Progress in the energy transition comes from practice: Practical solutions for the current energy moment.

Preface

(Cheng Maiyue: Consultant for GFM's " AI Energy" program, former RMI partner, and US power expert)

The energy transition has reached a point where it can no longer remain at the level of vision, slogans, or the worship of a single technology. What truly determines the success or failure of the transition is whether a system can be designed, built, financed, and operate stably in the market and the power grid. The value of RMI's " Progress Is Built " seminar lies in its ability to bring the energy issue back to the forefront of engineering, economics, and institutions.

Amory Lovins's "flexible energy transition path," proposed fifty years ago, is not primarily about using less energy, but about providing the same or even higher quality energy services through greater efficiency, better design, and less waste. Today, this idea has new practical significance in the AI era. Data centers, computing power expansion, grid connection, natural gas transition, methane control, distributed energy resources, and energy storage systems are collectively constituting a new round of energy stress testing.

Kevin Guo 's summary of the RMI seminar highlighted a key takeaway: energy transition does not happen naturally; progress comes from practice. Future energy competition will not only be about resources, nor simply about technology, but also about design capabilities, institutional capabilities, market organization capabilities, and engineering implementation capabilities.

For GFM 's * AI Energy*, this article is significant not merely for introducing a conference, but for reminding us that energy issues in the AI era have become fundamental questions concerning national capabilities, corporate capabilities, and the resilience of civilization. The future cannot be built through empty talk; it must be constructed.

—————————————————————

(RMI Introduction) RMI, formerly known as the Rocky Mountain Institute, is a non-partisan, non-profit energy transition think tank and action organization founded in 1982 and headquartered in Colorado. With "Energy. Transformed" as its core principle, RMI is dedicated to promoting the transformation of the global energy system towards a cleaner, more efficient, low-carbon, and more resilient direction. Its work focuses on energy efficiency, renewable energy, grid reform, transportation electrification, building energy conservation, industrial carbon reduction, methane control, and electricity demand in the AI era. RMI's distinctive feature is that it goes beyond environmental initiatives, using market, policy, technology, and capital synergies to promote economically viable and implementable energy solutions, following market mechanisms.

(Image caption) Scene from the RMI online seminar " Progress Is Built: Practical Solutions for Today 's Energy Moment ". RMI CEO Jon Creyts , co-founder Amory Lovins , and guests including Vikram Singh , Sarah Ladislaw , and Deborah Gordon participated in the discussion, exchanging ideas on clean energy deployment, energy efficiency, oil and gas emission reduction, AI- powered electricity, distributed energy, and practical solutions for the current energy transition.

Introduction: Energy transition is not just a slogan, but a means of building capacity.

On May 12, 2026 (US time), RMI hosted an online seminar titled "Progress Is Built: Practical Solutions for Today's Energy Moment." The seminar focused on the topic of "Progress in the Energy Transition Comes from Practice: Practical Solutions for Today's Energy Moment."

This is not an energy forum that only talks about vision.

From RMI CEO Jon Creyts' opening remarks, to RMI co-founder Amory Lovins' reflection on fifty years of thought on "flexible energy transition pathways," and then to Sarah Ladislaw, Deborah Gordon, and Vikram Singh's discussions on energy security, oil and gas emissions reduction, AI data centers, distributed energy, and the Global South energy transition, this conference truly addressed a very real issue:

With energy demand accelerating again, AI electricity consumption rising rapidly, geopolitical disturbances continuing, and oil and gas prices fluctuating repeatedly, what will enable the clean energy transition to take off?

RMI's answer is not a single technology, nor a single policy, but a set of solutions that are closer to institutional engineering:

Progress doesn't happen naturally. Progress only truly emerges after it has been designed, built, financed, deployed, regulated, and accepted by the market.

This is especially important for GFM's "AI Energy" section.

Because energy issues in the AI era are no longer just traditional environmental concerns. They are becoming a systemic problem involving computing power, power grids, natural gas, renewable energy, data centers, energy storage, geopolitical security, and industrial competitiveness.

