Meta rolls out its next generation infrastructure for AI

The parent company of Facebook, Instagram and WhatsApp, Meta has come up with the next generation of Meta Training and Inference Accelerator (MTIA). This step of Meta indicates that the company is committed to enhance AI-driven experiences across its products and services.

The next generation of MTIA is part of a broader full-stack development program for custom, domain-specific silicon that addresses Meta’s unique workloads and systems. This new version of MTIA more than doubles the compute and memory bandwidth of its previous solution while maintaining close tie-in to workloads. It is designed to efficiently serve the ranking and recommendation models that provide high-quality recommendations to users.

year, Meta rolled out the first-generation AI inference accelerator, Meta Training and Inference Accelerator (MTIA) v1 which was designed in-house with Meta’s AI workloads in mind.

According to Meta, “MTIA is a long-term bet to provide the most efficient architecture for Meta’s unique workloads. As AI workloads become increasingly important to our products and services, this efficiency will be central to our ability to provide the best experiences for our users around the world. MTIA v1 was an important step in improving the compute efficiency of our infrastructure and better supporting our software developers as they build AI models that will facilitate new and better user experiences.”

The new chip’s architecture is fundamentally focused on providing the right balance of compute, memory bandwidth and memory capacity for serving ranking and recommendation models.

MTIA has been deployed in Meta’s data centers and is now serving models in production. The company is already seeing the positive results of this program as it is allowing to dedicate and invest in more compute power for more intensive AI workloads.

The results so far show that this MTIA chip can handle both low complexity and high complexity ranking and recommendation models which are key components of Meta’s products.