Save StorySave this storySave StorySave this story<\/p>\n
Nvidia will spend $26 billion over the next five years to build open source artificial intelligence models, according to a 2025 financial filing. Executives confirmed the news, which has not been previously reported, in interviews with WIRED.<\/p>\n
The sizable investment could see Nvidia evolve from a chipmaker with an impressive software stack into a bona fide frontier lab capable of competing with OpenAI and DeepSeek. It\u2019s a strategic move that could further entrench Nvidia\u2019s place as the AI world\u2019s leading chip manufacturer, since the models are tuned to the company\u2019s hardware.<\/p>\n
Open source models are ones where the weights or the parameters that determine a model\u2019s behavior are released publicly\u2014sometimes with the details of its architecture and training. This allows anyone to download and run it on their own machine or the cloud. In Nvidia\u2019s case, the company also reveals the technical innovations involved in building and training its models, making it easier for startups and researchers to modify and build upon the company\u2019s innovations.<\/p>\n
On Wednesday, Nvidia also released Nemotron 3 Super, its most capable open-weight AI model to date. The new model has 128 billion parameters (a measure of the model\u2019s size and complexity), making it roughly equivalent to the largest version of OpenAI\u2019s GPT-OSS, though the company claims it outperforms GPT-OSS and other models across several benchmarks.<\/p>\n
Specifically, Nvidia claims Nemotron 3 Super received a score of 37 on the Artificial Intelligence Index, which scores models across 10 different benchmarks. GPT-OSS scored 33\u2014but several Chinese models scored higher. Nvidia says Nemotron 3 Super was secretly tested on PinchBench, a new benchmark that assesses a model\u2019s ability to control OpenClaw, and ranks number one on that test.<\/p>\n
Nvidia also introduced a number of technical tricks that it used to train Nemotron 3. These include architectural and training techniques that improve the model\u2019s reasoning abilities, long-context handling, and responsiveness to reinforcement learning.<\/p>\n
\u201cNvidia is taking open model development much more seriously,\u201d says Bryan Catanzaro, VP of applied deep learning research at Nvidia. \u201cAnd we are making a lot of progress.\u201d<\/p>\n
Meta was the first big AI company to release an open model, Llama, in 2023. CEO Mark Zuckerberg recently rebooted the company\u2019s AI efforts, however, and signaled that it might not make future models fully open. OpenAI offers an open-weight model, called GPT-oss, but it is inferior to the company\u2019s best proprietary offerings, not well-suited to modification.<\/p>\n
The best US models, from OpenAI, Anthropic, and Google, can be accessed only through the cloud or via a chat interface. By contrast, the weights for many top Chinese models, from DeepSeek, Alibaba, Moonshot AI, Z.ai and MiniMax are released openly and for free. As a result, many startups and researchers around the world are currently building on top of Chinese models.<\/p>\n
\u201cIt's in our interest to help the ecosystem develop,\u201d says Catanzaro, who joined Nvidia in 2011 and helped spearhead the company\u2019s shift from making graphics cards for gaming to making silicon for AI. Nvidia released the first Nemotron model in November 2023. He adds that Nvidia recently finished pretraining a 550-billion-parameter model. (Pretraining involves feeding huge quantities of data into a model spread across vast numbers of specialized chips running in parallel.) Nvidia has since released a range of models specialized for use in areas like robotics, climate modelling, and protein folding.<\/p>\n
Kari Briski, VP of generative AI software for enterprise, says Nvidia\u2019s future AI models will help the company improve not just its chips but also the super-computer-scale datacenters it builds. \u201cWe build it to stretch our systems and test not just the compute but also the storage and networking, and to kind of build out our hardware architecture roadmap,\u201d she says.<\/p>\n
Releasing models openly may have long-term strategic benefits for Nvidia, too. The company\u2019s chips remain the gold standard for training large AI models, with customers spending billions to acquire the company\u2019s hardware for their datacenters. But the rise of Chinese open models might at some point erode Nvidia\u2019s position if those models were to demonstrate dramatic improvements on rival hardware.<\/p>\n
In January 2025, DeepSeek released a cutting-edge open model using a more efficient approach that made its training far cheaper. But a variety of other Chinese models from big companies like Alibaba, as well as startups like Moonshot AI, Z.ai, and MiniMax, have also become popular in the West. Alibaba\u2019s model Qwen, which is easy to use and modify and is well maintained, is widely used by researchers and startups.<\/p>\n
A new DeepSeek model, expected to be released soon, is widely rumored to have been trained exclusively on chips made by the Chinese company Huawei, which is subject to US government sanctions. If true, the release could prompt more startups and researchers to try Huawei\u2019s hardware, particularly in China.<\/p>\n
In this respect, Nvidia may help shape AI competition between the US and China by providing a US-made alternative to open-weight Chinese models.<\/p>\n
\u201cWe're an American company, but we work with companies across the world,\u201d Catanzaro says. \u201cIt's in our interest to make the ecosystem diverse and strong everywhere.\u201d<\/p>\n
Some industry experts have warned that seeing open innovation shift to the other side of the world could be bad for the US in the long run.<\/p>\n
\u201cI'm a huge Nemotron fan,\u201d says Nathan Lambert, an AI researcher at the Allen Institute for AI (Ai2) who leads the ATOM (American Truly Open Models) Project. Lambert adds that the US government should also fund open models.<\/p>\n
Andy Konwinski, a computer scientist and entrepreneur who leads the Laude Institute, a nonprofit focused on promoting openness in AI, says Nvidia\u2019s investment is highly significant because of its position at the nexus of AI research. \u201cThey sit at the front of so many open and closed AI efforts,\u201d Konwinski says. \u201cThis is an unprecedented signal of their belief in openness.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"
Save StorySave this storySave StorySave this story Nvidia will spend $26 billion over the next five years to build open source artificial intelligence models, according to a 2025 financial filing. Executives confirmed the news, which has not been previously reported, in interviews with WIRED. The sizable investment could see Nvidia evolve from a chipmaker with […]<\/p>\n","protected":false},"author":1,"featured_media":47383,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[36],"tags":[],"class_list":{"0":"post-47382","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-business"},"_links":{"self":[{"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/posts\/47382","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/comments?post=47382"}],"version-history":[{"count":0,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/posts\/47382\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/media\/47383"}],"wp:attachment":[{"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/media?parent=47382"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/categories?post=47382"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/tags?post=47382"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}