Energy is once again becoming a fundamental variable in national, corporate, and civilizational capabilities.

⸻

(Image caption) Jon Creyts: RMI CEO Jon Creyts hosted and participated in the online seminar " Progress Is Built ," emphasizing that progress in clean energy is not just a slogan, but a practical, deployable, and buildable solution.

Jon Creyts : Real progress is putting feasible solutions into practice.

In his opening remarks, RMI CEO Jon Creyts pointed out that today's discussion is not about abstract energy transition, but rather about:

What is truly happening? What has already been implemented? Which clean energy solutions are being deployed?

He emphasized that the energy transition does not happen at the conceptual level, but rather through concrete improvements in laboratories, businesses, technologies, buildings, industries, transportation, and power systems.

This is crucial.

For many years, the energy transition has been easily placed within the context of "policy objectives" or "climate commitments." However, at this RMI conference, the energy transition was brought back into the language of engineering:

It's not about convincing the world to believe in the future, but about building the infrastructure for the future.

Jon Creyts also expressed his gratitude to RMI's supporters, partners, and donors, noting that the online conference had participants from 44 US states and 28 countries worldwide. This detail itself illustrates that energy transition is no longer a problem for a single country or market, but rather a global institutional issue facing simultaneously.

RMI also mentioned a limited-time matching donation program during the meeting:

Donating $1 will turn into $3.

Behind this is not just fundraising mobilization, but RMI's clear positioning of its own role: as a non-profit organization, it relies on charitable funds to drive market research, policy design, technology deployment and cross-sectoral collaboration.

This is also an often overlooked institutional fact in the energy transition:

Beyond policy, businesses, and capital, non-profit research institutions are becoming an important intermediary force in the clean energy transition.

⸻

(Image caption) Amory Lovins: RMI co-founder Amory Lovins participated in the seminar, reviewing the idea of "flexible energy transition path" over the past fifty years, and sharing his views from the perspectives of efficiency, overall optimization design principles and energy system restructuring.

Amory Lovins : The "flexible energy transition path" concept he proposed fifty years ago still holds true today.

The most important person at this meeting was Amory Lovins, co-founder of RMI.

Jon Creyts specifically mentioned that this year marks the 50th anniversary of the publication of Amory Lovins' important article, "Soft Energy Paths." This article raised a fundamental question against the backdrop of the energy crisis of the 1970s:

We don't consume energy for the sake of consuming energy. What humanity truly needs is not energy itself, but the services it provides.

Hot water, refrigeration, lighting, mobility, comfortable indoor temperature, industrial production, information processing, and quality of life.

This idea, which seems simple today, was revolutionary in the 1970s.

The prevailing thinking at the time was that increased energy demand meant increased energy supply: more coal, more oil, more nuclear power, and more large-scale centralized energy infrastructure.

Amory Lovins' question was:

If we redesign the system from the demand side and provide the same or even better service with less energy, is there another way?

This is the core of the "flexible energy transition path".

It's not anti-technology, it's anti-waste.

It doesn't mean abandoning energy; rather, it means making energy services smarter.

It's not simply about using less energy; it's about reshaping energy demand through design, efficiency, distributed systems, and overall optimization engineering.

Amory noted at the conference that the total energy used to produce every dollar (on a comparable basis) of GDP in the United States in 2023 was about 65% less than in 1975. This figure illustrates the significant role that energy efficiency improvements have played in the energy transition over the past half-century.

But his focus was not on reviewing achievements, but rather on pointing out:

We've only just scratched the surface of efficiency potential.

In other words, the first principle of the energy transition is not "build more first," but rather "ask:

Can we achieve the same or even better functionality with less energy, less materials, and less waste?

This issue is especially important in the AI era.

The increase in electricity demand brought about by AI will lead to new energy pressures if the answer is simply "build more power plants"; however, if the answer is based on Amory Lovins's system holistic optimization design approach, it will lead to a different set of questions:

Can data centers be made more efficient?

Is computing power better allocated?

Can the power grid be made smarter?

Can distributed energy resources take on part of the load?

