Tech companies accelerate nuclear deployment

Over the past couple of years, it has become increasingly clear that the futures of the tech and power industries are inextricably intertwined because of the expected growth of electricity demand for AI and data centres. It means that two industries with very different stakeholders, cultures and mindsets are being forced to work together to pursue their goals.

One critical difference is that the tech companies can be more agile than power producers, which – for good reasons – generally move much more slowly. The big tech companies’ ability to act rapidly was on display this week when Amazon and Google announced important moves to support new nuclear generation as an answer to their surging demand for electricity.

US utilities have given up on building new nuclear plants following the long delays and cost overruns at the Plant Vogtle Units 3 and 4 in Georgia. Even the UAMPS Carbon Free Power Project to build NuScale small modular reactors (SMRs) in Idaho, which was backed by both the Trump and Biden administrations, was abandoned last year after failing to secure enough utility buyers for its electricity.

Tech companies are much less price-sensitive buyers, and are showing that they are prepared to pay higher costs to secure 24/7 carbon-free power.

Google agrees to buy power from new SMRs

On Monday, Google announced an agreement to buy electricity from multiple SMRs developed by Kairos Power, an innovative advanced nuclear company. Its aim is to have the first of the new reactors online by 2030, adding more to reach a total of up to 500 megawatts of output by 2035.

Google thinks the next generation of advanced reactors, with simplified designs and robust inherent safety, will allow for faster construction and reduce delays and cost overruns. The Kairos SMR, which uses a molten-salt cooling system, was the first such advanced reactor design to receive a construction permit from the US Nuclear Regulatory Commission. The project to build a non-powered demonstration reactor in Tennessee broke ground over the summer.

By agreeing to buy power from a group of reactors, Google will help Kairos scale up and drive down costs. With any emerging technology, the first-of-a-kind project is almost always the most expensive. Learning by doing through multiple deployments should make each reactor project more efficient than the one before.

Michael Terrell, Google’s senior director for energy and climate, said advanced nuclear and other 24/7 carbon-free technologies, such as enhanced geothermal, were at the same early stage of their development that wind and solar power were when his company started signing power purchase agreements (PPAs) in 2010. “It’s great to see others getting in the game. The more the merrier,” he said. “I’d like to see a hundred more companies doing nuclear deals.”

Amazon invests in an SMR technology company

A day after the Google news, Amazon announced its own leap into advanced nuclear technology. It has agreed a series of deals intended to accelerate the deployment of a different SMR design: X-energy’s Xe-100 80-MW gas-cooled reactor, which uses TRISO fuel, described by the US Department of Energy as “the most robust nuclear fuel on earth”.

Amazon has gone a step further than Google in that it is actually investing in X-energy. Amazon’s Climate Pledge Fund is the anchor investor for a total of US$500 million that X-energy is raising in a Series C-1 financing round. The other investors include Ken Griffin, founder and CEO of investment firm Citadel; Ares and NGP, two other investment firms; and the University of Michigan. X-energy will use the money to fund the completion of its reactor design and licensing, and the first phase of its TRISO fuel fabrication facility in Tennessee.

Amazon has set a goal of deploying 5 gigawatts of new power projects with X-energy SMRs across the US by 2039. The first of those projects will be built, owned and operated by Energy Northwest, a consortium of public utilities in Washington state. Phase one will include four of the X-energy SMRs, giving a total capacity of 320 MW, with the aim of coming into service by the early 2030s. There is an option to expand the project to 960 MW over time. Amazon is making a direct investment to fund X-energy’s early development work.

Amazon has also signed an agreement with Dominion Energy to work on the development of an SMR project in northern Virginia. That area is famous for “data centre alley”, the world’s largest concentration of data centres, and Dominion projects that electricity demand in the region will rise by 85% over the next 15 years. The new SMR would be located near Dominion’s existing North Anna nuclear power station.

Wood Mackenzie view

Nuclear power enjoys bipartisan support across the political spectrum in the US, because of its importance both for climate policy and for national security. Its big problem has been that until now, there have been no buyers willing to pay premium prices for 24/7 carbon-free power. David Brown, director of Wood Mackenzie’s energy transition practice, estimates that the levelised cost of energy (LCOE) from a new SMR could be US$200 per megawatt hour. That is about four times the LCOE of power from utility-scale solar, and three times the LCOE for a new combined-cycle gas turbine plant.

The latest moves from the tech companies, along with other earlier announcements – including Microsoft’s intention to restart a reactor at the Three Mile Island plant and Oracle’s plans for a gigawatt-scale data centre powered by three SMRs – show that they are the customers the nuclear industry has been looking for.

The great drawback of nuclear from the tech companies’ standpoint is how long it takes to bring new capacity online, in a fast-moving market where the shortest possible “time to power” is important. But the Biden administration and many members of Congress support measures to accelerate nuclear development. Meanwhile, the tech companies are talking about some speedy timetables for SMR deployment, with Google aiming for first power in 2030 and Amazon in the early 2030s.

Since Wood Mackenzie started tracking the SMR market in 2021, the global project pipeline has grown from about 13 GW to about 22 GW. The announcements made by Google and Amazon this week could add 1.8 GW more to that total. Amazon’s 2039 goal for SMR deployment would add roughly 3.7 GW more.

