmo Simulations
     The smart-charging and home-battery result comes from a study by the author's company Zenmo for the Dutch government. Zenmo was NEON's modelling partner; the study itself was not NEON research.
   • ○ weak Work package: WP2 DC electricity grids
     Shared topic: local storage as an alternative to grid reinforcement, which WP2 was to quantify in scenarios.
3. 15. Grid-Forming Inverters

   Grid-forming inverters let solar panels and batteries stabilise the grid themselves, and what holds them back is institutional. The Iberian blackout of 28 April 2025 followed from missing spinning reserve and from grid codes that forced inverters to disconnect when voltage left the permitted band. A grid-forming inverter sets its own voltage and frequency, supports the wider grid during disturbances, keeps a local microgrid running when the wider grid fails, and makes black-start automatic. It synchronises through droop control, the same coupled-oscillation principle that keeps large generators in step. Its hardware is largely identical to that of grid-following inverters; the difference lies in firmware and certification. Faster transistors and higher efficiency halve residential inverter weight about every fourteen years, and if the price trend holds, wholesale prices fall from about 42 dollars per kW in 2025 to about 1 dollar in 2050. Modular multilevel converters bring inverters into high-voltage DC links, while the medium-voltage transformer is expected to remain.

   • ◐ medium Work package: WP5 Medium voltage fast charging
     WP5 developed control for grids dominated by power converters and fed its results into grid codes, the subject of this chapter. The chapter expects the medium-voltage transformer to remain, while WP5 aimed at chargers that connect to 10 kV without one.
   • ○ weak Work package: WP2 DC electricity grids
     Shared topic: power electronics and islanded microgrids. The chapter argues for AC grid-forming inverters, while WP2 worked on DC microgrids.
   • ○ weak Work package: WP8a Regulation
     The chapter blames grid codes for the Iberian blackout; WP8a analysed regulatory barriers to decentralised supply but not inverter grid codes.
4. 16. Computers: The Grid’s Distributed Brain

   The grid will get a distributed brain: every appliance will carry a small computer that plans its own energy use, and together they will make the grid resilient and cost-effective. Computing has become 30 to 40 percent cheaper every year for sixty years, so within ten years a 10-dollar computer can sit on every inverter, battery and solar array. Following the hierarchy of learning rates, smart use of EVs, batteries and other flexible devices comes first, grid-forming inverters in every building second, more batteries and vehicle-to-grid third, solar and wind fourth, and grid reinforcement last. Home energy management systems, EV charge controllers and heat pumps already have the hardware; what is missing is a system that uses it. Grid operators and their regulators hold this back: they set uniform prices years in advance, barely monitor grid use and do not share data. Telecom operators resisted the internet in the same way until cable companies threatened them.

   • ◐ medium Work package: WP3 Demand, storage and markets
     The chapter's first priority, smart use of electric vehicles, batteries and flexible appliances through home energy management, is WP3's subject: demand-side management, smart charging and aggregators.
   • ◐ medium Work package: WP8a Regulation
     The chapter's barrier is institutional: uniform tariffs fixed years ahead, little monitoring and no data sharing by grid operators. WP8a analysed the regulation of grid operators' roles and access to energy data.
   • ○ weak Work package: WP8b Standardization
     Making existing appliances cooperate requires interoperability standards, WP8b's subject; the chapter does not discuss standards.
   • ○ weak Work package: WP5 Medium voltage fast charging
     Distributed control of many devices for grid stability is shared with WP5's control algorithms for chargers.
5. 17. The Grid We Inherited

   The Dutch grid should be split into decentralised, semi-autonomous microgrids, because its centralised design is costly, ties the country to fossil fuels and is fragile. The grid operators' FIEN26 estimate adds about €5,000 per household to reinforcement costs already estimated at about €35,000, and lost growth from connection queues costs at least €1,000 per household a year. Zenmo's models indicate that local coordination, flexible demand and batteries could deliver the same transition at roughly half the cost. The grid was built for costly fuel, no storage, no computers, passive consumers and a copper-plate promise; all of these have changed, and the Dutch grid still has on average 80% of its physical capacity left. The Irish potato famine, the NotPetya attack and the Iberian blackout show how uniform, tightly coupled systems fail, while life survives through nested semi-autonomous parts. Drawing on years at Alliander, the author concludes that grid operators, the regulator ACM and the ministry cannot change course from within, so citizens must push for it.

