Blog

Electric motors were selected, in lieu of industry-standard gas turbines, to power the main refrigeration compressors at three of the four new LNG projects that took FID in 2024. Hence is a major change underway in the LNG industry? This 13-page report covers the costs of e-LNG, advantages and challenges, and who benefits from shifting capex.

Another capex cycle appears to be underway in global LNG, as the pace of new LNG project additions accelerates from 10MTpa/year in 2022-25 to 55MTpa/year on average from 2027-30. Annual LNG capex forecasts, and where the money goes, are thus summarized on pages 2-3.

Yet a shift also appears to be underway on the equipment side: electric motors are gaining share over gas turbines, to power the compression trains at LNG plants. An overview of how an LNG plant works, covering the Joule Thomson effect, LNG plant compressors’ energy use (in kWh/ton) and CO2 intensity (in tons/ton) are given on page 4.

Electric motors were selected, in lieu of industry-standard gas turbines, to power the main refrigeration compressors at three of the four new LNG projects that took FID in 2024. Specific examples of electric LNG plants (eLNG), and who won the order flow, are given on pages 5-6.

Economic costs of eLNG are modeled on page 7 (in $/mcf terms), with a sensitivity to CO2 costs, power prices and transmission distances on page 8; and a breakdown of the capex (in $/Tpa terms) on page 10, including the leading companies winning order flow on page 11.

Advantages and challenges of eLNG are discussed on page 9 and pages 12-13. Some proponents claim higher reliability, flexibility, lower maintenance costs, better start-up times and the chance to lower LNG emissions 50-90% below typical facilities. But past projects to electrify LNG plant compressors have not always performed perfectly.

Overall, a new LNG cycle is set to be a major theme for 2025, as are power grid bottlenecks, as ever more processes electrify. We will be looking for opportunities across the board in these process technologies.

This 15-page note outlines the largest changes to our long-term energy forecasts in five years. Over this time, we have consistently underestimated both coal and solar. So both are upgraded. But we also show how coal can peak after 2030, based on cost factors alone. Global gas is seen rising from 400bcfd in 2023 to 600bcfd in 2050.

Peak coal demand is a necessity for net zero scenarios. But it is an embarrassment for net zero modelers.

Our own numbers from 2022, for example, hoped global coal would peak at 8.2 GTpa in that year, run sideways for 1-2 years, then fall off a cliff by 2050. Yet global coal use hit 8.8GTpa in 2024, and has been revised upwards in five of the past six years (page 2).

There is a famous Albert Einstein quote about how insanity is doing the same thing over and over again and expecting different results. Thus, it is starting to feel insane to us, that every year, we update our models, smudge up prior year coal consumption in China and India, then insouciantly claim this year must be the year these countries will decide to stomach $50-80/ton CO2 abatement costs for gas switching (page 3).

Hence this 15-page note has followed some sober reflection in December-2024, as we updated all our energy supply-demand models for the past year. As a result, we are making some of the largest changes in our energy models since starting TSE, with materially more coal and solar in the long-term mix. But global coal use really and truly can peak.

How coal can peak is that coal costs will rise, solar costs will fall and eventually gas will be more competitive than coal for backing up the solar. The rising cost structure of China’s listed coal producers and broader coal-mining industry is shown on pages 4-6.

Solar capacity additions are the other line item in global energy balances that have consistently been upgraded, with every year’s trajectory seeming to defy even the most bullish forecasts from the previous year, for a decade. Hence this note contains large upgrades to our solar forecasts, which unlock fascinating new global energy demand, as shown on pages 7-9.

The key chart in this note is on page 10, which contrasts electricity costs in Asia from 2008 to 2050, across coal, gas and LNG. How coal can peak is that marginal solar has now deflated below marginal coal, coal costs keep rising, and from 2030, imported LNG starts to displace marginal coal in backing up the solar (see page 10).

Large updates to our global energy outlook follow. Coal demand does peak, but our 2050 coal forecasts are revised sharply upwards, especially in China’s energy supply-demand mix. Our forecasts for global electricity supply-demand and solar are also revised sharply upwards. Global gas supply-demand is still seen rising by 50% to 600bcfd by 2050. Key charts are on pages 11-15.

The note also dovetails with our top ten predictions for 2025.

This 11-page report sets out our top ten predictions for 2025, across energy, industrials and climate. Sentiment is shifting. New narratives are emerging for what energy transition is. 2025-30 energy markets look well supplied. The value is in regional arbitrage, volatility, grids, AI and solar.

