使用 Render 的轻量级 SDK 定义任务，并将它们链接成长期运行的分布式工作流。按需启动你的代理和批处理作业。Render Workflows 负责排队、编排和重试。
使用代码 BYTE 获取 50 美元积分：https://go.bytebytego.com/Render_072726
你可能知道 NVIDIA 销售 GPU。但你是否知道，它也是全球最大的开放 AI 模型发布者？其模型在 Hugging Face 上下载量名列前茅，而且阵容远不止聊天机器人：还有推理模型、世界模型、人形机器人模型、自动驾驶汽车模型，甚至药物发现、量子计算和全球预测模型。
这引出了两个问题。一家以硬件闻名的公司，如何在如此多的领域构建出如此强大的模型？以及为什么要免费提供这些模型，而这些模型却运行在 NVIDIA 销售的 GPU 上？
为了理解这两个问题，我们采访了 NVIDIA 应用深度学习研究副总裁 Bryan Catanzaro：https://www.linkedin.com/in/bryancatanzaro/。他向我们介绍了他的团队如何构建公司的开放模型：https://www.nvidia.com/en-us/glossary/open-models/，模型架构的背后原理，以及为什么 NVIDIA 开源了如此多的内容。感谢 Bryan 花时间如此详细地分享团队的工作。
NVIDIA 建立了哪些开放模型，以及它们的功能
NVIDIA 如何让其模型既快速又强大
一个基础如何让小团队构建许多模型
开放的真正含义，以及为什么发布数据和权重同样重要
为什么 NVIDIA 全部免费提供，以及团队在公开发布中学到的经验
NVIDIA 的开放模型位于 AI 的两个端点。一端是处理符号的模型，也就是在屏幕上工作的语言、代码和推理。另一端是处理物理世界的模型，即机器人和自动驾驶车辆（AV）必须感知周围环境并采取行动的模型。大部分阵容都处于这两者之间的某个位置。
观察生态系统的最简单方式是按每个模型家族的用途进行分组。
推理模型是大型语言模型，它们经过训练，可以在给出最终答案之前生成中间的标记。这使它们能够思考诸如数学和编码等任务。如今你看到的大多数前沿模型都是由推理模型驱动的，如 Claude Opus、GLM-5.2、Kimi K2.7。
NVIDIA 的推理模型系列被称为 Nemotron。首个 Nemotron 模型在2023年推出。2024年，NVIDIA 发布了下一代模型，拥有3400亿个参数，随后在2025年推出了更快的混合设计。当前一代 Nemotron 3 于2025年底至2026年推出，提供三个规格：小型 Nano 模型用于快速、简单的任务，中型 Super 模型用于更难的规划和推理，大型 Ultra 模型用于需要最复杂推理的任务。
语言模型擅长预测下一个标记，但它们并不理解物理世界。它们无法告诉你衬衫袖子如何折叠，也无法描述水杯倾倒时的样子。机器人或自动驾驶汽车需要这种理解，但从文本中是无法获得的。
这就是世界模型提供的功能。它不是预测下一个词，而是理解世界，并根据当前世界状态和一个动作预测下一个状态。它可以生成真实的、符合物理规律的视频，推理运动及因果关系，并生成机器人学习所需的动作数据。
世界模型非常有用，因为你可以用它们为那些在现实世界中捕捉缓慢、昂贵或危险的情况创建训练数据。
NVIDIA 的首席执行官黄仁勋称之为“物理 AI 的 ChatGPT 时刻”。NVIDIA 面向物理 AI 的前沿开放模型是 Cosmos，于2025年1月在 CES 发布，并于2026年统一为单一全能模型 Cosmos 3：https://www.nvidia.com/en-us/glossary/omni-model/，它可以生成场景、对场景进行推理并预测接下来会发生什么。Cosmos 3 还引入了世界动作模型：https://www.nvidia.com/en-us/glossary/world-action-model/，将预训练的视频骨干直接转化为机器人策略。还有一个40亿参数的 Edge 版本，小到可以在机器人上运行并实时控制它。
