Energy Sector Synthesis

Based on synthesis of 8 research explorations covering 921 concepts and 3,216 connections across the global energy system

---

The Big Picture in One Sentence

The world is actually building a lot of clean energy, costs are falling fast, and an emissions peak may be near — but the places, people, and industries that most need cheap clean energy are systematically cut off from it, while the infrastructure to move electricity around hasn’t kept up with the pace of building it.

---

How We Got Here: Eight Questions, One Story

These eight research explorations started as separate questions: How fast is the energy transition really moving? Is nuclear energy making a comeback? Are there enough minerals to build all these batteries and solar panels? Do carbon markets actually work? Can green hydrogen become a real fuel? How will climate change physically reshape the world? Who controls the global flow of energy? And what are companies like Shell, BP, and ExxonMobil actually doing?

On the surface, these look like eight different topics. But when you lay the findings side by side, the same handful of forces keep showing up — like characters in a story who turn out to be connected in ways you didn’t expect.

---

The Manufacturing Revolution Nobody Planned

Start with the most important fact: solar panels and batteries are getting cheaper at a pace that is genuinely historically unusual.

Every time the total amount of solar panels installed in the world doubles, the cost falls by roughly 20 to 25 percent. This is called Wright’s Law — the same pattern that made computer chips cheaper decade after decade. It has been running in solar for over thirty years, and it has not stopped.

The engine behind this is China. Chinese factories now produce the overwhelming majority of the world’s solar panels, wind turbines, and lithium-ion batteries. Chinese factories got very good at this very fast, and now they produce more than the world can immediately absorb — which pushes prices down further.

This matters because it was not a planned transition. No international agreement made solar cheap. No government target made batteries affordable. A manufacturing dynamic — driven largely by one country’s industrial policy and investment — did it.

The result is what researchers call the “renewables-coal crossover”: in a growing number of places, it is now cheaper to build a new solar plant than to run an existing coal plant. That crossover happened around 2025. It is a significant turning point, even if it doesn’t feel dramatic from the outside.

---

The Problem: Building Power Plants Is Easy. Plugging Them In Is Hard.

Here is the central frustration of the current moment: the cheap solar panels exist, but there are enormous lines to connect them to the electricity grid.

In the United States, there are hundreds of gigawatts of solar, wind, and battery projects waiting in “interconnection queues” — essentially permit lines to get connected to the grid. The wait can take years. A big part of the problem is that the rules for reviewing these connections (designed in an era when power plants were built once a century) haven’t adapted to an era when hundreds of projects are proposed every month.

Add to that: the physical wires to move electricity from sunny, windy places to where people actually live haven’t been built fast enough either. And local opposition to new power lines — the “not in my backyard” phenomenon — slows things further.

Think of it like this: a restaurant installed a kitchen that can cook meals ten times faster than before. But there’s only one waiter, the hallway is too narrow for more tables, and the neighbors complained about the new sign. The kitchen is ready. Everything else isn’t.

This is why falling costs don’t automatically translate into deployed clean energy everywhere. The constraint has shifted from cost to connection.

---

China’s Double Role: Most Important Helper and Most Significant Risk

China is at the center of the global energy story in a way that no single exploration fully captures.

On one hand, China’s manufacturing scale is the main reason clean energy is now affordable. Without Chinese factories, the energy transition would be decades slower and vastly more expensive.

On the other hand, China doesn’t just make the panels and batteries — it also processes most of the minerals that go into them. Lithium, cobalt, rare earth elements, and others are mined around the world, but the vast majority are refined (turned from raw ore into usable materials) in China. So even if a country mines its own lithium, it often sends it to China to be processed, then buys it back as battery-grade material.

This means China sits at two critical points simultaneously: the manufacturing of clean energy technology and the processing of the inputs that technology requires. No other actor comes close to this dual position.

This creates a geopolitical tension that only becomes visible when you look at the mineral exploration and the manufacturing exploration together. Either one alone understates the dependency. Combined, they reveal that much of the world’s clean energy supply chain runs through one country — which is both the reason costs are falling and the reason governments in the US, Europe, and elsewhere are scrambling to build alternative supply chains.

