A TechCrunch article reports that open-weight AI companies are becoming prime acquisition targets for tech giants, citing three large recent deals: Nvidia‘s rumored $13 billion acquisition of Hugging Face, a platform for sharing open-weight models and benchmarks; Nvidia‘s $6 billion deal with Poolside, an open-weight model builder; and Stripe‘s $7 billion acquisition of OpenRouter, a provider of open-weight models to businesses. The article frames this trend as a strategic move by Nvidia to reduce dependence on hyperscalers and frontier labs, especially as model builders like OpenAI and Google develop their own inference chips. By acquiring open-weight platforms, Nvidia gains access to a large developer ecosystem that can be driven to its chips and standards.

Adoption of open-weight models remains modest—just 6% of companies use them, according to a Ramp survey, and 2% of software engineers surveyed by Jellyfish. Nik Albarran, AI product lead at Jellyfish, notes that open-weight models are primarily used for high-volume, repetitive inference workloads like customer service chatbots, where they can be tuned cheaply. For coding and agentic tasks, frontier models often win due to easier access and token subsidies. However, as AI workflows mature and if proprietary model prices rise, more companies may consider open-weight alternatives.

Lin Qiao, CEO of Fireworks (a leading open-weight model router and potential acquisition target), says her company processes 40 trillion tokens daily, more than Gemini or OpenAI’s APIs. She argues for model diversity, predicting that every company will eventually train specialized models per use case. The article concludes that the dominance of OpenAI and Anthropic is not inevitable, and that the allure of open technology is proving tough for tech giants to resist.