一个世界模型可以预测下一个场景，但机器人必须采取行动。它必须将摄像头看到的内容实时转化为电机的运动，用于那些无人事先编写脚本的任务。
这就是机器人基础模型的工作，经常被称为视觉-语言-动作模型。这些模型的输入通常是摄像头图像和指令，而模型输出的是动作。
NVIDIA 面向类人机器人的开放基础模型系列是 Isaac GR00T。NVIDIA 在 2024 年公布了该项目，并在 2025 年初发布了首个开放类人基础模型 GR00T N1，此后不断迭代到 GR00T 1.7。
GR00T 模型使用 Cosmos 作为推理骨干，为机器人提供感知、推理和操作移动的共享基础：能够在操作物体的同时移动。这使机器人的决策更接近人类，并帮助机器人将复杂动作分解为结构化、逐步的计划，开发者可以针对特定任务进行专业化。1.7 版本还包含适合生产的商业许可，允许开发者将模型部署到实际应用中。
这一系列还在继续。Alpamayo 是面向基于推理的自动驾驶的开放模型家族。像 GR00T 一样，它使用 Cosmos 作为推理骨干，因此驱动机器人的相同物理-AI 基础也能让汽车应对它从未训练过的路况。它还展示了每个决策背后的因果链，开发者可以进行检查和扩展。BioNeMo 将开放模型引入生物学和药物发现领域，BioNeMo 代理工具包让开发者能够为生命科学工作构建自主代理。Ising 为量子计算机的构建者、操作员和开发者提供了 AI 工具，以将设备扩展到容错水平。Earth-2 是一个开放模型、库和框架的家族，使专业级的天气和气候 AI 可供任何人使用。在此基础上，NVIDIA 还提供用于生产的开放模型：过滤不安全内容的安全模型、语音模型以及图像视觉模型。
综合来看，这个强大的生态系统改变了社区推进和构建的方式。文章的其余部分将探讨 NVIDIA 如何构建这些模型，团队如何能够快速前进，以及他们学到了什么。
构建一个既快速又强大的模型通常是一项挑战。强大的模型通常更大，这使它们训练和回答问题的速度更慢。NVIDIA 的解决方案是同时追求前沿性和速度。Bryan 如此描述：“最快的模型就是最聪明的模型。”
虽然这听起来像是一句口号，但它实际上有深刻的见解。更快的模型可以在相同时间内处理更多数据进行训练。它可以在更多环境中进行后续训练。一旦部署，它可以在面对难题时以相同成本思考更久。因此，速度转化为能力，并在每个阶段不断累积。
基于这一策略，NVIDIA 在设计模型时侧重于效率和速度。以下设计选择使这些模型既快速又强大。前两个使模型更快、更高效。最后一个使其更强大。
大多数知名模型几乎完全基于 Transformer 的注意力机制。注意力机制允许每个输入的标记查看所有其他标记，这对于捕捉关系非常强大，但代价高。其成本随着输入长度的平方增长，因此输入长度翻倍，工作量大约增加四倍。模型在运行时必须将每个之前的标记保存在内存中。对非常长的输入，这会变得缓慢且昂贵。
另一类主要模型是状态空间模型，而 Mamba 是其最知名的例子。Mamba 层不是将每个标记与其他标记进行比较，而是一次读取序列并将所见的一切压缩到固定大小的内存中。这使得它的成本较低：成本随着长度线性增长，内存无论输入多长都保持不变。其权衡是固定大小的内存无法保留所有细节，因此 Mamba 在回忆埋藏得很深的具体事实时较弱。
Mamba 层高效，而注意力层更有效。将它们结合在一起可以兼得两者的优势。混合架构用 Mamba 层代替 Transformer 中的大部分注意力层，并保留少量完整注意力块。
NVIDIA的模型依赖于相同的混合模式。大部分层是Mamba，这使处理长输入的成本保持较低，也正是它让百万标记的上下文窗口成为可行而不仅仅是理论上可行。一些注意力层被插入其中，以恢复Mamba舍弃的精确回忆能力，使模型仍然能够从上下文中的任何位置检索到精确细节。
