Skip to Content Skip to Menu Skip to Footer

By Judson Jacobs and Peter Gardett

Highlights

Energy companies are poised to be early leaders in leveraging recent advancements in AI, a natural evolution for an industry rooted in data, analytics, engineering and complex processes.

Energy transition dynamics may incentivize, or even require, companies to test and use AI approaches to manage the increased complexity created by more renewable generation in the energy mix and accelerating technology and regulatory developments.

Many companies have already made rapid strides, establishing dedicated data science groups and leveraging machine learning. One of the main targets for AI to date has been boosting operating efficiencies, with S&P Global Commodity Insights documenting performance improvements in the 10%-25% range.

Outsized investor appetite for evidence of AI in company strategies will continue to give energy companies financial incentives to embrace this emerging technology in their operations. But there is mixed evidence that individual operators will be able to deploy the technology in a breakout way that redefines the industry, even as almost every role and responsibility within energy companies is remade by AI.

Gauging a surge in AI investment

A breakout in public company valuation and private company investment for business strategies linked to AI is creating broad incentives for energy market participants to rework their operations around AI implementation.

The appeal of technology additions as a productivity enhancer for existing revenue models and as a growth category for emerging revenue models is familiar to energy investors and energy company leadership. In an era where technology companies are disproportionately attracting capital, it can be difficult to remember that energy and industrials companies are also fundamentally technology driven. The origins of the modern energy sector in high-grade engineering, in repeatable processes underlying value extraction and in constant technological advancement mean the industry is primed to implement and leverage AI strategies.

The financial appeal in recasting businesses as AI-enabled is clear. While the eye-popping quintupling of chipmaker Nvidia Corp.’s valuation has attracted the most media attention, every business model closely tied to AI implementation has received a valuation premium from investors in a manner consistent enough to redefine company operator incentives. Valuations for AI-exposed businesses are consistently higher than similar businesses without AI market exposure (see chart).


AI is often too loosely defined and, for a technology perceived as “changing everything,” can be too loosely applied in businesses. The momentous release of a public large language model by Open AI in early 2023 highlighted the capabilities of AI in several professions. However, LLMs and the generative AI features that make these tools so impressive for human users have limited immediate applications for the energy sector overall, regardless of how much they could transform individual roles and workflows.

S&P Global Commodity Insights

AI is often too loosely defined and, for a technology perceived as “changing everything,” can be too loosely applied in businesses.

Judson Jacobs

Executive Director, Energy Technology and Innovation


Fitting AI capabilities to energy needs

Investor enthusiasm for AI will continue to pull company leadership and market attention toward implementing generative AI — including LLMs — in almost any function. However, the more “traditional” AI advancements in applying machine learning and expert systems to physical and market infrastructure are what could revolutionize energy companies.

The need for computer-assisted mechanisms and processes in energy markets to handle high levels of complexity on behalf of human decision-makers will become particularly acute as the energy transition proceeds. Key components of the energy transition impose additional potential for volatility, with more intermittent renewable energy on grids, more changes to market-setting regulations, and increased sensitivity regarding the security and reliability of energy supply. 

In the design and rollout of new products — particularly high-tech and often digitally coordinated energy infrastructure such as large-scale batteries — the role of AI in everything from initial design conception to timing of daily charging operations is increasingly assumed as underway by investors. This means project developers, project financiers and energy company operators need to be ready to explain AI deployment trade-offs and decisions.

Many of these AI advancements will not necessarily entail the kind of generative, self-originated processes that have so surprised the business world with the launch and rapid uptake of ChatGPT and other LLMs. Generative AI is, in many ways, not as much of a leap in functionality as an acceleration in the application of existing systems for managing large data. These developments were expressed first through expert systems and then via machine learning and data science, and they were adopted by energy companies and investors without significant disruption. 

S&P Global Commodity Insights

Generative AI is, in many ways, not as much of a leap in functionality as an acceleration in the application of existing systems for managing large data.

Judson Jacobs

Executive Director, Energy Technology and Innovation

 

 


In many cases, applications of AI in the energy sector — from improved battery-life performance and better oil refinery design to more targeted energy market financial hedging — will be inputs in other energy market functions rather than a wholesale replacement of an activity currently done by humans. 

