Science and Tech

Meta prepares for the next generation of AI: its own chip, data center and a supercomputer

Meta would be deliberately draining the battery of mobile phones to test its apps according to a former employee

Meta wants to design the next generation of artificial intelligence. It is developing its own chip, a next-generation data center, and a training supercomputer.

He was late to artificial intelligence, and he doesn’t want it to happen again. Meta has revealed its ambitious plan to develop its AI infrastructure, which includes the custom chip creation and an optimized data center. These efforts are intended to support the development of new applications based on AI and the metaverse.

With projected growth in demand for AI computing over the next decade, Meta is focusing its resources on building the next generation of AI infrastructure. One of the main innovations is the development of its own custom silicon chip to run artificial intelligence models.

This chip, called MTIA provides higher computing power and efficiency than conventional CPUs, and is designed specifically for Meta’s internal workloads. With it, Meta hopes to deliver improved performance, lower latency, and greater efficiency in every AI task.

In addition to the custom chip, Meta is building a next-generation data center designed for AI training. This data center will be optimized for artificial intelligence and will include liquid-cooled AI hardware and a high-performance network to connect thousands of AI chips in data center-scale training clusters.

This network will allow faster and more cost-effective developmentcomplementing other hardware solutions such as its first ASIC (Application Specific Integrated Circuit) called MSVP, which is intended to meet Meta’s growing video workload demands.

One of the world’s fastest AI training supercomputers

Another notable part of Meta’s plan is the supercomputer from IA Research SuperCluster (RSC), one of the fastest AI supercomputers in the world.

The RSC has 16,000 GPUs and was built to train artificial intelligence models that drive new augmented reality tools, content understanding systems, or real-time translation technology.

With this investment in your AI infrastructure, Meta seeks to lead in the field of artificial intelligence for the next decade.

Its focus on developing custom chips, an optimized data center, and a cutting-edge supercomputer demonstrate its commitment to advancing technology and creating new applications that benefit users around the world.

Source link