---

The New Player That Changed the Nuclear Math: AI Data Centers

Something unexpected showed up across multiple explorations: artificial intelligence may be responsible for a nuclear power revival.

Here is the connection. Training and running large AI models requires enormous amounts of electricity. Companies building AI infrastructure — Microsoft, Google, Amazon, and others — need reliable power, in massive quantities, delivered to specific locations. They don’t want to wait years for new power lines. They want guaranteed electricity at guaranteed prices.

Nuclear power plants, once built, produce electricity continuously regardless of weather. So these tech companies have started signing long-term contracts directly with nuclear operators — agreements called power purchase agreements, or PPAs — that essentially fund nuclear plants in exchange for dedicated power.

This creates a funding pathway for nuclear that bypasses the normal electricity market. It’s significant because nuclear construction is enormously expensive upfront. The financing cost — how much interest a developer pays while waiting years for the plant to generate any revenue — is often the biggest obstacle. A guaranteed long-term customer changes that math.

The catch: this creates a two-speed energy system within the nuclear industry. Hyperscalers with deep pockets get dedicated nuclear power. Everyone else waits. And the same AI data centers are also driving up demand for natural gas in the near term, because gas plants can be built much faster than nuclear. So AI is simultaneously a long-term driver of nuclear development and a short-term driver of fossil fuel lock-in.

---

The Developing World Gets Left Behind (And the Reasons Are Connected)

One of the clearest findings across these explorations is that the clean energy transition is moving at two very different speeds depending on where you are in the world.

In wealthy countries with stable governments and deep financial markets, clean energy investment is accelerating. In developing countries — much of Africa, South and Southeast Asia, Latin America — it is largely stalled.

The surface-level explanation is cost of capital. When a bank lends money for a solar project in Germany, it charges relatively low interest because Germany is considered a safe place to invest. The same bank, lending for a solar project in Nigeria, charges much higher interest because it perceives more risk — currency volatility, political uncertainty, the possibility that contracts won’t be honored. That higher interest rate makes the Nigerian solar project far more expensive, even if the panels and the sunshine are the same.

But the deeper pattern — which only emerges when you look across the finance exploration, the climate change exploration, and the mineral exploration together — is that these disadvantages are not independent. They stack and amplify each other.

Developing countries often sit on top of the mineral deposits the world needs for clean energy, but don’t capture the refining and manufacturing value. They face the most severe physical effects of climate change — floods, droughts, heat — which damages infrastructure and increases insurance costs, which further worsens investment conditions. The international finance mechanisms designed specifically to help — programs called JETPs, or “Just Energy Transition Partnerships” — have not delivered the promised funding on the necessary scale.

The result is a self-reinforcing trap: high physical risk raises capital costs, which reduces investment, which slows climate adaptation, which increases physical risk. No single exploration reveals this fully; it only becomes clear when the findings from finance, physical climate risk, and minerals are read together.

---

What Oil Companies Are Actually Betting On

The major oil companies — Shell, BP, ExxonMobil, Saudi Aramco — are not all making the same bet, and understanding their differences reveals something about how uncertain the future actually is.

Saudi Aramco is continuing to invest heavily in oil production. Its implicit bet is that demand from Asia — particularly India and China’s growing petrochemical industry — will remain strong long enough to justify that investment.

BP, after years of aggressive renewable energy pledges, has pulled back and is refocusing on oil and gas. Its implicit bet is that governments will not impose carbon pricing strong enough to make existing fossil fuel assets unprofitable within the relevant investment horizon.

This divergence reflects a genuine uncertainty: the fossil fuel assets that exist today might become “stranded” — worthless before they’ve paid back their investment — if climate policy tightens. Or they might not. The companies are making different predictions about which outcome is more likely.

There is also a political economy dimension: companies with large fossil fuel assets have structural incentives to lobby against the carbon pricing that would strand those assets. The policy outcome they’re betting on is partially something they can influence. This makes the forecast and the advocacy mutually reinforcing in a way that complicates clean prediction.