   • ◐ medium Work package: WP10 Integral models
     The chapter's central number, the same transition at roughly half the cost through local coordination, flexible demand and batteries, comes from Zenmo's agent-based models, the kind of model WP10 set out to build. The chapter does not attribute it to NEON.
   • ◐ medium Work package: WP2 DC electricity grids
     WP2 quantified bottlenecks and investments in Dutch low- and medium-voltage grids with grid operator data; the chapter's critique of reinforcement plans and its claim of unused grid capacity rely on that mechanism.
   • ◐ medium Work package: WP3 Demand, storage and markets
     The proposed architecture relies on local coordination and flexible demand from households, WP3's subject.
   • ◐ medium Work package: WP8a Regulation
     The chapter explains the lack of course correction by gridlock between grid operators, the regulator ACM and the ministry. WP8a analysed institutional barriers and regulation in the move to decentralised supply.
   • ◐ medium Work package: WP7 The human factor
     The chapter says people reject being told when to charge but provide flexibility when asked; the NEON plan's human-factor work starts from the same reactance mechanism.
   • ◐ medium Partner: Zenmo Simulations
     The chapter's main evidence is Zenmo's modelling and a Zenmo report for the Dutch government on local autonomy against grid congestion, which is not NEON research.
   • ◐ medium Partner: Alliander (Liander)
     The author draws on his years at Alliander, the parent company of NEON partner Liander, to argue that grid operators cannot change course from within.
   • ○ weak Work package: WP9 Societal values
     The affordability of grid costs for poor households is mentioned but not developed; WP9 studied the distribution of transition costs.
6. 18. Holons: The Universal Building Blocks of Life

   Complex systems built from nested, semi-autonomous modules, called holons, are both more efficient and more resilient, and the electricity grid should be organised that way. Herbert Simon and Albert Ando proved in 1961 that in near-decomposable systems, with strong links inside modules and weak links between them, local problems settle locally; the current grid, where everything couples to everything, is the opposite. Arthur Koestler named such a unit, a whole that is also part of a larger whole, a holon. The pattern recurs in evolution, the internet's end-to-end design, subsidiarity in EU law and Thorbecke's Dutch constitution, and Elinor Ostrom's polycentric commons. The grid's topology is already nested but lacks buffers, grid-forming inverters, computers and the will to share power with end users. An energy holon is a semi-autonomous microgrid that balances itself locally and offers a simple interface to the level above. The idea dates from Yoichi Kaya in 1987 and reached Dutch policy through Mart van Bracht in 2020.

   No direct link with NEON research found.

   • ○ weak Work package: WP3 Demand, storage and markets
     Citizen energy communities and local flexibility overlap with WP3's local market models; the chapter argues from theory, not from local-market evidence.
   • ○ weak Work package: WP2 DC electricity grids
     The energy holon is defined as a semi-autonomous microgrid; WP2 studied DC microgrids, which is a different question.
   • ○ weak Work package: WP9 Societal values
     Subsidiarity and polycentric governance share WP9's governance topic; no NEON work package tested holonic grids.
   • ○ weak Work package: WP10 Integral models
     NEON's integral model nests neighbourhoods in regions and the country, a modelling device rather than a test of holonic grids.
7. 19. The Grid Doesn’t Need a Commander

   Nested local markets can coordinate electricity grids better than a central operator. The grid in 2026 resembles telecom around 1993: control sits in the centre and devices at the edges are treated as dumb, so their intelligence and storage go unused. Hayek's knowledge argument and the zero-duality-gap result show that a well-structured multi-level market reaches the same optimum as a perfectly informed planner, while localising failures and reacting faster. Wholesale markets coordinate energy over time, but grid capacity has no locational or temporal price, which drives up Dutch grid costs. PowerMatcher, distribution locational marginal pricing and field trials such as Olympic Peninsula, with about 15% lower peak load, show the missing market works. Fairness concerns can be handled by pooling within neighbourhoods and by splitting tariffs into fixed and variable parts. The Dutch minister nevertheless rejected locational tariffs in 2024. Grid operators will not reform from inside; pressure will come from solar, batteries, EVs and aggregators, as cable companies once forced KPN onto the internet.