Can energy storage and demand response reduce peak pressure?

Could AI itself, in turn, optimize the energy system?

This is precisely the main theme that GFM's "AI Energy" has been focusing on for a long time.

⸻

From bananas to buildings: Energy efficiency is not a sacrifice, but better design.

Amory Lovins shared a highly symbolic example during the conference: his residential and office spaces.

He mentioned that his home, located at a high altitude in Colorado, was built in a cold environment and had faced extremely low temperatures. However, the building itself achieved a comfortable winter life without relying on traditional heating systems through passive solar design, overall building engineering, and efficient systems.

He even mentioned that the building's interior is kept warm by passive solar radiation, maintaining a tropical temperature and humidity environment year-round, making it ideal for growing bananas.

This is not an unusual or interesting story.

Its institutional significance lies in:

Truly advanced energy efficiency is not about enduring the cold, but about maintaining comfort without the need for traditional heating systems.

It's not about lowering the quality of life, but about designing the building itself into part of the energy system.

Amory noted that this type of holistic optimization design can eliminate the need for boilers for space and water heating, and also significantly reduce electricity demand. Its core is not a single device, but the entire building system:

Windows, walls, lighting, insulation, airflow, materials, equipment, and usage are designed together, rather than treated separately.

This is the integrated design principle that RMI has always emphasized.

In traditional engineering, each part is optimized separately, which often leads to overall inefficiency; while overall optimization design requires a system-wide rearrangement, so that improvements in one part reduce the demand for another.

This also applies to AI data centers.

If you only look at the chip, you'll think the energy problem is a chip efficiency problem; if you only look at the power grid, you'll think it's a power supply problem; if you only look at cooling, you'll think it's an air conditioning problem.

However, the real energy problem for AI is a holistic systemic issue involving chips, servers, data centers, cooling, power grids, electricity markets, energy storage, scheduling, and site selection.

This is precisely why Amory Lovins' ideas, published 50 years ago, are still valuable today.

⸻

Industry, automobiles, and pipelines: The cheapest energy is energy saved through increased efficiency.

Amory Lovins also mentioned several practical examples.

include:

Significant energy and capital costs are saved in industrial projects in China;

Improving safety and working conditions while saving energy in large-scale mining projects;

Collaborating with automakers to drive lighter, more efficient vehicle designs;

In industrial pumping systems, pumping energy consumption can be significantly reduced by shortening and redesigning pipelines.

The most noteworthy example is the pipeline and pumping system.

Amory points out that a large amount of electricity worldwide is used to drive motors, and many of these motors are used to pump fluids. If pipes are designed to be too long, too narrow, or have too many bends, they can create significant resistance, causing pumps to consume more electricity.

If the pipes are made shorter, straighter, and thicker from the design stage, energy consumption can be reduced significantly.

This may not sound like dazzling high technology, but it could potentially save a huge amount of electricity.

This aligns perfectly with the fundamental principles of RMI:

The energy revolution comes not only from new technologies, but also from redesign, fundamentally optimizing old systems that are obscured by habit.

This is even more important in the age of AI.

Many data centers waste energy not in the most obvious places, but in system design, cooling efficiency, backup power, load management, grid connection, peak-valley mismatch, and equipment utilization.

A true energy transition is not just about chasing the next revolutionary technology, but also about redesigning every wasteful step in the process.

⸻

(Image caption) Sarah Ladislaw speaks at the RMI online seminar, discussing energy security, policy transformation, and the practical path for current clean energy deployment.

Sarah Ladislaw : Today is not a return to the 1970s , but rather the beginning of a more complex era of energy security.

Sarah Ladislaw, a senior advisor at RMI, compared today's situation to the energy crisis of the 1970s during the discussion.

She pointed out that there are indeed many situations today that are similar to those of the 1970s: geopolitical shocks, energy price volatility, supply security anxieties, and accelerated policy responses.

But the difference today is that the energy transition is no longer just a supply adjustment under the oil crisis, but involves:

Energy security;

Industrial competitiveness;

Climate governance;

Infrastructure construction;

Speed of technology deployment;

Electricity demand is growing rapidly;

AI, data centers, and new industrial electricity demand.

