Meta Platforms is finalizing a supercomputing cluster of over 100,000 H100 server chips, costing more than $2 billion, to train the next version of its Llama AI model.
Meta invests $2B in 100K H100 chips for Llama AI
By
–
By
–
Meta Platforms is finalizing a supercomputing cluster of over 100,000 H100 server chips, costing more than $2 billion, to train the next version of its Llama AI model.