Is AI's Future Tied to Natural Gas, Coal, and Nuclear Energy?

Last week, one of the world's largest energy conferences, CERAWeek, hosted by S&P Global, took place in Houston, TexasThis event saw energy executives from around the globe sharing their insights on the transformative potential of artificial intelligence (AI) in the energy sectorThe discussions revealed that AI is not only poised to enhance exploration, drilling, and extraction techniques, but it is also generating a significant new demand in the energy industry.

During the proceedings, attendees reached a common understanding: the ultimate goal of AI lies in the realm of energyThis realization is rapidly becoming a societal consensus, demonstrating the interconnectedness between AI and energy.

But what is the essence of AI's relationship with energy? To comprehend this, one should consider three fundamental factors: computational power, algorithms, and data volume.

Computational power serves as the foundation of AI

This power primarily derives from advanced chip technology, which is complex due to its intricate design and manufacturing processesAlthough the challenges surrounding chip development are substantial, they are not insurmountable in the long run.

Next, we have algorithms, where approximately 20% rely on breakthroughs in underlying scientific theoriesWhile financial investment and technical expertise from engineers and programmers play roles, remarkable progress requires foundational shifts in scienceThe remaining 80% of algorithm development hinges on machine learning, which thrives on data inputThis pathway can evolve over time, leading to substantial advancements.

Lastly, we come to the matter of data volumeAI, like any powerful engine, demands fuel—in this case, dataThe quantity of available data relates closely to several factors, including a country's population size and economic depth

• VIX Calms Down: Is the Worst Over for U.S. Stocks?
• 50% Tariff Hike: A Heavy Toll on Chinese Firms in Mexico
• Alibaba Sees $40 Billion Quarterly Revenue Surge
• A Prosperous Landscape of Global Assets
• Nvidia's Market Value Surpasses $2 Trillion

Nations with substantial populations naturally accumulate more data, accelerating AI’s learning and improvementMoreover, the variety of real-world scenarios comes into playFor instance, while some economically advanced countries like Japan have robust offline retail markets, their underdeveloped online economies result in smaller data pools compared to countries like China.

A closer look reveals that the real, long-term competition in AI boils down to which entities can amass larger data quantitiesWhen examining the underlying layer of this data competition, it becomes evident that electricity is at the coreAs AI demand surges, the race for data translates into a struggle for energy resources.

John Ketchum, CEO of NextEra Energy, remarked that despite a plateau in electricity demand over the past decade, the projected growth rate for the next five years has dramatically increased by 81%. Following years of stagnation, American utility companies are revising their electricity consumption forecasts to unprecedented levels

Reports from Grid Strategies indicate that these revised forecasts have doubled since last year.

Emerging generative AI technologies require vast amounts of electricityFor example, OpenAI's chatbot, ChatGPT, consumes over 500,000 kilowatt-hours daily to manage around 200 million user requests, which is equivalent to more than 17,000 times the daily electricity consumption of an average American householdMoreover, Toby Rice, CEO of the largest natural gas producer in the U.S., cites predictions that by 2030, AI's electricity consumption could surpass that of household usage.

In China, energy consumption has followed a similar trajectoryLast year, the country consumed 9 trillion kilowatt-hours of electricity, reflecting a 7% increase year over yearIn the first two months of this year, power generation accelerated further, displaying an 8.3% increase, which exceeds the target growth of the GDP.

This points to a fundamental truth: to cultivate more robust AI, there needs to be a significant supply of data and abundant electric energy resources

This relationship underscores the urgency of building a more powerful and efficient energy infrastructure.

While it's widely accepted that AI requires significant power to operate, the debate over the sources of this electricity remains contentiousErnest Moniz, a former Secretary of Energy in the U.S., suggested that utility companies may need to become increasingly reliant on fossil fuels such as natural gas and coal, alongside nuclear powerThis might even necessitate the construction of new gas plants to meet the soaring demand.

For many tech companies with ethereal low-carbon emission strategies, solar and wind energy are the preferred renewable sourcesWhile these clean energy options are environmentally friendly, their inherent instability poses challenges, particularly for data centers that prioritize energy reliabilityUnpredictable power outages can prove catastrophic for tech infrastructure.

This concern has led to renewed interest in nuclear energy

However, the establishment of large-scale nuclear plants is not only costly but also time-consuming, taking somewhere between seven to ten years to completeThe technology sector, characterized by its need for rapid innovation, may find such timelines untenableThus, natural gas stands out as a viable option, although it carries significant costs and carbon emissions.

It's crucial to understand that the future of energy consumption cannot rely entirely on nonrenewable resources like coal, oil, and natural gasAI’s computational demands require constant and reliable electricity supply, a lesson highlighted by electricity shortages faced by countries with extensive data centers, like Ireland, demonstrating the elevated requirements for efficient power networks in the AI era.

As China emerges as the world's leading energy producer, it has established a diverse and clean energy supply system, generating nearly 9 trillion kilowatt-hours last year, approximately double that of the U.S

Within the solar energy sector, China leads globally, having transitioned from a reliance on foreign technologies, markets, and equipment to holding the world's largest market share and the most advanced technology in this industry.

Solar power has rapidly ascended to become China’s second-largest energy source after coal, surpassing hydropowerFrom an industrial perspective, storage technology also plays a critical roleWhile storage systems do not generate energy, they are essential for overcoming intermittency challenges posed by renewable sources like solar and windThis makes storage solutions a critical cog in the energy industry’s machinery, as without effective storage, the potential of renewable energy remains significantly unutilized.

Storage technology is evidently a linchpin in the green energy supply chainIn 2020, China announced its ambitious goals to peak carbon emissions by 2030 and achieve carbon neutrality by 2060, a vision commonly referred to as the “3060 strategy” in the renewable energy sector