Sarah's core argument is:

Energy efficiency doesn't mean we shouldn't build anything, but rather that every construction project should be smarter, more efficient, and have greater systemic value.

She cautioned that many people mistakenly believe that energy efficiency means "don't build anything anymore." However, the world currently needs a significant amount of new infrastructure, including power grids, renewable energy sources, energy storage, data centers, industrial facilities, and charging infrastructure.

The question is not whether to build it, but how to build it.

If we continue to build according to the inefficient logic of the old system, we will create new energy bottlenecks; if we incorporate efficiency, flexibility, demand response and system design into the initial planning, we can complete a larger-scale energy service with fewer resources.

This is especially crucial for AI.

The most common phrase in the AI industry right now is: speed to power, meaning faster access to electricity.

Data centers need rapid power access, and businesses want to deploy computing power as quickly as possible. However, Sarah warns that switching to natural gas or other fossil fuels solely for speed could create new lock-in effects.

The directions she proposed included:

Power supply and load share the same address;

Combining renewable energy with battery storage can rapidly create new power sources.

Make full use of the existing power grid supply capacity;

Consider new capacity at the regional planning level;

View data centers as part of the energy system, not as isolated users.

This is one of the key points that GFM's "AI Energy" should capture most:

AI is not simply a new industry that consumes energy. AI data centers themselves must become an integral part of the energy system.

⸻

(Image caption) Deborah Gordon: Deborah Gordon participated in the RMI Energy Transition Seminar, focusing on institutional challenges in the oil and gas industry, methane emissions, energy waste, and clean energy governance.

Deborah Gordon : The biggest injustice is that the oil and gas industry is still wasting a lot of natural gas.

Deborah Gordon of RMI is in charge of oil and gas solutions. She raised a very important point during the meeting:

The world wastes an enormous amount of natural gas every year, and this waste could have been avoided.

She pointed out that the current oil and gas industry, which values natural gas based on oil, results in significant natural gas leaks, venting, and operational waste, ignoring the inherent value of natural gas. If natural gas were viewed as a product with intrinsic value, like oil, it could potentially change the phenomena of waste and massive leaks.

This viewpoint is crucial.

In the energy transition, the public often focuses on solar energy, wind energy, electric vehicles, nuclear power and energy storage, but the waste and emissions of the oil and gas industry itself are also a core battleground for energy governance.

Deborah Gordon emphasized that methane control is one of the most important climate actions of the next decade.

Methane originates from oil and gas, coal, agriculture, and waste. Compared to carbon dioxide, methane has a stronger warming effect in the short term. Therefore, rapidly reducing methane emissions is a key means of mitigating climate risks in the near term.

She also pointed out that satellites, monitors, and data technologies are now available to more accurately track methane emission sources. RMI is also promoting related technologies and platforms to identify waste in high-emission facilities and oil and gas systems worldwide.

Here is a direction that GFM should pay special attention to:

Energy transition is not just about building new energy sources, but also about stopping the waste in old energy systems.

If the oil and gas industry itself can reduce leaks, electrify equipment, reduce combustion venting, and use cleaner electricity to drive production, then it can also significantly reduce emissions during the transition period.

Deborah pointed out that the oil and gas industry should electrify its operations. Oil and gas companies rely heavily on oil and gas to power drilling, compression, pumping, and processing equipment. Replacing some fuel and gas-fired equipment with electric equipment, coupled with low-carbon electricity, would significantly improve energy efficiency and drastically reduce operating costs and emissions.

This raises a systemic question for traditional energy companies:

If oil and gas companies claim to be energy companies, not just fossil fuel companies, are they willing to start by improving the efficiency, electrification, and methane control of their own production systems?

⸻

(Image caption) Vikram Singh: RMI online seminar guest Vikram Singh attends the discussion " Progress Is Built: Practical Solutions for Today 's Energy Moment ".

