The AI industry is seeing a wave of consolidation as some of the largest technology companies move to acquire open-weight model platforms and builders. According to a TechCrunch report, Nvidia is nearing a $13 billion acquisition of Hugging Face, the platform at the center of the open-weight AI developer ecosystem. The deal would follow Nvidia's $6 billion agreement with Poolside, an open-weight model builder whose employees will largely move to the chip maker, and Stripe's acquisition of OpenRouter, a provider of open-weight models to businesses, for more than $7 billion.
For Nvidia, the moves reflect a strategy to reduce dependence on major hyperscalers and frontier AI labs — especially as companies like OpenAI and Google develop their own inference chips. OpenAI recently unveiled its Jalapeño chip, whose capabilities were announced this week. "If model builders are making chips, Nvidia wants a chunk of the model-making business," the report notes. Nvidia already offers its Nemotron family of open-weight models, but uptake has been limited. Acquiring Hugging Face would give it access to a large user base it can steer toward its chips and standards.
Adoption of open-weight models remains relatively small but is growing. According to a survey of spending data by Ramp, just 6% of companies use open-weight models, while only 2% of software engineers surveyed by Jellyfish report using them. Nik Albarran, AI product lead at Jellyfish, told TechCrunch that open-weight models are primarily used by companies with high-volume, repetitive inference workloads — such as customer service chatbots — where they can be tuned to answer questions cheaply.
Stripe's rationale for the OpenRouter deal was framed similarly. "Tokens are the central currency for companies building with AI, and it's clear that the real-world economic potential will depend on making good use of scarce compute resources," Patrick Collison, Stripe's co-founder and CEO, said in a statement.
For more complex tasks like coding and agentic workflows, Albarran noted that frontier models often still win out due to easier access and token subsidies. However, he suggested that as companies refine their AI workflows, open models will become more viable. Control and configurability remain the main drivers for companies adopting open-weight models today, rather than cost concerns. "There are not many companies where that is the case yet … [but] if the prices continue to go up from the frontier labs," Albarran said before the quote was cut off in the source material.