There is always a delicate balance in outlook notes. Erring on the side of caution yields predictions that are guaranteed to be true, but nevertheless add no value (‘up will be up, and down will be down’). Erring in the other direction can yield very interesting hypotheticals, but they will never happen (e.g., ‘zombie apocalypse’).

Hence this note is our attempt to reflect on everything that happened in 2024, everything we have learned as energy analysts over the past 15-years, and thus make predictions for what will happen in 2025+, which will be at least 50% novel and at least 50% probable.

Each page of the report covers a single one of our ten themes for energy in 2025, a prediction for what we think will happen, and is substantiated with our single best ‘killer chart’.

Four of our themes cover energy supply-demand, especially global energy balances, especially coal, LNG and solar where particularly interesting market trends are afoot.

Three of our themes cover how the narrative is changing in energy transition. ‘Decarbonize the world at any cost’ is starting to feel like a futile folly, a kind of environmental Afghanistan. We argue the narrative will shift sharply in 2025, as new themes emerge.

Our outlooks for AI and power grids in 2025 are on pages 7 and 8, including a preview for how our numbers will change, when we revisit these topics in greater depth in 1Q25.

Our wildcard scenario for 2025 looks at trade tensions, and how they may create not just downside for energy markets, but potentially interesting upside?!

Companies that do well amidst our themes for energy in 2025 will be those that focus on competitiveness, efficiency, low-cost energy, and building out capacity that their customers will pay for, per page 10. Our TSE company database has been updated for 4Q24 and now features 1,700 mentions of 600 core companies.

A Kardashev scale civilization uses all the energy it has available. Hence this 16-page report explores ten futuristic uses for global energy, which could absorb an additional 50,000 TWH pa by 2050 (60% upside), mainly from solar. And does this leap in human progress also allay climate concerns better than pre-existing roadmaps to net zero?

Most long-term energy forecasts simply lack imagination. In particular, most energy transition scenarios leave little room for new demand, which is why AI was a shock in 2024. But what if civilization was capable of harnessing vastly more energy?

The Kardashev scale was proposed by Nikolai Kardashev, in 1964. It measures the technological advancement of a civilization according to the amount of energy it is capable of harnessing and using. Kardashev Level 1.0 equates to a civilization that can use all the available energy on its planet. Currently, the useful energy consumption of all human civilization is equivalent to about 0.01% of the solar energy reaching the Earth’s surface at ground level, as discussed on pages 2-3.

In this note, we will go full sci-fi, and indulge the fantasy of near-infinite energy, e.g., from vast quantities of future solar available at 1c/kWh? How much incremental energy demand might human civilization want? Where could it go? And does this produce better human outcomes than limiting global energy demand in order to reach net zero by 2050?

Incremental demand for living space and material possessions are probably the two most obvious yet boring use cases, with demand sensitivities on pages 4-5.

More interesting and futuristic, however, the bulk of this note explores advanced materials that push the limits of engineering (page 6), an unstoppable rise of AI energy potentially culminating in Matrioshka Brains powered by Dyson Spheres (!) (page 7), a return of supersonic aviation (page 8), aerial vehicles (page 9), greening 1bn acres of desert (page 10), infrastructure projects that transform urban landscapes (page 11), electrochemical DAC to construction materials (page 12) and of course space-faring (page 13).

We propose how low-cost solar would provide the vast majority of the energy needed for these futuristic new energy uses, yet oil runs sideways and gas use still rises, in this future energy system (chart below), based on the economic reasoning on pages 14-16.

We started this note as a science fiction fantasy. But after writing it, we think this kind of energy transition is actually more likely to play out than our last published roadmap to net zero, whose deliverability has recently started to seem less likely.

Absorption chillers perform the thermodynamic alchemy of converting waste heat into coolness. Interestingly, solid oxide fuel cells and absorption chillers may have some of the lowest costs and CO2 for powering and cooling AI data-centers. This 14-page report explores the opportunity, costs and challenges.

Some power generation sources produce both electricity and waste heat. Absorption chillers can convert that waste heat into coolness. Hence could this combination provide both data-center power and data-center cooling, more economically and with lower carbon, than the traditional approach of using electrically-driven HVAC? This question felt interesting to explore in a dedicated research note.

A fascinating avenue to get net zero back on track, more broadly, while also enhancing energy security and competitiveness, would be to capture more waste heat, including by turning heat into coolness, via absorption chillers. Market sizes are quantified on pages 2-3.

How does an absorption chiller work? The four key stages, in the evaporator, absorber, generator and condenser, are described clearly and concisely on pages 4-5.