Vikram Singh : The problem isn't a lack of ideas, but rather the progress in turning them into a real-world phenomenon.

In the discussion, Vikram Singh built upon the ideas of Jon Creyts and Amory Lovins, raising a practical question:

Many ideas about energy efficiency, system design, and clean technologies have existed for years. The real question is how to translate these ideas into real progress in today's policy, market, and geopolitical environment?

This is the core position of RMI as an institution.

It doesn't just propose theories or make simple initiatives; it promotes implementation across markets, policies, businesses, and projects.

Vikram's question brought the discussion to the present:

Against the backdrop of increasing energy demand, trade tensions, geopolitical volatility, and rising energy security pressures, how can the world build a clean, reliable, and affordable energy system more quickly?

This is precisely the question that GFM's "AI Energy" journal needs to explore in the long term.

Because the energy crisis cannot be solved by just one technology.

it takes:

Market incentives;

Policy reform;

Power grid planning;

Capital investment;

Enterprises' responsibility for electricity consumption;

Data center site selection strategy;

Renewable energy and energy storage are integrated;

Emissions reduction for old energy systems;

Improved efficiency on the demand side.

This is not a breakthrough in a single area, but a systemic project.

⸻

AI Data Centers: A New Stress Test for the Clean Energy Transition

AI was explicitly included in the energy discussion at this meeting.

Sarah Ladislaw answered a question about how the AI industry uses renewable energy and batteries. She pointed out that AI companies are currently most concerned with obtaining electricity quickly, but this does not mean they can only rely on natural gas or fossil fuels.

She mentioned that colleagues at RMI are researching co-location, generation and load coupling, and grid interconnection issues. The idea is to add renewable energy and battery storage near existing power generation assets to solve the problem of data center grid connection difficulties.

Here is a very important institutional judgment:

AI data centers should not see themselves merely as electricity consumers, but rather as an integral part of energy system design.

If AI companies are only focused on getting electricity as quickly as possible, they could drive a new wave of natural gas infrastructure expansion; if they are willing to collaborate with power grids, renewable energy sources, energy storage, and regional planning, they could become a significant force in accelerating energy system upgrades.

The impact of AI on the energy system is mainly reflected in three aspects:

First, the new electricity demand is huge and concentrated in specific areas.

Second, data centers require high reliability, low interruption risk, and fast access.

Third, the AI industry has strong capital capabilities, enough to influence the direction of investment in the power market and energy infrastructure.

Therefore, AI companies face a new responsibility:

It's not enough to just say you need electricity; you also need to explain whether your electricity usage helps the entire system become cleaner, more reliable, and more efficient.

This is the judgment that GFM's "AI Energy" column should establish most:

The energy issue for AI is not a logistical problem for AI companies, but a cutting-edge issue for the next generation of energy systems.

⸻

Distributed energy: not small and beautiful, but a new foundation for resilience and efficiency

Later in the meeting, Amory Lovins and Jon Creyts discussed distributed renewable energy and DERs, or Distributed Energy Resources.

Amory emphasizes that the value of distributed energy should not be simply compared to centralized systems in terms of cost. Distributed energy is more resilient because many power outages originate in the power grid, especially the distribution network.

If homes, communities, buildings, and businesses can install solar power, batteries, smart inverters, and local microgrids, they can maintain basic power supply and improve energy security during grid failures.

Jon Creyts also pointed out that low-cost renewable energy, batteries, and distributed technologies are creating tremendous opportunities. If distributed energy can be combined with data, AI, aggregated scheduling, and market mechanisms, it won't just benefit individual households or buildings, but rather allow thousands of distributed assets to collectively become grid resources.

This is an important direction for future energy systems:

A single rooftop solar panel is a device;

Millions of rooftop solar panels plus batteries form a system;

With the addition of AI scheduling and market mechanisms, it could become a new type of virtual power plant.

This also has important implications for AI energy.

In the future, not all electricity will have to come from large, centralized power plants. Distributed energy resources, energy storage, demand response, and smart dispatch can work together to reduce peak-hour pressure and make the grid more flexible.

