---

Discussion

Red Velhouse:

Eric, the headline numbers sound extraordinary: about 692 gigawatts of renewable capacity were added globally in 2025, roughly 85.6 percent of all power-capacity additions. Is the clean-energy transition breaking through, or are we misreading the scoreboard?

Eric Arcan:

It is a genuine breakthrough, but the scoreboard needs a label. That is installed capacity, not electricity generated every hour. Solar panels and wind turbines can be added quickly, but their output varies. The harder question is whether grids have enough transmission, storage, flexible demand and dependable generation to deliver power when people need it. So yes, renewable growth is real. But capacity growth alone does not tell us whether coal and gas plants are actually running less.

Kate Burvish:

And demand is growing at the same time. Renewable power can expand rapidly while new factories, cooling systems, electrification and data centers absorb much of the increase. That does not make the investment pointless; it may prevent a much dirtier expansion. But preventing additional fossil use is different from displacing existing fossil use. The economic result depends on both sides of the ledger: clean supply and total consumption.

Red Velhouse:

So the first distinction is capacity versus generation, and the second is new demand versus displacement. Ann, that distinction collides with the politics at Climate Week. The event emphasizes implementation, while the United States is simultaneously emphasizing fossil-fuel production and pursuing another withdrawal from the Paris Agreement. What does that contradiction do politically?

Ann Tofado:

It creates a split-screen politics. Internationally, leaders and companies can frame clean electrification as climate protection, energy security and industrial competitiveness. In Washington, the federal administration is framing expanded energy production—including fossil fuels and nuclear power—as an affordability and national-security priority. Those frames compete, but they can also overlap around the need for a larger, more reliable grid. The political conflict is less about whether society needs more power than about which technologies receive support, who pays, and whose authority sets the rules.

Red Velhouse:

That larger grid brings us to the fastest-moving source of new demand. Kate, the International Energy Agency, or IEA, estimates that global data-center electricity use rose 17 percent in 2025, while electricity use by artificial-intelligence-focused centers rose about 50 percent. Is AI becoming a climate problem, an infrastructure opportunity, or both?

Kate Burvish:

Both, and the economics are unusually concentrated. The IEA's current projection puts data-center consumption near 950 terawatt-hours in 2030, almost twice the estimated 2025 level. That demand can attract investment in generation, transmission, storage and equipment. But data centers are geographically concentrated, so the benefits and costs land unevenly. A community might receive jobs and tax revenue while facing congestion, water pressure or higher costs if new infrastructure is spread across ordinary customers.

Eric Arcan:

I agree with the distributional warning, but there is a physical opportunity here too. A large, predictable load can help finance new power plants and transmission that would otherwise be difficult to build. The catch is timing. A data center can be ready before a clean project, a transformer or a transmission line is ready. If planners fill that gap with existing gas or coal, the load becomes an argument for fossil lock-in rather than a catalyst for cleaner supply.

Ann Tofado:

And that timing problem quickly becomes a legitimacy problem. People may accept national arguments about artificial intelligence, competitiveness or security until their electricity bill rises or a water permit becomes controversial. Local opposition is not necessarily anti-technology. It can be a demand for a visible bargain: the project gets access to power, but it also pays for the reliability and environmental systems it requires.

Red Velhouse:

Eric, let’s make that bargain more concrete. Should data centers pay the costs they cause—generation, transmission, reliability upgrades and water systems—or could that slow useful investment?

Eric Arcan:

They should pay the costs they cause, especially for dedicated upgrades and reliability. Otherwise we privatize the benefit and socialize the infrastructure bill. But “full marginal cost” needs careful engineering. A new line may serve several customers, and a power plant may improve the whole region's reliability. The fair rule is not to punish large loads; it is to make cost allocation transparent and prevent everyone else from becoming the involuntary backstop.

