Nvidia is moving faster in an effort not only to help drive the spread of open-weight models, but also to lead technological advances.

Nvidia recently signed a $6 billion licensing deal with coding AI startup Poolside. The deal also includes hiring most of Poolside's engineers. Nvidia will invest $1 billion, valuing Poolside at $12 billion pre-money.

In connection with the deal, The Wall Street Journal reported Nvidia will use Poolside to speed development of a world-class open-weight model that can compete with Chinese AI companies such as DeepSeek and Kimi K3.

Open-weight models have lower operating costs and are easier to customize, which could allow them to directly challenge closed-model players such as OpenAI and Anthropic, the WSJ reported, citing sources familiar with internal discussions.

Poolside is a startup founded in 2023 by software developer Eiso Kant (에이소 칸트) and former GitHub chief technology officer Jason Warner (제이슨 워너). In a letter to shareholders, Poolside said the deal with Nvidia was a choice aimed at making artificial general intelligence, or AGI, an open technology built by many rather than monopolised by a few.

Poolside's management will not join Nvidia and will continue separate research. About 100 Poolside employees, including engineers, will move to Nvidia and be assigned to develop Nemotron, an open-weight project Nvidia unveiled in 2023.

Nvidia in March launched the Nemotron Alliance, an open-model coalition that includes Mistral, Thinking Machines Lab and Perplexity. Last month, CEO Jensen Huang posted on social media to stress the importance of open-weight models. Since DeepSeek unveiled a low-cost open-weight model in January 2025 and sent shock waves through U.S. stocks, U.S.-China competition in AI has continued, the WSJ reported.

Even if it does not plan to compete directly with AI companies, Nvidia's push toward its own open-source AI models is increasingly shifting into an aggressive mode. That could complicate its relationships with AI companies developing models on Nvidia GPUs, whether open-source or closed, a growing number of observers say.

ㆍ[Tech Inside] Why is Nvidia trying to develop large AI models itself?

More companies are adopting Nvidia AI models. Palantir, for example, announced in June that it would cooperate with Nvidia to use Nemotron models for U.S. government customers.