---

Green Hydrogen’s Circular Problem

Green hydrogen — hydrogen made by splitting water using renewable electricity — is supposed to be a clean fuel for industries that are hard to electrify: steel, shipping, long-haul aviation. But it faces an unusual obstacle.

Green hydrogen needs large industrial customers to justify the scale of production required to bring costs down. But large industrial customers — steel mills, fertilizer plants, shipping companies — won’t commit to hydrogen until costs are lower. This is a classic “chicken and egg” problem.

What the explorations reveal is that this problem has three locks, not one. Carbon pricing would make green hydrogen competitive with fossil alternatives, but carbon pricing is weak in most places. Electrolyzer costs (the equipment that splits water) need to fall further, but they fall faster at higher volumes, which requires the customers who don’t yet exist. And renewable electricity — the main input cost — needs to be cheap in the same geographic location where the industrial customer operates, which is not always the case.

The green hydrogen valley of death is not just a financing problem. It requires at least two of these three conditions to change at roughly the same time. Understanding that only becomes clear when the hydrogen exploration, the carbon market exploration, and the transition timeline exploration are read together.

---

The Carbon Market Problem

Carbon markets — systems where companies can buy and sell credits representing avoided greenhouse gas emissions — sound elegant in theory. In practice, the research shows significant gaps between design and function.

Many carbon credits have been issued for forests or projects that didn’t actually reduce emissions in the way claimed. The accounting is genuinely hard: how do you prove a forest wouldn’t have been cut down anyway? Regulatory oversight varies enormously across countries and programs. And carbon prices in most markets remain far below the level economists calculate would actually change corporate behavior.

The EU’s market is the most developed and has the highest prices, but even there, the price has been volatile and occasionally low enough that it doesn’t meaningfully constrain emissions. The EU is attempting to extend carbon cost pressure to imports through a “carbon border adjustment” — essentially a tariff on goods from countries with weaker carbon rules — but this is new and its effectiveness is unproven.

The finding across explorations is that carbon pricing, as actually implemented globally, is not yet a reliable mechanism for driving the pace of decarbonization that physical climate models say is necessary.

---

Bottom Line: What the Data Actually Shows

Several structural insights only become clear when these eight explorations are read as a connected system rather than separate questions.

The transition is real but geographically bifurcated. Global clean energy deployment is accelerating, and 2025 may represent an actual peak in global emissions. But this transition is concentrated in a small number of well-capitalized markets. The developing world — which contains most of the world’s people and will generate most of future emissions growth — is largely excluded by financial structures, not by technology costs.

China’s role is more compound than it appears. China is not just a manufacturer of solar panels. It sits at the mineral refining stage, the manufacturing stage, and increasingly the project financing stage of the global clean energy system. This compound position means that efforts to “de-risk” supply chains by diversifying away from China face a more complex challenge than any single policy instrument can address.

AI has introduced a new, time-split dynamic. In the near term, AI data center demand is pulling gas infrastructure investment. Over a longer horizon, the same demand is driving the most significant nuclear financing development in decades. Whether AI is net positive or negative for decarbonization depends almost entirely on the gap between those two timescales — and that gap is currently unresolved.

The policy and physical infrastructure are the binding constraints, not technology cost. For most of the last decade, the primary obstacle to clean energy was that it cost too much. That obstacle has largely been removed for solar and batteries. The new binding constraints are permitting rules, transmission capacity, interconnection queues, and financial architecture for developing countries. These are slower to change than technology costs, which is why the transition is moving faster in some dimensions than others.

The feedback loops that trap developing countries are not independent problems. Higher climate risk, higher capital costs, lower refining value capture from minerals, and failed international finance mechanisms are not parallel misfortunes. They are causally linked in ways that make each worse. Solving any one in isolation does not break the loop. That structural interdependence is the most important finding that does not appear in any single exploration but emerges clearly when all eight are read together.