The Biden administration has been supporting a variety of SMR designs under its Advanced Reactor Demonstration Program, and has also been moving to ease concerns over the availability of the high-assay low-enriched uranium (HALEU) that many Generation IV reactors, including X-energy’s Xe-100, need to make their fuel. The Department of Energy this week announced that it had selected four companies to bid for up to US$2.7 billion of work enriching uranium, to create a US supply chain for HALEU.

It is still very early days for SMRs, and the industry has many challenges to overcome before it can achieve liftoff. These innovative reactor designs will face the difficulties that plague all first-of-a-kind projects. But a pathway is emerging for advanced nuclear power to play a greater role in energy supply in the US and globally in the 2030s and beyond.

In brief

Crude prices fell after the Washington Post and other outlets reported that Benjamin Netanyahu, Israel’s prime minister, had told President Joe Biden that his country would not attack Iran’s oil or nuclear infrastructure. On Friday morning, Brent crude was trading at about US$73 a barrel, down from a peak of about US$81 a barrel 11 days ago.

The Biden administration has proposed a deregulation that would accelerate the development of geothermal energy resources. The Bureau of Land Management (BLM) wants to declare what is known as a Categorical Exclusion for many geothermal resource confirmation wells, removing the need to carry out an environmental assessment before being awarded a permit to drill. The BLM said the reforms “could significantly reduce permitting timelines and capital costs of geothermal deployment”.

General Motors is making the largest-ever investment by a US car manufacturer in a lithium carbonate project. It is putting US$625 million in cash and letters of credit into a joint venture with Lithium Americas to develop the Thacker Pass lithium mine in Nevada. The investment comes on top of US$320 million that GM committed to the project in 2023.

Jeff Morrison, a senior vice president at GM, said the company was building a resilient supply chain for battery materials to cut costs, create jobs and deliver better value to customers and investors. GM has agreed to buy 100% of the volumes from phase one of Thacker Pass for 20 years.

Back in March, the US Department of Energy’s Loan Programs Office (LPO) extended a US$2.26 billion conditional commitment to Lithium Americas to help finance the construction of a lithium carbonate processing plant at the Thacker Pass site. The LPO said developing the project would strengthen domestic battery supply chains and “reduce reliance on unreliable foreign sources” of lithium carbonate.

The US Supreme Court has declined to block the implementation of the Biden administration’s new rules for greenhouse gas emissions from power plants but showed signs that the regulations could be struck down in the future. West Virginia brought a legal action backed by 21 other states, along with allies including power co-operatives, over the new emissions standards for fossil fuel power generation that the Environmental Protection Agency announced in April.

Two Supreme Court justices, Brett Kavanaugh and Neil Gorsuch, indicated in the court order that they thought West Virginia and its allies had “shown a strong likelihood of success on the merits” for at least some of their challenges to the EPA’s regulations. But they joined the majority on the court in deciding that there was no need to block the new rules at this stage. Companies do not have to start compliance work until June 2025, so they are unlikely to suffer irreparable damage before the DC Circuit Court of Appeals has had a chance to rule on the substantive issues in the case.

Other views

Future energy: biomethane’s time has come – Simon Flowers, Kateryna Filippenko and Gavin Thompson

Growing pains: how corporate power and renewables strategies are adapting to energy transition challenges – Norman Valentine

Four key takeaways from LME 2024 – Nick Pickens and others

Renewable-battery hybrid power plants in congested electricity markets: Implications for plant configuration – James Hyungkwan Kim and others

Can energy-rich Pennsylvania chart a path toward decarbonization? – Julian Spector

Geospatial variation in carbon accounting of hydrogen production and implications for the US Inflation Reduction Act – Valeria Vallejo, Quoc Nguyen and Arvind Ravikumar

A roadmap for designing hydrogen projects for 45V compliance – Nathan Iyer and Taylor Krause

Powering up the Global South: The cleantech path to growth – Vikram Singh and Kingsmill Bond

Quote of the week

“Demand is a challenge for all climate technologies… But it’s particularly acute for carbon removal and direct air capture, because it’s a public good. We’re producing a waste management service that no one currently has to pay for, and that makes commercializing this particularly difficult.”

Giana Amador, executive director of the Carbon Removal Alliance, an industry group, explained the fundamental challenge the sector faces. In last week’s Energy Pulse, I talked about some of the issues facing the CCUS sector in general, and what it needs to accelerate its growth.

Chart of the week

This chart is a history lesson, showing the prevailing trends in electricity demand growth in the US, right up to the point where they are starting to change. It comes from our new Horizons report: Gridlock – the demand dilemma facing the US power industry, written by Chris Seiple, Wood Mackenzie’s vice chairman for the energy transition and power and renewables.

The two lines show US electricity demand and real GDP since 1950. You can see that there was a gradual decoupling over the decades, with electricity demand lagging behind GDP growth, before a complete breakdown of the relationship in the 2000s. Over the past decade, electricity demand in the US has been broadly flat, not growing at all. But as Seiple puts it: “Burgeoning data-centre development, a resurgence in energy-intensive US manufacturing, and greater transport and heating electrification will result in electricity demand growth not seen since the 1990s.”

Get The Inside Track

Ed Crooks’ Energy Pulse is featured in our weekly newsletter, alongside more news and views from our global energy and natural resources experts. Sign up today via the form at the top of the page to ensure you don’t miss a thing.