(Art direction and design incl. AI :Jean-Dominique Lavoix-Carli)
Power for AI Power
In both the United States and China, the exponential growth of the AI sector is putting immense pressure on the energy systems of both major powers. This demand for electricity affects the entire AI production and usage chain.
The increase in usage by individuals, businesses, and government agencies is draining smartphone batteries, and the power consumption of laptops and desktop computers, along with the proliferation of data centers, requires generating quantities of electricity that were never anticipated by the designers and managers of the power grid.
Finally, semiconductor manufacturers and suppliers are facing ever-increasing demands, and their electricity consumption is also skyrocketing. Conversely, the oil and gas and nuclear sectors are investing heavily in AI to optimize their operations, from exploration to production, equipment use, and administrative and financial tasks.
Hélène Lavoix : Portal to AI: Understanding AI and Foreseeing the AI powered world
- RTAS Expertise Highlighted on Italian Television Channel LA7
- The Houthis, Bab el-Mandeb, and Maritime Transport
- The Strange Mystery of Bolivia’s Lithium Reserves
- Beyond Geopolitics: Understanding the Critical Pressures Reshaping Our World.
- WATCH OUT – The Iran War, Saudi Arabia, and the Houthis
- Chronicles of the Hyperwar – AI: AI, the New Energy Frontier (1)
- SIGNAL – The Golden Dome, a New Technological Frontier
However, the AI sector is developing too rapidly compared to the energy sector’s capacity to adapt. Hence the already underway plans to deploy new generations of data centers in orbit, in order to benefit from solar energy and space cooling. This results in a radical transformation of the space, AI, and energy production sectors, a transformation that is also influencing the geopolitical competition between China and the United States.
In this first part, we examine the American dimension of this new AI-energy-space geopolitics. We argue here that this new energy frontier of AI is transforming global geopolitics, giving rise to “spatiopolitics,” whose feedback effects are accelerating further transformations.
I. Energy for AI…
The Great Devourer of Electricity
The International Energy Agency estimates that in 2024, the global electricity consumption of data centers was equivalent to 415 terawatt-hours, or 1.5% of global electricity consumption.
This figure was achieved after five years of a 12% annual increase in electricity consumption due to the rise of AI. If this rate of increase continues, it implies a doubling of electricity consumption every five years. Given that this rate is highly likely to increase, the doubling of electricity consumption is likely to be much faster than anticipated.
But in the United States, these figures take on a whole new dimension. Considering only the electricity consumption of data centers , this reached 32 gigawatts in 2024, and the anticipated increase in 2030 could reach over 124 gigawatts. For reference, 1 gigawatt represents 1 million kilowatts, or 1 billion watt-hours. 1CSIS, “The Electricity Supply Bottleneck On U.S. AI Dominance,” March 3, 2025
Electricity generation is therefore a major physical limitation to the development of AI, while generation capacity can quickly become a crucial competitive advantage in the AI race between the US and China.
The Data Center Season
Furthermore, the electricity consumption of the AI sector directly competes with the needs of both individuals and all other sectors of activity, including all types of industry and the global rise of electric vehicles, which are themselves inextricably linked to the spread of new AI applications. In January 2026, President Trump even issued an executive order decoupling residential electricity bills from data center consumption.
In the United States, the combination of massive investments and substantial support from the federal government has led to a proliferation of data center construction by the various GAFAM companies. For example, on February 4, 2026, Google announced that it had achieved annual revenue of $402.8 billion in 2025.
This result represents a 15% increase compared to 2024, while also including a net profit of $132 billion. According to Sundar Pinchas, CEO of Alphabet, Google’s parent company, these staggering figures are attributable to the penetration of enterprise AI solutions by Gemini, Google’s AI platform. Between November 2025 and January 2026, the number of monthly Gemini users jumped from 650 million to 750 million.
Furthermore, Google’s other activities, including its search function, YouTube, and advertising, generated hundreds of billions of dollars. The Google Cloud division alone generated $70 billion.
But the “cloud” is by no means a metaphysical concept, as the metadata “clouds” are stored in Google’s network of data centers. Maintaining Google’s expansion therefore requires the expansion of the company’s physical infrastructure, particularly in the form of data centers.