在混合设计的基础上，模型还使用了专家混合层（MoE）。一个MoE层由许多小型专家层组成，并将每个标记路由到其中的少数几个层。因此，模型并不是对每个标记运行所有参数，而是只激活其参数的一小部分。这使我们能够以低标记成本使用大的总容量，从而保持模型运行速度快。
按Bryan的描述，Mamba构建了整个序列的全局视图。注意力机制总是可以回溯并找到精确的事实。结合起来，它们比单独使用任何一个都更智能。NVIDIA在2024年发布了这一混合成果，其他实验室，包括Qwen和Kimi团队，也随后走向了相同方向。
第二个选择是NVIDIA两大业务的交汇点。其更大的模型使用一种名为NVFP4的4位数格式进行预训练，这意味着大部分训练计算每个数值只使用四位。位数少意味着内存占用少、数据传输少，因此运算更快且功耗更低。
需要注意的是，四位意味着精度非常低。大多数团队会用较高精度训练，之后再缩小模型，这会损失准确性。NVIDIA从第一步就使用4位训练，因为它知道下一代GPU Blackwell将配备快速4位硬件。模型与芯片是为彼此设计的。
副总裁描述了背后的思维方式。摩尔定律已经不再能在每一代轻松带来收益，所以进展必须来自于模型与硬件的共同设计。团队之所以尝试4位预训练，是因为他们相信这一结果是可能的。他的话是，发明出惊人事物的第一步就是相信你能做到。
在做出这些效率选择之后，模型在后训练阶段获得其能力。
后训练配方本身是大多数其他团队所趋同的。它从监督微调开始，模型从经过精心挑选的优秀答案示例中学习，掌握所期望的格式和行为。然后，它转向强化学习，模型在真实任务中练习，并因达成正确结果而获得奖励。它不是从示例中学习，而是从自己的尝试中学习。
已证明扩大这一强化学习阶段可以显著提高模型的能力。在这里，NVIDIA 对效率的关注再次得到回报。由于其模型运行成本较低，它可以以较低成本扩展强化学习，让模型在多种环境中并行练习，并从超过一百万次的试验中学习。
这正是副总裁指出更好模型的真正瓶颈所在。不是数据，也不是计算资源。而是这些训练环境的多样性。模型在越多样的情况下练习，就变得越有能力，因为这正是它获得实际问题解决经验的地方。
到目前为止，我们已经了解了构建模型的设计决策，以及训练单个前沿模型所涉及的内容。但他们是如何构建如此多的模型，并如此迅速地持续改进它们的呢？
在一个领域构建强大的模型是困难的。在语言、视频、机器人、车辆和生物学等多个领域构建模型需要巨大的工程和计算资源。NVIDIA 的方法是确定一个可复用的基础，先构建一次，然后在不同的项目中重复使用。
副总裁将这一方法追溯到 CUDA，这是使 NVIDIA GPU 可编程的软件层。“U”代表统一，他称这是最重要的字母。其理念是构建一个基础，并在其上运行所有内容，而不是维护多个独立的堆栈。
同样的本能也塑造了这些模型。上一节介绍的混合架构不会为每个模型重新构建。相同的骨干网络在所有三种 Nemotron 尺寸中都能扩展。整个组件也会在不同项目间共享。Cosmos Reason，这是一个用于推理物理世界的模型，被重新作为 NVIDIA Isaac GR00T 基础模型的推理核心使用，因此机器人团队不必从零开始构建这一部分。另一种方式是十个平行项目各自重复发明相同的部分，这会使人员和计算资源过于分散。据 Bryan 说，NVIDIA 试图尽可能“懒惰”，因为这是在有限资源下获得最大效益的方式。
团队效率高、迭代快的另一个原因是文化。NVIDIA 大多不是自上而下的公司；它是一个志愿者组成的公司，团队可以选择自己的工作方向，统一是通过邀请而非控制实现的。协作文化认为，如果团队相互合作并完成更多工作，他们将获得比单独行动更多的资源。当每个人都共享这种信念时，一个小团队也能产生大量成果。
你可能听说过很多关于开放模型的事情。几个 AI 实验室发布了功能强大的模型。你可以下载权重并使用它们。但那并不是真正的开放模型。这里是 NVIDIA 方法与其他实验室显著区别的地方。正如 Bryan 所说，Nemotron 不仅仅是一个模型。它还包括数据、工具、训练配方，以及论文中公开的思想。最终的权重只是最后一步。
