FrontierThe story, in brief
Salesforce and Nvidia’s new reasoning model is everything the AI labs should fear
Salesforce + Nvidia just shipped a reasoning model that does sales work out of the box. The frontier labs should be nervous.
The KeyNews take
Why it matters
A domain-specific reasoning model built on open weights and trained for real enterprise workflows signals that frontier capability is flowing downstream faster than expected—and that enterprise vendors can now compete on model quality, not just integration.
The key facts
5 to knowSalesforce Koa model built on Nvidia Nemotron open-weight base
Trained specifically for sales, marketing, customer-support tasks
Domain-specific reasoning model (not a generic LLM)
Open-weight foundation enables rapid customization
Enterprise vendor shipping proprietary reasoning capability (historical frontier lab domain)
Go to the source
TechCrunch AItechcrunch.com
Publisher excerpt: Salesforce Koa is built on Nvidia's open-weight Nemotron model and is trained to do sales, marketing and customer-support tasks.