   • ◐ medium Work package: WP3 Demand, storage and markets
     The chapter's argument rests on local transactive markets in which electric vehicles, heat pumps and batteries bid against grid limits; WP3 designed and compared local market models and aggregator roles.
   • ◐ medium Work package: WP8a Regulation
     The chapter's diagnosis is regulatory: the Dutch refusal of locational network tariffs. WP8a analysed the regulation of distribution grids, demand response and network tariffs.
   • ◐ medium Work package: WP9 Societal values
     The chapter has to answer the fairness objection that locational pricing hurts households without flexibility; the distribution of costs per household type was WP9's research subject.
   • ○ weak Work package: WP8b Standardization
     The telecom analogy and the call for simple open protocols share WP8b's topic of open standards.
   • ○ weak Work package: WP2 DC electricity grids
     Shared topic: Dutch grid congestion and reinforcement costs.
8. 20. Your Home Is a Holon Waiting to be Kissed Awake

   Buildings and neighbourhoods equipped with batteries, grid-forming inverters, solar panels and EVs can become semi-autonomous energy holons, and this is already happening worldwide. Pay-as-you-go solar kits reach off-grid households in Africa and Asia, and households in Pakistan and South Africa install solar and batteries to escape expensive or unreliable grids. In the Netherlands, companies short of grid capacity use batteries for peak shaving, and a shared container battery would cost a neighbourhood of 200 households about €10 a month each, under one tenth of planned reinforcement costs. A Zenmo simulation for the Dutch government found that rewarding avoidance of the worst peaks cut a street's highest annual peak by over 50%, against 15% when devices only followed existing energy markets. Pilots from Hoogkerk to the Philippines show the same pattern. Holons can nest up to the national level, and rooftops, facades, parking and 10% of the North Sea could cover an electrified Netherlands apart from bunker fuels. Regulation and the resistance of incumbent grid operators hold this back.

   • ◐ medium Work package: WP3 Demand, storage and markets
     The Zenmo street simulation, with devices reacting to markets against rewards for avoiding the worst peaks, studies the mechanism WP3 covers: demand-side management, smart charging and neighbourhood storage.
   • ◐ medium Work package: WP10 Integral models
     The chapter's key number comes from a local agent-based simulation of a Dutch street, the kind of model WP10 planned; the national electrification balance is also an integral-model question.
   • ◐ medium Work package: WP2 DC electricity grids
     The claim that a neighbourhood battery costs less than one tenth of planned reinforcement relies on grid investment figures of the kind WP2 quantified.
   • ◐ medium Work package: WP8a Regulation
     The chapter ends on missing permission: Dutch rules block energy communities and shared batteries. WP8a analysed regulation for prosumers, storage and energy communities.
   • ◐ medium Work package: WP1b Building integrated PV
     The national balance assumes solar on half of all roofs, 30% of facades and most parking areas with 30%-efficient panels, the building surfaces WP1b worked on.
   • ◐ medium Partner: Zenmo Simulations
     The key number comes from a Zenmo simulation for the Dutch government, not from NEON research.
   • ○ weak Work package: WP1a Airborne wind energy
     North Sea supply is counted with conventional turbines; airborne wind plays no role.

Part 3: Flourishing

1. 21. Human flourishing

   An energy system meant to produce human flourishing must be designed for humans as they are: cooperative, fair-minded and capable of self-government. Re-examined evidence, from the Tongan castaways to the Stanford Prison Experiment and the Blitz, shows people pulling together under stress, which undercuts the assumption that local energy communities will collapse into free-riding. Fairness experiments show that people accept inequality earned through effort, redistribute inequality caused by luck, and, deciding behind a veil, choose a maximised average with a guaranteed floor. Universal entitlements keep more political support than aid targeted at the poor. Because most decisions are unconscious and poverty taxes bandwidth, the energy system should run mostly automatically. Warm cooperation works in groups of up to about 150 people; larger scales need impersonal rules, so institutions should act as neutral referees in Ostrom's sense. Self-determination theory supplies the design targets: autonomy, competence and relatedness. An energy system needs users in charge to deliver lasting flourishing; the author says he cannot yet fully prove this.