What does an absorption chiller cost? Capex, opex and total costs of cooling are drawn from our economic model of absorption chillers, in cents per ton-hour and in $/kW-th, and compared with mechanical HVAC equipment on pages 6-8.

Hence how do the costs compare for powering and cooling a data-center using (i) grid power and mechanical HVAC (ii) CCGTs and mechanical HVAC (iii) simple cycle gas turbines and absorption chillers (iv) Solid Oxide Fuel Cells and absorption chillers? The answers on this comparison surprised us, per pages 9-11.

Challenges with fuel cells and absorption chillers should be considered, before getting overly excited, hence some recent successes and issues are summarized on pages 12-13.

Companies producing absorption chillers and solid oxide fuel cells, including our review of Bloom Energy’s patents, are on page 14.

What if achieving Net Zero by 2050 and/or reaching 1.5ºC climate targets now has a <3% chance of success, for reasons that cause decision-makers to backtrack, and instead focus on climate adaptation and broader competitiveness? This 14-page report reviews the challenges. Can our Roadmap to Net Zero be salvaged?

The goal of research is neither to cheerlead for what you want to happen, or to whine about what you don’t want to happen. It should be to predict what will happen. Even when you don’t like the predictions.

Hence every December we have attempted to distil our research from the previous year, into a Roadmap to Net Zero, which suggests the most likely trajectory where the world could reach zero net CO2 emissions by 2050, thereby limiting climate change to 1.5 – 2.0ºC of warming.

Unfortunately, this year, we increasingly fear our Roadmap to Net Zero is not what will happen. The purpose of this note is to explain why.

The first challenge is that we are seeing lower willingness to pay for decarbonization than we expected, per the evidence on pages 5-6.

The second challenge is a more adversarial world, where issues such as defence, self-sufficiency and competitiveness threaten themes such as coal-to-gas switching and climate coordination, per pages 7-11.

The third challenge is slow progress with CCS and CDRs. We find it unlikely that gross emissions will fall below 30GTpa by 2050, but can anything close to 30GTpa be captured and/or offset, per pages 12-13?

Hence our most likely scenario is now for Net Zero to be delayed by 2-3 decades and for 2.5-3ºC of warming by 2100. Around 1.3ºC of this warming has already happened.

What could still salvage a 1.5-2.0ºC Climate Scenario, versus the 2.5-3ºC world that increasingly looks more likely, could be some game-changing technology, emerging at the bottom of the cost curve: AI breakthroughs, thermo-electrics, solar + battery costs collapsing sharply, fusion, electrochemical DAC.

And maybe we should not fixate too much on achieving Net Zero by 2050, or the precise level of warming in 2100, which no one really knows anyway. If you can find good opportunities, which boost competitiveness (and are not overly reliant upon fickle policy support!!), then these are the ways to improve the world’s energy system from the bottom up.

Solar trackers are worth $10bn pa. They typically raise solar revenues by 30%, earn 13% IRRs on their capex costs, and lower LCOEs by 0.4 c/kWh. But these numbers are likely to double, as solar gains share, grids grow more volatile, and AI unlocks further optimizations? This 14-page report explores the theme and who benefits?

A solar module is a 2.7 m2 rectangle, whose internal semiconductors convert incoming electromagnetic radiation into a direct current via the photovoltaic effect. To maximize energy production, ideally, the entire 2.7 m2 rectangle will be pointed directly at the sun and receive full sunlight. But this is challenging as the sun arcs across the sky, tracing a different path every day of the year, and varying with latitude, as shown on page 2.

Solar trackers orient solar modules towards the sun. The market size, key parameters of different systems, and “how solar trackers work” are succinctly explained on pages 3-4.

The energy uplifts from solar trackers have been estimated at 10-50% in different studies. But we can do better than this broad range, and actually calculate both the energy uplift and the revenue uplift from first principles, on pages 5-8.

The economics of solar trackers can therefore be modeled more effectively. Our base case yields 13% IRRs and deflates solar LCOEs by 0.4 c/kWh. We can also model how steepening duck curves, battery co-deployments, and AI optimizations will further improve the case for solar trackers, on pages 9-10.

The solar tracker industry is worth $10bn pa, relatively concentrated, and relatively unusual for a solar supply chain in that it is still dominated by US companies. We discuss key conclusions from our screen of solar tracker companies on pages 11-13.

A key mega-theme that has permeated our 2024 research has been the rise of AI, and the benefits of greater digitization and optimization. It is interesting to end by noting that solar trackers, once again, fit this trend, and amplify demand for sensor equipment.