Jorge F. Negrete P.
To understand Artificial Intelligence (AI), it is essential to grasp a syllogism: there is no Artificial Intelligence without infrastructure, and that infrastructure is Digital Infrastructure. Period.
AI is not a technology isolated from the digital world; it is enabled by digital infrastructure and by the evolution of connectivity networks that make it viable and vital for this historic moment in our global society. “The coming wave” (Mustafa Suleyman) is the child of connectivity—of the old telecommunications networks that have evolved into a complex and robust concept: digital infrastructure. The stronger and broader the infrastructure, the greater the release of AI. Conversely, mediocre digital infrastructure with limited connectivity for citizens guarantees little or no access to AI.
Governments that fail to be competitive in adopting AI will be left behind in digital economy, competitiveness, innovation, and the opportunity to provide economic and social development to their citizens. Chiapas and Oaxaca, with 20 Mbps bandwidth and just over 50% internet coverage, stand in sharp contrast to Chile’s 400 Mbps and 97% coverage. Some remain trapped in digital marginalization, while others ride the superhighway of global competitiveness.
The relationship between digital infrastructure and AI. There is no public policy, digital policy, science and technology strategy, or economic and investment policy for the digital ecosystem that does not begin with the promotion and deployment of digital infrastructure. Talking about AI in healthcare, education, finance, public security, or computing without digital infrastructure is a fallacy—if not mere propaganda.
AI depends on a solid, high-performance digital infrastructure to function efficiently. The relationship between AI and 5G networks, Data Centers, and processors is symbiotic. AI needs this infrastructure, and in turn, helps manage and optimize it.
AI models, especially Deep Learning, require astronomical processing capacity. Traditional processors are not efficient for AI. This is why Graphics Processing Units (GPUs) and, more recently, Tensor Processing Units (TPUs) are used, as they are designed to accelerate the training of large-scale AI models.
Data Centers are the factory and training ground of AI. They are the brain of its infrastructure. Data Centers host thousands of processors and high-performance storage systems. They are vital for training AI models with massive data volumes: they provide computing power, storage, and real-time predictions.
5G networks. Connectivity “is the nervous system” (Carlos Slim) that get together all components of AI. 5G networks, with their high speed, bandwidth, and low latency, are vital for AI. They make AI stable, efficient, and integrated into the digital ecosystem. This allows devices and sensors in factories, smart cities, ports, or autonomous equipment to send and receive vast amounts of data. 5G ensures that AI makes decisions in real time.
Digital infrastructure goes beyond Data Centers; it extends across all telecommunications networks, energy grids, and supply chains.
As AI models grow larger and more complex, AI’s efficiency will depend on the hugeness and capacity of digital infrastructure: more fiber optics, more radio base stations, more spectrum, more powerful processors, larger Data Centers, and faster 5G networks.
The challenge is enormous. Where does that leave Chiapas, Oaxaca, Guerrero, and Michoacán? With the worst connectivity in Mexico? And what about Nuevo León or the Bajío, which need bandwidths like those of Chile or Brazil to remain competitive for foreign investment?
An AI society requires the best digital infrastructure.
President of Digital Policy & Law
X / @fernegretep