   • ◐ medium Work package: WP7 The human factor
     The chapter's design rules rely on behavioural mechanisms: people cooperate when they can see each other, information changes little and automation works. WP7 studied what makes people support and adopt energy technology.
   • ◐ medium Work package: WP9 Societal values
     The fairness requirements, a guaranteed floor, earned inequality and universal entitlements instead of means testing, are energy-justice questions of the kind WP9 studied.
   • ○ weak Work package: WP3 Demand, storage and markets
     Profit sharing from a community battery touches WP3's local market models.
2. 22. Economics, the Broken Compass

   Standard economics favours the rich, the present and fossil fuels, and this bias helps explain why the world still burns fossil fuels and why the internet was captured. Economics separates growing the pie from dividing it, so it does not register extraction such as Apple's 30% app-store fee as a cost. It underestimates externalities: the IMF puts fossil-fuel damage at $7 trillion a year, about 7% of world output. Discounting shrinks future damage; at the 2% rate economists are converging on, the social cost of carbon rises from $60 to $190 per tonne, and weighting rich and poor lives equally would push it above $1000. Since Lionel Robbins, comparing wellbeing between people has been ruled unscientific, which hides that a euro does more good for the poor. The homo economicus model ignores flourishing, and companies bound to shareholder value act it out. Models also ignored learning curves, so the IEA kept forecasting the end of solar growth and clean energy was under-built.

   • ◐ medium Work package: WP10 Integral models
     Only the section on learning curves matches NEON: the chapter blames models that ignore learning for under-forecasting solar, batteries and electric vehicles, and WP10 planned endogenous learning curves.
   • ○ weak Work package: WP12b NEON annual outlook
     The chapter attacks the IEA's outlook; NEON planned its own annual outlook.
   • ○ weak Work package: WP8b Standardization
     The warning that the energy system could become a platform with a toll booth shares WP8b's aim of avoiding platform monopoly.
3. 23. How to Bake a Doughnut

   Economics should first lift everyone above an energy floor, then keep the system below the planetary ceiling, and optimise for flourishing only inside that space. A decent life needs only about 15 GJ of final energy per person per year, and sun, wind and batteries can supply far more within planetary boundaries. Flourishing becomes a target function: material capability is maximised, autonomy, competence and relatedness act as constraints, and labelled dials set the weight of inequality, future generations and rights. This makes energy justice checkable in a model run. Field experiments show that automation, more than information, makes households respond to prices. Equilibrium models cannot represent the energy transition as a complex adaptive system, as a 2017 paper by Hoekstra with Maarten Steinbuch and Geert Verbong argued. Agent-based models with learning curves can, and they forecast well: a simple 2017 model predicted 1.1 million Dutch EVs in 2025 against 1.05 million actual. Such models can expose chokepoints and lobbying, and open code lets domain experts audit them.

   • ◐ medium Work package: WP10 Integral models
     The chapter's programme of agent-based, learning-curve-driven, open models matches the WP10 plan, and the 2017 paper it cites is the one the NEON plan names as the basis of its integral model. Most of the chapter's evidence predates NEON, and the author writes that most of the models he envisions are not yet built.
   • ◐ medium Work package: WP9 Societal values
     The inequality dial, which scores tariffs and subsidies by their effect on the bottom of the income distribution, needs the welfare effects per household type that WP9 set out to measure.
   • ◐ medium Researcher: Maarten Steinbuch
     Co-author of the 2017 paper on modelling the energy transition as a complex adaptive system that the chapter cites; the NEON plan names the same paper as the basis of its integral model. Steinbuch was NEON's main applicant.
   • ◐ medium Researcher: Geert Verbong
     Co-author of the same 2017 paper, cited for its 343 references; Verbong leads the work package on demand, storage and markets in the NEON plan.
   • ○ weak Work package: WP12b NEON annual outlook
     The 2017 forecast of Dutch electric vehicle numbers is the kind of figure an annual outlook tracks.
4. 24. Preventing extraction

   Extraction in energy systems is a product of system design: decent people inside an extractive architecture produce extractive outcomes. Useful technologies follow a predictable slide in eight steps, through dependence, chokepoints and property to finance-driven rent, rewritten rules, weakened counterpower and hardening. Prevention is cheap only in steps two to four. Historical cases include Samuel Insull, the East India Company, Standard Oil and the Phoebus light-bulb cartel. The electricity grid is essential, monopolistic and opaque, an ideal substrate for extraction. British network owners expected 9.2% real returns where the regulator judged 5.25-5.75% normal. Dutch grid operators earn a return on capital in the ground, so copper pays better than flexibility, and their cable and transformer data stay secret. Dutch net metering ends without a dividend or published arithmetic of who pays. Loss aversion and system justification make people defend extraction once it exists, so fairness has to be designed in from the start. EU rules such as the single charging port show that reversal is possible.