This also carries another layer of meaning to the saying "progress is built up":

We need to build not only large projects, but also small units;

We must not only build supply systems, but also systems for scheduling.

We need to build not only power sources, but also resilience.

⸻

Global South and India: Energy Transition is Also a Development Issue

In response to questions about the Middle East crisis, oil and gas supplies, and energy security in countries like India, Jon Creyts pointed out that RMI has been working in India for many years, driving the transformation of electrification, vehicles, energy systems, and supply chains.

The significance of this section of the system lies in:

The energy transition cannot be understood solely from the perspective of Europe and the United States.

For India, Southeast Asia, Africa, and the Global South, energy is simultaneously a development issue, an import dependency issue, a foreign exchange issue, an industrialization issue, and a livelihood issue.

If renewable energy, electric transportation, and energy efficiency can reduce dependence on oil and gas imports, this is not just a climate policy, but also an economic security policy.

Sarah also mentioned that many countries are now rethinking energy security. Energy crises are no longer seen as isolated incidents, but rather as an increasingly frequent source of stress.

Therefore, energy policies in various countries are shifting from simply "emissions reduction" to more complex objectives:

Reduce import dependence;

Enhance local supply capacity;

Enhance industrial competitiveness;

Control energy prices;

Reduce emissions;

Improve system resilience.

This is also of great significance to Chinese companies, Asian companies, and Chinese capital.

The future energy market is not simply a green story, but a new core of industrial chain restructuring, capital reallocation, and geopolitical competition.

⸻

GFM's institutional assessment: The next stage of the energy transition is a competition of design and institutional capabilities.

The biggest takeaway from this RMI conference for GFM's "AI Energy" column wasn't a particular technology, but rather a higher-level judgment:

The next stage of the energy transition will not only be a competition of technology, but also a competition of design and institutional capabilities.

Whoever can better design buildings, data centers, industrial systems, power grids, and distributed energy can generate greater value with fewer resources.

Whoever can establish market mechanisms, policy frameworks, financing tools, and regulatory standards faster will be able to bring clean energy to fruition more quickly.

This is especially true in the age of AI.

Because AI is both a new pressure on energy demand and a new tool for optimizing energy systems.

It may generate significant electricity demand, and it may also help with grid scheduling, load forecasting, energy storage management, data center cooling optimization, and industrial efficiency improvement.

The problem is:

Will AI become a burden on the energy system, or will it become the intelligent layer of the energy system?

It depends on the design, and it also depends on the system.

⸻

True energy progress never happens naturally.

The theme of this RMI seminar is simple:

Progress is built.

Progress in the energy transition is built.

In the context of energy transition, this statement implies three things.

First, clean energy won't succeed through slogans, but through practical deployment, cost reduction, engineering capabilities, and market acceptance.

Second, energy efficiency is not about cutting back on food and clothing, but about eliminating waste through better design and making life better.

Third, the energy problem in the AI era cannot be solved by simply increasing supply. Instead, demand-side efficiency, distributed energy, energy storage, oil and gas emission reduction, power grid planning, and data center responsibilities should be addressed within the same system.

The question Amory Lovins posed fifty years ago still holds true today:

What we really need is not energy itself, but energy services.

In the AI era, this statement can be further rewritten:

What we really need is not more electricity consumption, but more efficient, reliable, clean, and intelligent energy services to support the next stage of human productivity.

This is why GFM's AI Energy focuses on the RMI conference.

Because it reminds us:

Energy transition is not a distant vision.

It is today's buildings, today's power grids, today's data centers, today's oil and gas facilities, and today's policy choices.

The future cannot be built on empty talk.

The future is built.

Disclaimer:
This article is a compilation and analysis of publicly available content from the RMI online seminar, compiled by GFM's " AI Energy" column . It is for research, exchange, and informational purposes only and does not constitute investment, energy project decision-making, legal, policy, or technical advisory advice. The views expressed in this article do not represent the official position of RMI , nor do they represent the complete original intent of any participating speakers. Readers should consult professional organizations for specific project, investment, or policy judgments.