实际上，这意味着 NVIDIA 会发布训练数据集、训练后数据集、强化学习环境，以及重现工作所需的配方。团队不仅可以运行模型，还能看到它是如何构建的，并训练自己的版本。
对于机器人和自动驾驶汽车也是如此，开发者可以使用开源模拟和学习框架，以及数据集来训练和评估模型性能，然后再部署。
以这种规模发布数据是不寻常的。对一家公司来说，这是一件令人害怕的事情。但 NVIDIA 仍然会这样做，甚至还分享了巧妙的构建模块，比如统计上真实但完全隐私的合成“人格”数据，这样模型可以学习世界知识而无需记住真实的人。
你可以从社区对它所做的事情中看到这一点。NVIDIA 的开放数据和模型催生了大量的社区微调和衍生作品。其中一个机器人数据集成为 Hugging Face 上下载量最高的数据集之一。其 Nemotron 模型的许多变体每个都有数百万次下载，而 Nemotron 系列最近总下载量超过了 1 亿。仅凭权重，你无法构建出这样的成果。
你可能认为发布开放模型听起来很棒。但是，为什么一家销售 GPU 的公司会免费提供在这些 GPU 上运行的 AI？
Bryan 分享了两个原因。第一个是 NVIDIA 需要深入理解 AI 才能为其构建合适的硬件。为了设计未来的芯片，你必须知道 AI 的发展方向。真正了解的最好方式是自己构建模型并将其展示给真实用户。保持模型私有很容易自欺，因此公开发布能保持工作的透明度，使 NVIDIA 的研究人员与他们从中学习的更广泛社区保持联系。开放模型如此热门，其真正的价值在于让开发者能够以极低的成本对这些模型进行后续训练。
第二个原因涉及商业。每当 AI 发展，NVIDIA 也随之增长。其目标是支持生态系统，而不是与构建在其上的公司竞争。每一个采用开放模型并构建自己 AI 的团队都可能成为未来的计算客户。因此，公司宁愿把模型交给生态系统，让全球对 AI 的需求推动其业务发展。
还有一个附带好处：安全。开放技术往往更安全，因为更多人可以检查它并发现问题，就像开放的互联网受到所有用户的强化一样。
公开发布如此多的模型可以提供封闭实验室永远无法面对的教训。Bryan 分享了几个宝贵的经验。
在这个发展如此迅速的领域，每一个模型在发布前就已经过时。因此，没有单一的发布是关键。获得信任的是一个稳定、持续的计划。当社区相信一个模型系列将长期存在时，他们会关注每一次新发布，而不会把它当作一次性的事件。
一次干净的发布是许多小事情都做对的结果，从许可到文档再到可直接运行的示例。副总裁承认，这是不引人注目的部分。倾听社区反馈并优化第一次运行体验，才会把一次下载转化为持久使用。
NVIDIA 放弃了自家的原生许可证，转而使用 OpenMDW，这是 Linux 基金会针对开放 AI 专门制定的社区许可证。使用一个共享的、专为目的设计的许可证可以减少摩擦，并表明这些模型是为了自由使用的。
接下来的方向是更早地开放。NVIDIA 不打算私下构建模型再在完成后发布，而是希望更早地引入合作伙伴，因为它计划无论如何都会开放所有内容。目标是让合作伙伴在模型仍在构建时就参与塑造模型。为实现这一点，NVIDIA 组建了 Nemotron 和 Cosmos 联盟。
最显著的是这种信念：这是长期承诺，而不是一个阶段性行为。副总裁提到了 CUDA，它花了十多年时间，世界其他地方才明白它的重要性。NVIDIA 坚持下去，它成为了公司的基础。
Define tasks with Render’s lightweight SDK and chain them into long-running, distributed workflows. Launch your agents and batch jobs on demand. Render Workflows handles queuing, orchestration, and retries.