   • ◐ medium Work package: WP8a Regulation
     The Dutch half of the argument is regulatory: an allowed return on grid assets that favours copper over flexibility, the regulator ACM, net metering and energy communities in the new Energy Act. WP8a analysed such regulatory barriers.
   • ◐ medium Work package: WP9 Societal values
     Who pays for the end of net metering, and energy poverty as a symptom of extraction, are distribution questions WP9 studied.
   • ◐ medium Work package: WP7 The human factor
     The chapter explains why extraction defends itself through loss aversion, status-quo bias and the way carbon-price revenue is returned; WP7 studied loss aversion and acceptance.
   • ◐ medium Researcher: Derk Loorbach
     The chapter cites Kemp, Loorbach and Rotmans (2007) for the finding that Dutch transition practice gave incumbents too central a role. Loorbach leads the societal-values work package in the NEON plan; the paper predates NEON.
   • ◐ medium Researcher: Jan Rotmans
     Co-author of the same 2007 paper on transition management cited in the chapter; Rotmans is a co-applicant of NEON's societal-values work package. The paper predates NEON.
   • ○ weak Work package: WP8b Standardization
     Standards that keep new entrants out are mentioned; the full argument follows in chapter 27.
   • ○ weak Work package: WP3 Demand, storage and markets
     The claim that storage and aggregators could undercut grid reinforcement if grid data were open touches WP3's comparison of flexibility options.
5. 25. Power to the People

   Users keep power over the energy system only through exit and voice, and both have to be built on purpose. Independence comes down to five requirements: no chokepoint users cannot route around, switching providers as easily as telecom operators, data that travels with the user, flexibility value that lands with the household, and the right to choose and fire whoever trades on its behalf. Telephone history shows easy exit was forced on incumbents. Data portability works only as live, standardised access, as in Open Banking and the new Dutch data hub Het Normo, which grid operators may control in practice. Cloud-tethered devices such as the JuiceBox charger, which lost its smart functions in 2024, reflect a business model, since time-critical control already runs locally. Decisions belong at the lowest competent level, with an overrule button the system obeys. Long-horizon choices need paid, representative deliberation with binding output. Unbundling, interoperability and public grid data prevent capture. In Schönau, residents bought their local grid in 1997; their cooperative now serves 170,000 customers.

   • ◐ medium Work package: WP3 Demand, storage and markets
     The argument that flexibility value should land with the household, which can choose and fire its aggregator, relies on how aggregators and local markets work, WP3's subject.
   • ◐ medium Work package: WP8a Regulation
     Data rights, switching rules and citizen energy communities are regulation for prosumers, smart meters and data protection, all topics in WP8a's plan.
   • ◐ medium Work package: WP7 The human factor
     The overrule button as a condition for accepting automated control, and deliberation as a route to accepting long-term trade-offs, rely on acceptance and reactance mechanisms that WP7 studied.
   • ○ weak Work package: WP8b Standardization
     Standardised data interfaces as a defence against capture share WP8b's topic.
   • ○ weak Partner: ElaadNL
     ElaadNL is named as the place where the 'overrule button' got its Dutch name.
6. 26. Radical simplicity through holons

   Complexity gives extractors the opacity of secrecy while staying legal, so the energy system should stay close to physical reality and add abstractions only where they simplify. Any layer whose deletion leaves the electricity flowing only supports someone's income. Electricity markets stack capacity markets, scarcity pricing and financial hedges on the spot market, and each layer opens room for gaming: the GreenHat default left $179 million to other PJM members. Single-price bidding zones plus redispatch invite inc-dec gaming, and pricing by location removes that gap. Once the cost of learning and applying rules is counted, simple rules beat complex ones. Held complexity is acceptable when it offers defaults, full data for experts, one override button, one overview screen and one-sentence explanations. Energy holons balance themselves and present a thin interface upward: a price, a limit, an offer. Such interfaces must be mandated, as the EU did with USB-C. A grandma test, certified by a laboratory label, checks that a household can override, go offline, refuse data sharing and leave.

