Meta is taking a significant step to optimize its AI infrastructure by extending the life of its AI servers and reducing costs by $2.9 billion in 2025. This move is part of the company's broader effort to improve efficiency and reduce expenses.
To achieve this, Meta is redesigning its data centers to be cheaper and faster to build, with a focus on supporting both AI and non-AI workloads. The company has also revised its capital expenditure forecast, expecting to spend between $30-33 billion, down from the initial estimate of $34-37 billion.
This cost-cutting measure is notable, especially considering Meta's ambitious plans to invest between $60-65 billion in AI infrastructure by 2025. The company aims to build a state-of-the-art data center with over 2 gigawatt capacity, acquire over 1.3 million advanced graphics processing units (GPUs), and refine generative AI technologies.