Get $50 credits with code BYTE ：https://go.bytebytego.com/Render_072726
You probably know NVIDIA sells GPUs. But did you know it is also the largest publisher of open AI models in the world ? Its models rank among the most downloaded on Hugging Face, and the lineup goes way beyond chatbots: reasoning models, world models, models for humanoid robots, self-driving cars, even drug discovery, quantum computing, and global forecasting.
This raises two questions. How does a company best known for hardware build so many strong models across so many domains? And why give them away for free, when those same models run on the GPUs NVIDIA sells?
To understand both, we spoke with Bryan Catanzaro ：https://www.linkedin.com/in/bryancatanzaro/ , VP of Applied Deep Learning Research at NVIDIA. He walked us through how his team builds the company’s open models ：https://www.nvidia.com/en-us/glossary/open-models/ , the reasoning behind their architecture, and why NVIDIA open-sources so much of it. Our thanks to Bryan for taking the time to share the team’s work in such detail.
What open models NVIDIA has built, and what they do
How NVIDIA makes its models both fast and powerful
How one foundation lets a small team build many models
What open really means and why releasing the data matters as much as the weights
Why NVIDIA gives it all away, and the lessons the team has learned shipping in the open
NVIDIA’s open models sit at two ends of AI. At one end are models that work with symbols, the language, code, and reasoning that lives on a screen. At the other are models that work in the physical world, where robots and autonomous vehicles (AVs) have to perceive their surroundings and act. Most of the lineup falls somewhere along that line.
The simplest way to see the ecosystem is to group them by what each family is for.
Reasoning models are large language models that are trained to produce intermediate tokens before the final answer. This is what lets them think through tasks such as math and coding. Most frontier models you see these days are powered by a reasoning model. Claude Opus, GLM-5.2, Kimi K2.7.
NVIDIA’s reasoning model family is called Nemotron. The first Nemotron model was introduced in 2023. In 2024, NVIDIA released the next generation with 340 billion parameters, followed by a faster hybrid design in 2025. The current generation, Nemotron 3, arrived across late 2025 and 2026 in three sizes. A small Nano model for fast, simple tasks, a mid-size Super model for harder planning and reasoning, and a large Ultra model for complex tasks that need the heaviest reasoning.
Language models are great at predicting the next token, but they do not understand the physical world. They cannot tell you how a shirt sleeve folds, or what a glass of water looks like the moment it tips over. A robot or a self-driving car needs that kind of understanding, but it cannot get it from text.
That is what a world model provides. Instead of predicting the next word, it understands the world and can predict the next state given the current state of the world and an action. It can generate realistic, physically plausible video, reason about motion and cause and effect, and produce the action data a robot needs to learn.
World models are extremely useful because you can use them to create training data for situations that are slow, expensive, or dangerous to capture in the real world.
NVIDIA’s CEO, Jensen Huang, has called this “the ChatGPT moment for physical AI.” NVIDIA’s frontier open model for physical AI is Cosmos, launched at CES in January 2025 and unified in 2026 into a single omni-model ：https://www.nvidia.com/en-us/glossary/omni-model/ , Cosmos 3, that can generate a scene, reason about it, and predict what comes next. Cosmos 3 also introduces world action models ：https://www.nvidia.com/en-us/glossary/world-action-model/ , which turn the pretrained video backbones directly into robot policies. There is also a 4-billion-parameter Edge version small enough to run on the robot and control it in real time.
A world model can predict the next scene, but a robot has to act. It has to turn what its cameras see into how its motors move, in real time, on tasks nobody scripted in advance.
That is the job of a robot foundation model, often called a vision-language-action model. The inputs to these models are usually camera images and instructions, and the model produces the movement.
