Open-weight AI companies are emerging as significant acquisition targets for major tech firms, with several high-value deals reported recently. Nvidia is reportedly pursuing a $13 billion acquisition of Hugging Face, a platform for sharing open-weight AI models and benchmarks.
This follows Nvidia's $6 billion agreement with Poolside, an open-weight model builder, which will see most of Poolside's employees move to the chip-making giant. Two weeks prior, Stripe acquired OpenRouter, a leading provider of open-weight models to businesses, for more than $7 billion.
This trend reflects a shift in the AI sector, as companies like Nvidia seek to reduce dependence on major hyperscalers and frontier labs, especially as model builders like OpenAI and Google develop their own inference chips. Nvidia aims to gain access to a large user base for its chips and standards by controlling a significant developer space for open models.
The increasing cost of AI inference is also prompting companies to explore cheaper open-weight models. While adoption is currently small, with 6% of companies using open-weight models according to Ramp data, or 2% of software engineers surveyed by Jellyfish, it is growing. These models are primarily used by companies with high-volume, repetitive inference workloads, such as customer service chats, where they can be tuned for cost-effectiveness.
Patrick Collison, Stripe's cofounder and CEO, stated that the economic potential of AI depends on efficient use of compute resources. Nik Albarran, AI product lead at Jellyfish, noted that while frontier models often prevail for complex tasks, open models offer control and configurability. Albarran suggested that if prices from frontier labs continue to rise, more companies may consider open models, particularly as their AI workflows mature.