Faced with this need, Alphabet announced it would invest $175 billion in building new data centers by 2026, almost double the amount allocated to data centers in 2025. It’s worth noting that Microsoft, Meta, OpenAI, and Amazon are announcing similar investments to build new data centers by 2026.
Electricity in America
2026 will therefore be dominated by the need to generate the electricity required for these new data centers, while continuing to power the existing infrastructure.
This pressure from the AI sector will largely contribute to the more than 16% increase in electricity demand by 2030. Gas-fired power plants already provide 43% of electricity production in the United States, while coal accounts for 16% and nuclear power 18.6%.
The remainder coming from renewable energy sources. U.S. Energy Information Administration
Moreover, it is a mature, well-mastered technology. Faced with the explosion in the number of data centers, US electricity companies have also embarked on a race to build gas-fired power plants.
Hélène Lavoix: US Nuclear Renaissance, 1 and 2
This has resulted in a profound reorganization of the American landscape, with the emergence of what we call “The Great Disruptor,” namely the coupling of the “California-Arizona” GAFAM-TSMC technological duo with the oil&gas “Texas-Louisiana” energy duo.
Arizona, the new AI Frontier
Indeed, the establishment in Arizona of the first factory of TSMC, the Taiwanese semiconductor giant, and the construction of three more factories by 2030, representing an investment of $200 billion, is transforming this semi-desert state into a new AI hypercenter.
This is attracting all the GAFAM companies and investors. But at precisely the same time, Texas and Louisiana are once again becoming major oil and gas powers. This phenomenon is due to the ongoing combination of the oil and shale gas revolution, of which Texas, with the exploitation of the gigantic Permian Basin, has become the center.
In other words, Texas and Louisiana produce the gas necessary to meet the energy needs of the data centers and supercomputers of the GAFAM companies located in Arizona and California.
Thus, shale gas production is literally the energy base of the American AI power race. Furthermore, Venezuela’s colossal natural gas reserves are already attracting the attention of American companies.
II. AI for Energy
With the global rollout of the artificial intelligence (AI) revolution, all major oil companies—whether American, French, Saudi, Russian, Chinese, British, etc.—are integrating these technologies into all their operations, from exploration and production to downstream.
AI is becoming a tool for optimizing efficiency, from geological exploration to equipment management, without neglecting the financial aspect.
AI for Oil
At the same time, the oil and gas industry is generating colossal revenues, reaching $5.95 trillion in 2024. These revenues are also funding the development of the artificial intelligence sector. Indeed, the oil industry is heavily dependent on data.
Oil companies need geological data to prospect for new deposits and manage currently operating fields. Oil refining requires the production of a highly accurate and constant data stream.
The Everything AI Optimization
The same applies to the storage of oil and petroleum products, as well as to reserve management. This is also true for transportation, by pipeline, tanker, carrier, and truck fleets. Furthermore, the sale and exchange of oil and petroleum products at the global, regional, and local levels generate a considerable amount of financial, legal, and regulatory data.
In other words, on a global scale, the oil industry constitutes a massive engine of extremely complex data. This is why the emerging family of artificial intelligence (AI) technologies is intersecting with the oil and gas industry: technologies such as deep learning rely on the classification and interpretation of big data.
Jean-Michel Valantin, Petrodollars for AI – Is AI the Future of Oil and Gas Industry ?
Thus, the central role of data production, management, and application makes the oil industry an ideal catalyst and integrator for AI.
Hélène Lavoix, “Sensors and Actuators for AI(1): Integrating AI into Reality,” The Red Team Analysis Society, January 14, 2019
The same applies to the need to anticipate demand and new reserves.
Oil as Data
For example, Exxon Mobil, the American giant, uses AI to optimize its energy consumption, particularly at refineries. British Petroleum integrates AI algorithms into its exploration services. Shell uses predictive AI to improve the utilization and efficiency of its equipment while reducing downtime.
Through this predictive maintenance approach, AI algorithms analyze sensor data from equipment and machinery at refineries and drilling sites to anticipate breakdowns, thus enabling timely maintenance and avoiding costly unplanned shutdowns.