   • ◐ medium Work package: WP10 Integral models
     The chapter claims that holons plus agent-based models close to physical reality make corrective market layers unnecessary. WP10's agent-based integral model could test that claim; the chapter cites no result.
   • ◐ medium Work package: WP3 Demand, storage and markets
     Holons with local storage and a thin interface, a price in and an offer out, depend on local market models and demand-side management, WP3's subject.
   • ◐ medium Work package: WP8b Standardization
     The interfaces between holon levels must be thin, standardised and mandated, as the EU did with USB-C; open standards and interoperability are WP8b's subject.
   • ○ weak Work package: WP7 The human factor
     Defaults, an override button and a 'grandma test' concern acceptance, WP7's topic.
7. 27. Sunlight is the best disinfectant

   Data, standards and source code in the energy system should be 'open unless', with the burden of proof on whoever wants secrecy. Energy software costs less than a thousandth of the hardware it steers, yet it holds the decisions, and rent migrates to the decision layer. Transmission operators have had to publish their data since 2015; distribution grid operators still withhold data on cables, transformers and congestion. Personal consumption data stays private. Open protocols beat walled gardens, as the Web did. ElaadNL's charging protocol OCPP, given away in 2009, became international standard IEC 63584. Two habits remain harmful: paywalled standards and open protocols anchored in one company's cloud, such as plug-and-charge certificates that allow tracking of drivers. Open-source dispatch software lets anyone verify what acts for the user and is a security requirement, as the xz case shows. Closed systems mainly serve investors. Grid operators and battery makers fund open energy software as a complement, as in LF Energy and Alliander's forecasting software, and the EU can mandate open interfaces.

   • ◐ medium Work package: WP8b Standardization
     The chapter's evidence is standards practice (OCPP, S2, ISO 15118, paywalled standards, central certificate anchors), WP8b's subject. It goes further than the plan by asking for free-to-read standards and open-source dispatch software.
   • ◐ medium Partner: ElaadNL
     ElaadNL's charging protocol OCPP, given away in 2009 and now standard IEC 63584, is the chapter's main example of an open protocol that won.
   • ◐ medium Partner: Alliander (Liander)
     Alliander's open load-forecasting software is cited as a grid operator funding open software, while distribution grid operators are criticised for withholding grid data.
   • ○ weak Work package: WP8a Regulation
     Open grid data against private consumption data touches WP8a's work on data access and data protection.
   • ○ weak Work package: WP10 Integral models
     The chapter links open data to open models; its 'open unless' rule would apply to NEON's own models.
8. 28. Universal access to energy

   Universal energy access is affordable, and households without a grid are now easiest to serve. 666 million people have no grid connection, 2.1 billion cook on open fires, and 9.2% of Europeans cannot keep their home warm. Off-grid families already pay more for kerosene than a solar kit would cost: three panels, an inverter and a 5 kWh battery come to about $700 per family, or $5 per person per year. Delivery needs competition, open reference designs, cheap finance and efficient appliances people want. A remote kill switch is acceptable as loan collateral but should expire by law at payoff or after five years. For poor households on the grid, means-tested schemes fail because the neediest rarely apply. The alternative is a basic energy income: a free kilowatt-hour block on every connection, automatic, funded outside the bill, below typical use and visible on every bill, with throttling instead of disconnection. Universal access by 2030 could cost under $45 billion a year, two and a half days of fossil fuel subsidies.

   • ◐ medium Work package: WP9 Societal values
     Only the section on poor households connected to the grid matches NEON: energy poverty and the distribution of costs per household type were WP9 subjects. The off-grid half of the chapter has no NEON counterpart.
   • ○ weak Work package: WP7 The human factor
     Defaults, visible benefits and dignity in design are behavioural mechanisms; WP7 studied acceptance of technology.
   • ○ weak Work package: WP8a Regulation
     Protection against disconnection and current-limiting meters touch WP8a's consumer-protection topic.
9. 29. Conclusion

   The concluding chapter is being rewritten; the current version of the book has only its heading.

   No direct link with NEON research found.