As an operational matter, the need to leverage AI to stay competitive will almost certainly accelerate the trend of enhanced digital monitoring and “digital twinning” of physical energy assets. The deployment of monitoring devices on physical infrastructure allows for the creation of so-called digital twins to real-world assets, letting engineers and operators monitor and control those assets remotely, as well as permitting digital experimentation that would be too costly or dangerous with a physical energy asset. This will transform physical jobs such as inspecting and maintaining infrastructure into increasingly digitally interfaced and office-based jobs. 

Energy companies have been working for years to better integrate digital networks and the resulting data insights into their operations, with mixed results. As firms evaluate ways to more actively deploy capital expenditure against AI strategies and to garner investor attention amid the broader AI revolution, energy companies are following a handful of strategies to incorporate AI into their businesses.

Building on AI progress at energy companies

Data-driven approaches to improving the efficiency and outcomes of analyses are nothing new to energy professionals, particularly individuals with high exposure to large data sets. What has changed over the past decade-plus, however, is the establishment of formal groups to advance overall company capabilities in this rapidly evolving technology area. These dedicated groups are proving invaluable in accelerating the uptake of AI solutions across their organizations, with the aim of moving the needle on overall corporate performance. They also serve as a natural landing spot for the generative AI concepts (and whatever might come next) that offer such great promise.

This centralization of AI efforts is allowing energy companies to take a more structured approach to building data science capability, including via the following:

  • Establishing formal relationships with a diverse set of partners (e.g., computing infrastructure providers, platform vendors and technology startups) that augment internal capabilities and bring in new ideas from outside the industry.

  • Assessing build versus buy criteria and decisions and establishing related intellectual property protection protocols.

  • Positioning AI resources optimally within the organization to ensure effective engagement with the business during solution development, which in turn leads to greater workforce acceptance of the resulting products.


Perhaps the most important responsibility of these groups is developing a prioritized portfolio of solutions that best meet companies’ needs — delivering near-term “wins” that build momentum for the technology while also pursuing higher-risk opportunities that can transform the business. Activities tend to fall into three broad buckets:

  • Improving efficiency. AI in energy received a big boost during the oil price downturn of 2014–2015 that prompted struggling unconventional operators to be among the first to launch formal data science programs, targeting lower well costs and higher productivities. The concepts these programs pursued — applying AI algorithms to learn and automate repetitive tasks, to predict and avert equipment failures, to optimize supply chain and logistics networks, and to assist in other efficiency-boosting activities — are spreading rapidly. Deployed effectively, S&P Global documents operational performance improvements in the 10%-25% range.

  • Managing large and complex systems. The expanding energy value chains of the energy transition (e.g., renewables-heavy power grids and green hydrogen networks) are proving difficult to manage and optimize using traditional means. Applying AI solutions in conjunction with other digital concepts (e.g., digital twins) allows operators to calculate and then autonomously implement optimal configurations, driving further efficiency improvements and capacity expansions. 

  • Accelerating the innovation cycle. High costs are impeding the uptake of certain clean energy technology segments (e.g., carbon capture, utilization and storage, and small modular nuclear reactors) and therefore creating the imperative to reduce the learning curve more quickly in future projects. AI can quickly identify suboptimal design features once plants are operational and rectify them in the next project iteration. Additionally, a machine learning-enabled, drug discovery-like approach can speed the search for new materials to advance the energy transition, such as higher-efficiency carbon capture materials.

Looking forward

The energy sector continues to seek the right approach to exploiting AI capabilities within its businesses, balancing centralization with innovation at the edges, open innovation with proprietary technology development, and incremental gains with game-changing solutions. Companies that established formal data science groups and that are actively engaging external partners ahead of the curve appear to be getting it right, as these are the businesses that identified generative AI’s potential during its early stages and have been running agents for several years. Those with less structured approaches have only recently begun to identify use cases and are now working with their legal and IT organizations to gain approval to proceed with pilots. These strategies will continue to unfold in the coming years. 


Learn more

 


Next Article:
The materials transition: Ensuring we build with low-carbon materials

This article was authored by a cross-section of representatives from S&P Global and in certain circumstances external guest authors. The views expressed are those of the authors and do not necessarily reflect the views or positions of any entities they represent and are not necessarily reflected in the products and services those entities offer. This research is a publication of S&P Global and does not comment on current or future credit ratings or credit rating methodologies.

Content Type

Special Reports

Language

English
ai_footer