NVIDIA’s open foundation model line for humanoid robots is Isaac GR00T. NVIDIA unveiled the project in 2024, released the first open humanoid foundation model, GR00T N1, in early 2025, and has since iterated to GR00T 1.7.
GR00T models use Cosmos as a reasoning backbone, giving robots a shared foundation for perception, reasoning, and locomanipulation: the ability to move while manipulating objects. This enables more human-like decision-making and helps robots break complex actions into structured, step-by-step plans that developers can specialize for specific tasks. The 1.7 release also includes a production-friendly commercial license, allowing developers to deploy the model in real-world applications.
The lineup keeps going. Alpamayo is an open model family for reasoning-based self-driving. Like GR00T, it uses Cosmos as its reasoning backbone, so the same physical-AI foundation that powers robots also lets a car work through road situations it was never trained on. It also exposes the chain of cause and effect behind each decision, which developers can inspect and build on. BioNeMo brings open models to biology and drug discovery, and the BioNeMo Agent Toolkit lets developers build autonomous agents for life sciences work. Ising gives quantum computer builders, operators, and developers the AI tools needed to scale devices to fault tolerance. Earth-2 is a family of open models, libraries, and frameworks that makes professional-grade weather and climate AI available to anyone. On top of these, NVIDIA offers open models for production use: safety models that filter unsafe content, speech models for voice, and vision models for images.
Taken together, this powerful ecosystem changes how the community can advance and build. The rest of the article looks at how NVIDIA builds these models, how the team is able to move fast, and what they have learned.
It is often a challenge to build a model that is both fast and capable. Capable models are usually larger, which makes them slower to train and to answer questions. NVIDIA’s answer is to chase the frontier and speed at the same time. Bryan put it this way: “the fastest model is the smartest model.”
While this may sound like a slogan, it has a real insight. A faster model can be trained on more data in the same amount of time. It can be post-trained across more environments. Once deployed, it can think for longer on a hard problem at the same cost. So speed translates into capability and compounds at every stage.
Given this strategy, NVIDIA designs its models with a focus on efficiency and speed. The following design choices are what make these models both fast and capable. The first two make the model faster and more efficient. The last one makes it more capable.
Most well-known models are built almost entirely on the Transformer’s attention mechanism. Attention lets every token in the input look at every other token, which is powerful for capturing relationships but expensive. The cost grows with the square of the input length, so doubling the input roughly quadruples the work. The model has to keep every earlier token in memory as it runs. On very long inputs, that gets slow and costly.
The other main class is the state space model, and Mamba is its best-known example. Instead of comparing every token with every other one, a Mamba layer reads the sequence once and compresses everything it has seen into a fixed-size memory. That makes it cheap: the cost grows linearly with length, and the memory stays constant no matter how long the input. The tradeoff is that a fixed-size memory cannot keep every detail, so Mamba is weaker at recalling one exact fact buried far back.
Mamba layers are efficient while attention layers are more effective. Combining them gives the best of both worlds. A hybrid architecture swaps most of the attention layers in the Transformer with Mamba layers and keeps a few blocks with full attention.
NVIDIA’s models rely on the same hybrid pattern. Most of the layers are Mamba, which keeps long inputs cheap to process and is what makes a million-token context window practical rather than just theoretical. A few attention layers are placed in between to restore the pinpoint recall that Mamba gives up, so the model can still retrieve exact details from anywhere in the context.
On top of the hybrid design, the models also use mixture-of-experts, or MoE, layers. An MoE layer consists of many small expert layers and routes each token to only a few of them. So instead of running every parameter on every token, the model activates only a small subset of its parameters. That allows us to use a large total capacity at a low per-token cost, which keeps the model fast.
The way Bryan describes it, Mamba builds a global view of the whole sequence. Attention can always go back and find an exact fact. Put together, they are smarter than either one alone. NVIDIA published this hybrid result in 2024, and other labs, including the teams behind Qwen and Kimi, have since moved in the same direction.
