When a company stops building servers and starts building territory, the balance of power shifts.
Menlo Park, November 2025.
Inside meeting rooms where decisions sound like mergers of physics and finance, Meta approved an investment that rewrites the scale of private technology. The company plans to deploy approximately six hundred billion dollars to expand a global network of data centers designed to feed artificial intelligence systems that will demand energy, cooling and processing power at a level previously linked only to nation states. Meta does not frame the move as expansion. It calls it preparation for the next intelligence era.
The strategy is based on a simple but brutal idea. Artificial intelligence is no longer about training models. It is about who controls the computational infrastructure that allows those models to exist. The company has already begun acquiring land for facilities capable of managing gigawatt level energy consumption and millions of advanced chips working simultaneously. These centers are not designed for social media. They are designed for models that learn from billions of parameters and improve without human prompts.
In Europe, analysts note that the proposed scale rivals the infrastructural footprint of national telecommunications projects. Governments once led this frontier. Meta is acting as if the frontier is already private. In Asia, experts warn that computing power is becoming a new form of geopolitical influence, comparable to the control of minerals or shipping routes. In Latin America, observers worry that the gap in access to artificial intelligence will deepen not because of lack of talent, but because of lack of electricity and data pipelines to support this kind of infrastructure.
The technical challenge is staggering. A facility of this scale devours power. It demands industrial cooling systems that behave like artificial climates and fiber networks able to transmit enormous traffic without latency collapse. The company has negotiated long term supply contracts for renewable generation, signaling that the future of AI will not be limited by algorithms but by megawatts. At a strategic level, Meta is positioning itself not as a software company, but as an energy and infrastructure operator disguised as a software company.
The decision also reveals a psychological shift. Instead of waiting for artificial intelligence demand to explode, Meta is building as if that future has already arrived. It is not reacting to competitors. It is attempting to set the conditions of the market. The message is clear. Whoever owns the processing layer will own the intelligence layer. Whoever owns the intelligence layer will own the next economy.
The scale forces uncomfortable questions. What happens when private corporations build infrastructures larger than those of many countries? Where does oversight exist when decision making is not public but corporate? The answers are not ready. The infrastructure is arriving faster than the rules.
Meta calls this an investment. In reality, it is a bet on a new world order built from electricity, cooled by water and governed by algorithms. The company believes that the future will not be defined by the artificial intelligence we use, but by the artificial intelligence we can afford to run.
The world is about to learn that intelligence has a physical cost.
Global narrative resilience.
Resistencia narrativa global.