Kate Burvish:

That is also the best investment signal. If a data center pays nothing for congestion, it will locate where power appears cheapest even when the system cost is high. If it pays every conceivable future cost, investment may flee or become unnecessarily expensive. Regulators need contracts that distinguish firm service from interruptible service, and they need to show ratepayers what they are receiving in return. The Federal Energy Regulatory Commission, or FERC, has begun pushing regional grid operators to justify or reform these tariffs, but that process guarantees neither speed nor fairness.

Red Velhouse:

The cost question is local, but the policy signal is national. Ann, the federal government has also changed the investment backdrop. Public Law 119-21 ended the residential clean-energy credit after 2025 and restricted the clean-electricity credit for some new wind and solar projects after 2027. How much political damage does that do?

Ann Tofado:

It raises uncertainty more than it ends deployment. Existing projects, state policies, corporate procurement and remaining incentives still matter. But long-lived infrastructure depends on confidence that the rules will survive the construction period. A deadline can produce a rush to qualify, followed by hesitation on projects that need years to permit and build. Politically, the reversal also tells investors that federal climate policy can swing sharply between administrations, which makes every project carry a larger policy-risk premium.

Kate Burvish:

And that premium is not abstract. It can affect financing costs, supply-chain commitments and which technologies get built first. Some developers may accelerate projects before a credit expires; others may wait until demand, permitting and financing line up. The market can preserve momentum, but policy volatility makes the transition more expensive than a stable framework would.

Red Velhouse:

Let’s connect that uncertainty back to the engineering question. Eric, can the United States meet rising demand without keeping substantial gas capacity?

Eric Arcan:

We should stop treating this as a one-technology morality play. The system needs much more renewable generation, but also transmission, storage, efficiency, demand response and firm power. Nuclear can contribute where projects are viable. Gas may provide backup, but building unabated gas as the default answer risks locking in emissions and fuel-price exposure. The engineering test is hourly reliability at reasonable cost, with a credible plan for reducing fossil operation over time—not simply announcing clean capacity on one side and backup capacity on the other.

Kate Burvish:

And fossil-fuel price shocks strengthen the economic case for diversification. War-related disruptions have reminded governments that volatile fuel prices are not just an environmental concern; they are also an affordability and security concern. But renewable generation is not free insurance. Grids, storage, permitting and manufacturing require large upfront spending. The relevant comparison is not “free wind versus expensive fuel.” It is the total cost and risk of each system over decades.

Ann Tofado:

That broader framing may be the most politically durable one. Climate arguments remain important, especially as climate damage worsens, but coalitions often expand when clean energy is connected to jobs, energy independence, reliability and industrial capacity. The danger is that national-security language can justify speed without accountability. Local communities still ask who controls the land, who receives the benefits and who absorbs the risks.

Red Velhouse:

So AI can be both a new load and a possible tool for managing that load. Eric, could artificial intelligence help decarbonize the system through forecasting, grid management and efficiency, or are those promises ahead of demonstrated emissions reductions?

Eric Arcan:

The potential is credible, but potential is not a power plant. Better forecasting can reduce renewable curtailment. Smarter controls can coordinate batteries, flexible loads and transmission. But those gains require access to operational data, upgraded equipment and rules that let the system respond. Meanwhile, the electricity demand from AI is immediate. I would judge the claim by measured outcomes: lower emissions per unit of computing, more flexible demand, fewer hours of fossil backup and actual grid projects delivered.

Kate Burvish:

I would add that efficiency can produce a rebound effect. If computing becomes cheaper, people and firms may use more of it, so efficiency per task does not automatically mean lower total energy use. The IEA also identifies AI profitability and capital-market conditions as major uncertainties. A boom could intensify infrastructure bottlenecks; a slowdown could reduce projected demand. Utilities are being asked to build assets for a forecast, not for a settled fact.

