Personally I feel like while the generative AI bit makes for a good headline, the better summary is buried later in the article:
Nvidia also announced its commitment to offer accelerated computing and a comprehensive suite of AI software as part of the collaboration. The company stated that substantial co-engineering efforts are underway, intending to integrate the Nvidia AI engine into Snowflake’s Data Cloud.
So this isn't just NeMo for LLMs, it's also likely to be a tight integration with a bunch of NVidia's AI/ML stack.
Disclosure: I work for Snowflake, but I didn't work on this and I'm not speaking for them.
Comments
Personally I feel like while the generative AI bit makes for a good headline, the better summary is buried later in the article:
So this isn't just NeMo for LLMs, it's also likely to be a tight integration with a bunch of NVidia's AI/ML stack.
Disclosure: I work for Snowflake, but I didn't work on this and I'm not speaking for them.
Compute is definitely a bottleneck right now. Did they quantify this commitment in any way?