It is worth noting that, since the 2000s, the United States has been the country where the two major revolutions in the oil and shale gas industry and artificial intelligence have taken place. Moreover, these two revolutions followed parallel timelines.2Loretta Napoleoni, Technocapitalism: The Rise of the New Robber Barons and the Fight for the Common Good, Seven Stories Press, 2024
Furthermore, the shale oil technological revolution, with horizontal drilling and hydraulic fracturing, has made the United States a net producer and exporter of oil and gas.3Daniel Yergin, The New Map: Energy, Climate, and the Clash of Nations, 2020.
AI for Hydraulic Fracking
However, this new approach relies on heavy equipment that requires significant and ongoing investment. Shale oil companies face highly complex challenges, such as optimizing extraction, modeling reservoirs, forecasting production, and adapting to evolving demand. Integrating AI allows these companies to optimize various aspects of their operations, thereby securing revenue while mitigating risks.
The US shale oil and gas industry, heavily reliant on technology, is integrating machine learning, particularly to accelerate drilling operations.
This alliance between the American AI and shale oil sectors also guarantees oil and gas production for the US domestic market. This development is of strategic importance for the growth of the AI sector.
Indeed, the proliferation of data centers, the training of AI models, and the widespread adoption of AI are generating increasing demand for electricity.
Hélène Lavoix, “Uranium for the American Nuclear Renaissance (2): Towards a Global Geopolitical Race”; Goldman Sachs, “GS Sustain: Generational Growth: AI, Data Centers, and the Next Explosion in US Electricity Demand,” April 29, 2024.
III. The Space Race for energy
However, the dual race for electricity production and AI optimization cannot fully meet the anticipated increase in electricity consumption by the US AI sector.
This results in a new race to orbital space to install new generations of data centers. The installation of data centers in space aims to address the physical constraints that the proliferation of data centers is encountering on Earth, namely the limited availability of electricity and water.
Space is energy
That’s why SpaceX is considering developing specialized versions of its Starlink satellite constellations, which, equipped with interconnected AI, could achieve the computing power of terrestrial data centers. These satellite-data center constellations will be powered by solar panels, and the satellites will be cooled by sub-zero space temperatures. Cesareo Contreras, The next Great Space Race : Building data centers in Orbit, Phys.org, 7 January, 2026.
Furthermore, as the only company in the world to have made spaceflight profitable through the number of flights and its mastery of reusable rocket landings, SpaceX is uniquely positioned to undertake the effort of launching data centers into orbit.
These satellites will be an improved version of its Starlink constellations, equipped with Nvidia’s H100 super semiconductors, powered by solar panels and passively cooled by space temperatures.
Finally, since November 2025, an American satellite has been equipped with Nvidia H100 chips and runs “Gemma,” Google’s AI model. Furthermore, SpaceX is considering developing specialized versions of its Starlink satellite constellations, which, equipped with interconnected AIs, could achieve the computing power of terrestrial data centers.
Titan
To acquire the technological and capital resources necessary for this strategy, Elon Musk has just merged SpaceX and xAI (the parent company of its AI, Grok). Meanwhile, the June 2026 IPO of this entity, which effectively merges the launch vehicle, satellite (Starlink, SpaceX’s satellite division), and AI sectors with xAI-Grok, could reach $1.5 trillion.
In other words, the race for AI power is transforming American and global energy geopolitics, turning it into an AI-energy “space politics.” This transformation is beginning to materialize as early as 2026. It will trigger a triple growth in the AI, space, and energy sectors, and thus in the associated infrastructure and mineral resources, while also transforming the very nature of American power.
Why is this important for French and European companies?
- Many French companies possess cutting-edge expertise in aerospace, energy, and electronics. These sectors are experiencing rapid growth in the US due to the convergence of AI, energy, and space. They therefore need to establish a presence in the market, and to do so, they need to learn about these new, high-potential markets.
- Furthermore, the same logic applies to the financial sector, particularly in the world of investment funds. North American companies, facing difficulties, are engaging in a “rush for critical minerals” central to the manufacture of AI hardware.
- Finally, entering the US market will also allow them to discover the opportunities presented by the military dimension of the AI/space/AI vortex.
