One of the sources said Meta’s new training chip is a dedicated accelerator, meaning it is designed to handle only AI-specific tasks. This can make it more power-efficient than the integrated graphics processing units (GPUs) generally used for AI workloads.

The Facebook owner is currently one of Nvidia’s biggest customers, amassing an arsenal of GPUs to train AI models.

The push to develop in-house chips is part of a long-term plan at Meta to bring down its mammoth infrastructure costs as the company places expensive bets on AI tools to drive growth.

Any setbacks in new chips meant that companies were forced to use older chips, which might not sit well with customers. Perhaps the AI industry is learning from this, because, according to a report from Reuters,