Speaking at the conference, Russo also disclosed that OpsRamp, the SaaS-based IT operations platform it bought earlier this year, is now available via Greenlake.
According to Russo, it can take 1,200 MWh to train a large language model (LLM), and 250 MWh per day to operate one for inferencing purposes, contributing significantly to the energy bills an organization's IT systems generate.
[
add
]
[
|
|
...
]