AI Pulse · Business Models & Winners

Nvidia has reportedly taken an unusual route with AI startup Poolside. Rather than buying the company outright, it paid $6 billion for a non-exclusive licence to Poolside’s internal “Model Factory” software, then invested a further $1 billion at a $12 billion pre-money valuation. More than 100 Poolside engineers are moving to Nvidia’s Nemotron team, while Poolside’s three founders remain independent.

The non-exclusive part matters. Poolside can still sell the same software to other customers. Nvidia gets access to a training pipeline and a team that has built it, without taking on the full cost and complexity of an acquisition. Poolside gets badly needed capital after a failed $2 billion fundraising effort left it unable to fund its planned GPU cluster commitment.

For business owners, this is a useful signal about where value may sit in AI. It is not only in the final model or chatbot. The systems used to train, test, manage and improve models are becoming valuable assets in their own right. Large firms may increasingly buy access to those capabilities, hire the teams behind them, and leave the original business operating separately.

That creates an opening for specialist AI companies, but also a warning. If your company depends on expensive infrastructure commitments before revenue arrives, funding risk can quickly shape your strategic options. This is the kind of thing we build into an AI command centre, tracking where the real commercial value is moving, not just which model made the latest headline.