Data center load forecasts sharpen as Germany crosses a clean-power threshold

Key Developments

Data-center load forecasts widen as projects get larger

Utility Dive reported that BloombergNEF’s base case now puts U.S. data-center electricity demand at 118 GW in 2030 and 194 GW by 2035, representing upward revisions of 52% and 83% from BNEF’s December outlook, while a chip-delivery scenario puts U.S. data-center demand at 207 GW by 2033 (Utility Dive). The article said installed U.S. data-center capacity topped 47 GW by the end of 2025, 16% above BNEF’s forecast, and quoted BNEF as saying it added about 100 GW of project capacity across the U.S. over the last year (Utility Dive). Utility Dive also reported that more than 70 projects in BNEF’s pipeline are at least 1 GW, with some projects up to 10 GW, and that the gap between two 2030 demand scenarios is 42 GW (Utility Dive).

The grid-planning read-through is that data-center demand uncertainty has moved from a timing problem to a capacity-sizing problem. A 42 GW forecast spread is not a rounding issue for utilities, transmission planners, or generation procurement; it is large enough to change reserve-margin math, queue-screening rules, and customer-financing terms. The more consequential angle is that the growth case is increasingly shaped by project scale and chip-delivery assumptions, not only by historical load growth.

What to watch: Track whether utilities and RTOs begin using chip-supply-linked scenarios alongside signed-interconnection queues, because a project pipeline with 1 GW to 10 GW facilities can make conventional load forecasts stale before assets enter service (Utility Dive).

CenterPoint turns Texas large-load rules into a near-term capital question

Utility Dive reported that CenterPoint Energy submitted more than 17 GW of large-load projects to ERCOT’s new large-load interconnection process and expects 14 GW to qualify as base load or studied load in the first “Batch Zero” group (Utility Dive). CenterPoint said the 14 GW would represent more than a 65% increase from its current Houston-area peak system demand of 21 GW, and the company expects 50% load growth by the end of 2029 (Utility Dive). The article said about 10 GW of the 14 GW have all required ERCOT studies approved and are eligible for base-load designation, while 4 GW have one of two required studies approved and are positioned to qualify as studied load (Utility Dive). Utility Dive also reported a $1.2 billion increase to CenterPoint’s 10-year capital plan, bringing planned investment to $66.7 billion through 2035, and approximately $900 million of customer cash commitments and deposits tied to the 14 GW (Utility Dive).

The operational implication is that Texas’ large-load screening process is already feeding directly into utility capital allocation. Customer deposits and ERCOT study status help separate speculative demand from load that can move into system planning, but the scale still stresses sequencing: transmission, materials, generation adequacy, and cost recovery must align before late-decade energization dates. The CenterPoint case makes data-center growth less abstract than national forecasts because it ties megawatts to a named service territory and a published capital plan.

What to watch: Watch Batch Zero approvals, facility-extension agreements, and CenterPoint’s future capital-plan updates; changes in those items will show whether the 14 GW converts into executable utility work or remains partly contingent pipeline demand (Utility Dive).

Germany’s wind-and-solar crossover shifts the policy debate toward backup and grid rules

Carbon Brief reported that wind and solar generated 225 TWh of electricity in Germany in 2025, equal to 44% of total generation, while fossil fuels generated 217 TWh, or 43%, marking the first year in which wind and solar overtook fossil fuels in the country (Carbon Brief). Carbon Brief said the EU also saw wind and solar overtake fossil-fuel power generation in 2025 for the first time, and that Germany aims to raise renewables’ share of electricity consumption to 80% by 2030 while targeting a largely climate-neutral power system by 2035 (Carbon Brief). The same analysis said Germany targets 115 GW of onshore wind by 2030 and approved a record 20.8 GW of new onshore-wind capacity in 2025 (Carbon Brief).

Figure 1 — Wind and solar generated 225 TWh (44% of total generation) in Germany in 2025, edging past fossil fuels at 217 TWh (43%) — the first year renewables overtook fossil generation. Source: Carbon Brief.

The read-through is that Germany’s transition debate is no longer only about whether variable renewables can become the largest power source; it is about how the system prices backup, curtailment, and connection priority after the crossover. Carbon Brief noted that Germany’s government has pursued support for new gas-fired plants intended to convert to green hydrogen by 2045, while planned grid reforms have drawn renewables-industry criticism because developers in constrained areas could need to waive future curtailment compensation for automatic grid connections (Carbon Brief). That makes the next stage less about headline generation share and more about the market design for firm capacity and grid congestion.

What to watch: Watch Germany’s August coal-phaseout timeline review and grid-connection reform details; those decisions will indicate whether the 2025 crossover accelerates clean-capacity deployment or pushes more cost into backup and curtailment management (Carbon Brief).

Subsidy disclosure gaps keep fossil-fuel support in the biodiversity policy frame

Carbon Brief reported that only 21 countries, or 16% of the 134 national reports submitted to the UN Convention on Biological Diversity by July 1, 2026, appeared to have identified their harmful subsidies by the 2025 target date (Carbon Brief). The analysis found that 32 countries identifying some or all such subsidies spend almost $270 billion annually on biodiversity-harming incentives, while the Kunming-Montreal framework calls for countries to identify harmful subsidies by 2025 and eliminate, phase out, or reform at least $500 billion per year by 2030 (Carbon Brief). Carbon Brief said harmful subsidies can be found in fossil fuels, agriculture, forestry, mining, and fishing, and quoted Prof. Jessica Dempsey saying fossil-fuel subsidies can be biodiversity-harming because they drive climate change and can cause localized extraction harms (Carbon Brief).

For energy markets, the policy signal is disclosure before reform. If most countries have not identified the relevant subsidy base, then the 2030 phaseout-or-reform target is difficult to translate into fuel-price, fiscal, or demand assumptions. The fossil-fuel angle is not that every biodiversity subsidy maps cleanly to energy, but that fossil-fuel support remains inside a broader nature-and-climate subsidy reform agenda, which can resurface in fiscal-policy negotiations even when climate legislation stalls.

What to watch: Track COP17 in Armenia in October 2026 and any country-level subsidy inventories published before then; the practical energy-market effect depends on whether fossil-fuel support is quantified separately or buried inside broad multi-sector subsidy categories (Carbon Brief).

This is an AI Briefing — AI-generated analysis published under TLCapital.AI. It is not personal research or positions, and it is not investment advice. Figures are sourced to primary filings with dates noted throughout. Do your own diligence.