Sustainable Energy - February Commentary

This is a marketing communication. Please refer to the prospectus, supplement and KID/KIID for the Funds before making any final investment decisions. The value of this investment can fall as well as rise as a result of market and currency fluctuations. You may not get back the amount you invested.  

Past performance does not predict future returns. 

ENERGY ISSUES FOR THE TRUMP ADMINISTRATION

One month into his presidency, Donald Trump is already having an impact on global energy markets. This month, we review the executive orders that he announced on his first day in office and how the recent news from DeepSeek might affect the outlook for AI-related power supply and the need for significant grid upgrades.

Executive orders relating to energy

Many of the executive orders from 20th January, the first day of President Trump’s new term, related to the energy sector. Some were specific to the fossil fuel industry while others reflected the broader need for greater access to cheap energy to satisfy estimates of growing demand. We provide a summary here of some of the key executive orders for the sustainable energy sector:

• Declaring a national energy emergency
  This order calls for using emergency powers to facilitate the leasing, siting, production, transportation, refining and generation of energy. In essence, Trump is seeking the authority to i) reduce environmental restrictions on existing energy infrastructure and ii) ease permitting for new transmission and pipeline infrastructure to expedite the completion of infrastructure and natural resource projects. The order specifically refers to the need to “swift and decisive action” to remedy an “increasingly unreliable grid”.
• A withdrawal from the Paris Agreement
  The withdrawal from Paris was much expected and was indeed a repeat of the withdrawal made in his first term. Trump has previously called climate change a hoax, and says the accord puts the United States at a competitive disadvantage to geopolitical rivals like China. Trump said "I'm immediately withdrawing from the unfair, one-sided Paris climate accord rip-off," and "The United States will not sabotage its own industries while China pollutes with impunity."
• Revoking Biden’s 2021 electric vehicle (EV) targets
  The revocation of EV targets was also widely expected and was accompanied by a halting of unspent government funds (from a $5bn government fund) for vehicle charging stations and the removal of a waiver which allows certain states (California and 11 others) to adopt zero-emission vehicle rules by 2035. Trump also said his administration would consider ending EV tax credits.
• Offshore wind suspension
  This order suspends new federal offshore wind leasing pending an environmental and economic review. Trump described wind turbines as “ugly, expensive” and “harmful to wildlife”.

Many of his energy-related executive orders targeted the fossil fuel sector more than the sustainable energy sector. Fossil fuel related orders included a repeal of Biden's efforts to block oil drilling in the Arctic, a plan to fully refill the US Strategic Petroleum Reserve (SPR) and an order to resume liquefied natural gas (LNG) export permit applications from new LNG projects supplying Asia and Europe. We felt that sustainable energy activities were notably absent from his day one executive orders and, bar the suspension of some IRA-related loans, there was no mention of plans to repeal the Inflation Reduction Act itself.

We see increasing urgency over these broader energy-related issues, and in particular, we stress that surging US electricity demand (as a result of the growth of artificial intelligence querying and data centres as well as the wider trend of electrification) is a critical issue. It is imperative for Trump to deal with this if he is to win the ‘AI arms race’, requiring him to oversee significant grid upgrades and near-term growth in both renewable and natural gas-based power generation, as referred to in his ‘energy emergency’ executive order.

The impact of DeepSeek on AI power demand growth

A few days after Trump’s inauguration, Chinese startup DeepSeek announced significant computational and energy efficiency improvements in its latest AI model, causing market concerns around the long-term outlook for AI energy demand. The DeepSeek news implies that language learning (AI training) could potentially be significantly less energy intensive than anticipated, with knock-on effects for global AI-related power demand. We believe that there are two factors to consider here in thinking about the longer-term implications:

• First, the relative importance of AI training within overall AI power demand. We understand that the vast majority of forecast US data centre computational and power demand is for the use of AI (AI inference) as opposed to AI training. We note that some analysts, for example Cushman & Wakefield, indicate that 90% of capacity will be used for inference although many commentators suggest a level more like 75%. Thus, significant efficiency improvements in training (which DeepSeek claim and which represent a small relative share of power demand) would not have a material impact on data centre capacity or power consumption.
• Secondly, the potential for vast latent demand for AI in the event that it becomes cheaper and more accessible. This is often referred to as the Jevons Paradox (or the rebound effect) which states that increases in efficiency can lead to an increase in consumption, thereby offsetting the gains made by the efficiency improvements. The step change announced by DeepSeek could simply pull forward the opportunity set for deploying AI across a wider range of industries and demand cases. This would clearly be a net positive for the majority of power demand that is related to inference, potentially leading to much greater overall power demand.

Overall, we see good chance that energy demand for AI actually increases post the DeepSeek news. This will likely place more pressure on the US administration to facilitate greater, diverse forms of electricity supply and to more rapidly build a stable power grid.

Since the DeepSeek news, a number of US tech companies have reported full-year results and provided 2025 guidance. So far, it seems that AI investments continue to accelerate sharply with capex from the four leading US tech companies (Microsoft, Alphabet, Amazon and Meta) likely to be $320bn in 2025, up 30% from $246bn in 2024 and more than double the spend delivered in 2023. Since the start of 2024, cloud capex forecasts for 2025 have been revised higher by around 60% as the companies fight to maintain their positions at the forefront of AI large language model research.

Capital expenditure of US hyperscalers

Source: Nat Bullard, January 2025

On balance, it seems that a sharp slowdown in AI development seems unlikely, meaning that new power sources and grid upgrades remain just as critical as before. While the outlook for AI power demand growth might be volatile around such newsflow, we note that c.60% of our forecasted US power demand growth comes from non-AI drivers (onshoring of manufacturing, electrification, electric vehicles, etc.).

Increase in US electricity demand (2023-2030) split by source of demand

Source: Nat Bullard, January 2025

With respect to sustainable energy and electrification, it is clear that Trump’s executive orders indicate a focus on delivering cost competitive new power and investing to build a more stable grid. News from DeepSeek does not appear to have de-railed the strong demand outlook and, after 20 years of flat electricity consumption, we see demand growth of around 2-3% per year due to data centres, AI querying, reindustrialization and electrification. There is an urgency for Trump to resolve these issues and we would expect this action to provide some positive relief for the sustainable energy sector.

READ FULL COMMENTARY