Hydrogen sector, office transformation and AI’s demand for nuclear power

AI, powered by data centres, needs massive amounts of energy to function – and demand is growing. Moving forward, finding efficient and sustainable energy solutions for the ever-growing demand of these companies will be key.

In this edition, we take a look at:

• Why do hydrogen companies fail to ignite, despite a positive long-term outlook for the sector?
• Will the office real-estate sector regain traction?
• Can nuclear power be the solution for tech companies’ energy needs to power AI?

Investor Perspective: Hydrogen sector still failing to ignite

There’s still plenty of interest in hydrogen. Total announced investments across all project phases have increased from USD90bn in 2020 to USD680bn in 2024 (although committed investments are lower). In the period up to 2030, 1,125 projects are under consideration, with an expected annual output of 48 Mt of clean hydrogen.[1]

But despite this apparent interest, many hydrogen companies are continuing to struggle. The S&P Kensho Hydrogen Economy Index (tracking companies across the value chain for low-carbon hydrogen) has fallen by around 18% this year and is now down around 60% from its February 2022 peak.[2]

The sector’s woes in the U.S. may be due (in part) to the slow rollout of tax credits, as well as lower short-term demand for hydrogen. In Europe, slow and limited government funding, paired with high regulatory barriers in some countries has put an even greater brake on decision making.[3]

In the short and medium term, these problems are unlikely to be overcome (and could get worse). So there’s plenty of room for further downside risks, meaning investors should tread carefully. The key question is whether truly green hydrogen can really be produced at a lower price. The IEA’s Net Zero Emissions by 2025 Scenario assumes an expected 50% drop in the cost of green hydrogen for end users by 2030 in to 2-9 USD/kg.[4] If this is going to happen, more policy support would need to be forthcoming.

Our view: Office real estate has to transform

When we look at real estate from an ESG perspective, we usually focus on its environmental (E) implications.  But I think it’s also important to consider real estate’s social (S) impact too, in commercial as well as residential sectors.

One recent study[5] based on cell phone data to track foot-traffic patterns in U.S. cities, for example, reckons that while the number of office visits is increasing, they are still down 28% from 2019 levels.

My concern is that we shouldn’t assume that office usage will eventually return to pre-pandemic levels, given sufficient policy encouragement. Rising use of AI could change the game again in the future, reducing or changing the nature of office use, and city and social planners need to consider some possible social scenarios here.

DWS and other real-estate forecasters remain cautious on the economic outlook for the office sector, and I think there are good reasons for this.

Under debate: AI and nuclear power

Recent research points to the potential of artificial intelligence (AI) to help drive net-zero solutions and combat climate change.[6]  But AI also creates a carbon problem, through its high and increasing energy consumption.  AI data centres consumed around 275 TWh of electricity in 2022, almost 1.2% of total global electricity demand. This share could rise to 3% by 2027.[7]

Tech companies are therefore looking for new energy sources to meet anticipated demand[8]. One result has been some initial agreements with firms in the nuclear sector, based on the apparent potential of smaller, “next generation” nuclear power plants, located close to data centres (and hence mitigating fears around electricity grid capacity). In addition to the lack of carbon emissions, much has been made of the new nuclear technologies’ claimed safety and the relatively small amount of nuclear waste generated.

So will this marriage between tech and nuclear work? It’s very early days. The technology involved in many of these projects is still speculative and it is not clear what the real costs (financial and other) will be. Regulators and surrounding populations will also likely continue to focus on the safety aspect. Some factors that have limited the growth of the U.S. and European nuclear industries for decades[9] will remain. But it is definitely an area to monitor – whatever your views on nuclear power.

References

[1] Hydrogen Insights September 2024 (hydrogencouncil.com)

[2] S&P Kensho Hydrogen Economy Index | S&P Dow Jones Indices (spglobal.com)

[3] US and European hydrogen stock prices collapse as prospects deflate (ft.com)

[4] Global Hydrogen Review 2024 (iea.blob.core.windows.net)

[5] Study finds hybrid work benefits companies and employees | Stanford Report

[6] Artificial intelligence potential for net zero sustainability: Current evidence and prospects - ScienceDirect

[7] Carbon Emissions from AI and Crypto Are Surging and Tax Policy Can Help (imf.org)

[8] Google joins Big Tech's move into nuclear power, and other top energy stories | World Economic Forum (weforum.org)

[9] Exclusive | Wall Street Giants to Make $50 Billion Bet on AI and Power Projects - WSJ