Microsoft does not have enough energy to power NVIDIA chips

Recently, Microsoft unexpectedly announced that it did not have enough energy capacity to connect all existing processors for AI needs.

Reasons for energy problems

The situation in the field of AI is mainly related to the shortage of energy resources to support the infrastructure, rather than the amount of computing power.

The problem with data center energy consumption for AI is not new; it has been discussed by experts for a long time. Some technology companies have previously announced their investments in small nuclear reactor research in order to easily scale data centers around the world.

In the United States, for example, ordinary consumers are already seeing significant increases in their electricity bills. OpenAI is also calling on the US government to invest in the construction of new power plants to gain significant advantages in the battle for the AI market.

However, there are also certain risks associated with such trends. For example, the infrastructure for training new models will certainly be in demand, but the demand for data centers may not be as significant as some experts expect. Demand may not reach colossal levels if semiconductor technology allows new AI models to be launched locally. This could accelerate the collapse of AI and affect the market as a whole, and even non-technology companies in particular.