Big tech are playing everything at AI

Goal is negotiating the construction of a campus for a gigantic data center. It would be intended for its artificial intelligence projects, and the potential cost could exceed 200 billion dollars, sources close to those plans indicate In The Information.

Potential locations. The project seems to have the construction of that campus in states such as Louisiana, Wyoming or Texas. Those responsible for the goal have already visited possible locations, they indicate these sources. These three states have land at affordable prices, Abundant and cheap energy and even proximity to large Internet communications infrastructure.

Meta denies it. A Meta spokesman has denied these rumors, and has indicated that the company already communicated its forecast of capital expenses (CAPEX). In fact, it describes these rumors of “pure speculation.”

Zuckerberg already gave data about CAPEX. The truth is that Mark Zuckerberg effectively gave precisely information about his Capex forecast by 2025. The figure is very high, 65,000 million dollars, but it is approximately one third of what these new rumors would point out.

Data centers everywhere. The fever for investment in data centers seems to have been infected among large technology companies, and all of them have recently announced extraordinary projects to build data centers dedicated to AI. Thus, we have:

• Amazon: 100,000 Millions of dollars
• Microsoft: 80,000 Millions of dollars
• Google: 75,000 Millions of dollars
• Goal: 65,000 Millions of dollars
• Apple: 12,000 Millions of dollars

And Stargate what? Meanwhile, the other great project in this area It is undoubtedly Stargateled by OpenAI and Softbank as main investors and technological partners, and that theoretically raises An investment of 500,000 million dollars in four years.

But AI models want to be more efficient. That tendency to create gigantic data centers and that will consume vast amounts of energy contrasts with the fact that AI models try to be increasingly efficient and capable, as we have seen with Deepseek. So, what happens to all that investment, where is it addressed?

Demand (theoretically) will grow big. What Big Tech are doing is preparing for a massive AI adoption by both private clients and companies. Although today use is relatively modest, these investments raise a clear future. One in which we will use the AI ​​constantly, as we now use the Internet or our mobiles. And 8,000 million people using AI at all hours will need a lot of computing capacity.

And the AI ​​in local, what? It is also expected that part of that load will not be executed in these large data centers, but on our devices. The small models – as Gemini Nano in Android or Apple Intelligence on the iPhone – will be able to replace part of the functions we will need, which will “download” partly to the data centers.

But. Of course, industry and Big Tech forecasts could fail. These immense investments are a clear commitment to the revolution that the AI ​​poses, but it remains to be seen if this technology will be infiltrated in our lives as mobiles or internet have done. The latter, for example, caused the bubble of the Puntocom in 2000, and that has made the debate about a potential bubble of AI There is also.

Image | Goal | Wikimedia

In Xataka | We already know how much spending on AI by companies in 2024 has grown. An absolute barbarity