Ann Tofado:

And internationally, the proposed Turkish initiative on managing the powering and deployment of AI is worth watching precisely because its shape is unknown. It could become disclosure guidance, a voluntary framework or something more enforceable. The political challenge is getting governments and companies to accept common transparency rules without treating them as a brake on technological competition. For now, the announcement is a signal, not a binding outcome.

Red Velhouse:

Then what would distinguish implementation from another week of impressive commitments? Give me concrete indicators for the next year—things that would show whether the capacity, demand and politics are actually lining up.

Eric Arcan:

Start with transmission and interconnection milestones: permits issued, transformers procured, storage actually installed, and power plants or large loads connected under rules that preserve reliability. Then measure generation, not just capacity. Are fossil plants running fewer hours? Are emissions falling as electricity demand rises? Those are harder facts than a pledge.

Kate Burvish:

I would watch who pays and what prices do. Are data centers signing contracts that cover the infrastructure they trigger? Are household and industrial customers protected from hidden cost shifts? Is private capital still flowing when incentives change, or are projects surviving only through increasingly expensive public support? Durable economics should show up in investment, construction and affordability—not just announced spending.

Ann Tofado:

Politically, watch whether governments turn broad alignment into durable institutions. The United Nations climate-plan synthesis covered only 64 new national plans, representing about 30 percent of 2019 global emissions. That limited coverage is a reminder that international momentum is uneven. The real test is whether national commitments become budgets, permits, grid decisions and policies that survive electoral change.

Red Velhouse:

The unresolved issue is whether clean-energy growth is finally displacing fossil fuels, or mainly serving a larger and more demanding electricity system—especially one reshaped by artificial intelligence. Watch the next year for grids and storage actually built, data centers paying transparent infrastructure costs, emissions falling despite rising demand, and whether international AI-energy proposals become rules or remain rhetoric. Climate Week can focus attention, but implementation will be measured in power delivered, costs allocated and fossil capacity genuinely retired. Sources and references for this discussion are available with the episode at Factolio.com.

---

Sources and References

These sources supported the factual material used in this discussion. Factolio’s panel discussion is AI-generated from researched evidence and is written in original language.

1. Associated Press — Fuel prices, AI and disasters complicate UN climate efforts (NEWS)
2. Climate Week NYC / The Climate Group — Highlights: Climate Week NYC 2026 (PRIMARY)
3. Climate Week NYC / The Climate Group — Agendas for Climate Week NYC 2026 (PRIMARY)
4. International Renewable Energy Agency — Near-700 GW Surge in 2025 Proves Renewable Energy Resilience (PRIMARY)
5. International Renewable Energy Agency — Renewable Capacity Statistics 2026 (DATA)
6. UN Climate Change / UNFCCC — Renewable power generation records its fastest growth ever (PRIMARY)
7. International Renewable Energy Agency — Delivering on the UAE Consensus: Tracking progress toward tripling renewable energy capacity and doubling energy efficiency by 2030 (ANALYSIS)
8. International Renewable Energy Agency — Tripling renewable power and doubling energy efficiency by 2030: Crucial steps towards 1.5°C (ANALYSIS)
9. International Energy Agency — Key Questions on Energy and AI (ANALYSIS)
10. International Energy Agency — Energy and AI: Energy supply for AI (ANALYSIS)
11. U.S. Energy Information Administration — Annual Energy Outlook 2026 (DATA)
12. Federal Energy Regulatory Commission — FERC Launches Aggressive Targeted Action to Speed Large Load Integration (PRIMARY)
13. U.S. Congress — Public Law 119–21, July 4, 2025 (PRIMARY)
14. Internal Revenue Service — FAQs for modifications under Public Law 119-21 (PRIMARY)
15. The White House — American Energy Dominance Is Back Under President Trump (PRIMARY)
16. Associated Press — Trump signs executive order directing US withdrawal from the Paris climate agreement (NEWS)
17. UNFCCC — 2025 NDC Synthesis Report (PRIMARY)
18. World Meteorological Organization — State of the Global Climate 2025 (DATA)