The second choice is where NVIDIA’s two businesses meet. Its larger models are pretrained in a 4-bit number format called NVFP4, meaning most of the math during training uses just four bits per value. Fewer bits means less memory and less data to move, so the math runs faster and on less power.
Note that four bits is very little precision. Most teams train in higher precision and only shrink the model afterward, which loses accuracy. NVIDIA trained in 4 bits from the first step because it knew its next GPU generation, Blackwell, was being built with fast 4-bit hardware. The model and the chip were designed for each other.
The VP described the mindset behind it. Moore’s law is no longer giving easy gains every generation, so progress has to come from designing the model and the hardware together. The team only attempted 4-bit pretraining because it believed the result was possible. In his words, the first step to inventing something amazing is believing that you can.
With those efficiency choices in place, post-training is where the model gains its capability.
The post-training recipe itself is what most other teams converged to. It starts with supervised fine-tuning, where the model learns from curated examples of good answers and picks up the format and behavior expected of it. Then it switches to reinforcement learning, where the model practices on real tasks and is rewarded for reaching the right outcome. Instead of learning from examples, it learns from its own attempts.
Scaling up this reinforcement learning stage has been shown to improve a model’s capability significantly. Here NVIDIA’s efficiency focus pays off again. Because its models are cheaper to run, it can scale RL at a lower cost, letting the model practice across many environments in parallel and learn from over a million rollouts.
This is where the VP placed the real bottleneck for better models. Not data or compute. The diversity of these training environments. The more varied the situations a model practices in, the more capable it becomes, because that is where it gains real problem-solving experience.
So far we have looked at design decisions to build a model and what is involved to train a single frontier model. But how do they build so many models, and keep improving them so quickly?
Building strong models in one domain is hard. Building them across language, video, robotics, vehicles, and biology needs huge engineering and compute resources. NVIDIA’s approach is to identify a reusable foundation, build it once, and reuse it across different efforts.
The VP traced this back to CUDA, the software layer that made NVIDIA’s GPUs programmable. The “U” stands for Unified, and he called it the most important letter. The idea is to build one foundation and run everything on top of it, rather than maintaining many separate stacks.
The same instinct shapes the models. The hybrid architecture from the last section is not rebuilt for each model. The same backbone scales across all three Nemotron sizes. Whole components cross between efforts too. Cosmos Reason, the model built to reason about the physical world, was reused as the reasoning core inside NVIDIA’s Isaac GR00T foundation model, so the robotics team did not have to build that part from scratch. The alternative, ten parallel efforts each reinventing the same pieces, would spread people and compute too thin. According to Bryan, NVIDIA tries to be as lazy as possible, because that is how it gets the most out of limited resources.
The other reason the team is so productive and iterates quickly is cultural. NVIDIA is mostly not top-down; it is a company of volunteers, where teams choose what to work on, and unification happens through invitation rather than control. The culture of collaboration is that if teams work with each other and get more done, they will get more resources than if they each go alone. When everyone shares that belief, a small group can produce a lot.
You may have heard a lot about open models. Several AI labs released capable models. You can download the weights and use them. But that is not truly an open model. This is where NVIDIA’s approach really separates it from the other labs. As Bryan puts it, Nemotron is not just a model. It is also the data, the tools, the training recipes, and the ideas published in the papers. The finished weights are just the last step.
In practice that means NVIDIA publishes the training datasets, the post-training datasets, the reinforcement learning environments, and the recipes needed to reproduce the work. A team can not only run the model, but also see how it was built and train their own version.
The same is true for robotics and self-driving cars, where developers can use open source simulation and learning frameworks, datasets to train and evaluate the model performance before deployment.
Releasing data at this scale is unusual. It is a scary thing for a company to do. NVIDIA does it anyway, and even shares clever building blocks like synthetic “persona” data that is statistically realistic but fully private, so a model can learn about the world without memorizing real people.
You can see this in what the community has done with it. NVIDIA’s open data and models have spawned a large set of community fine-tunes and derivatives. One of its robotics datasets became one of the most-downloaded datasets on Hugging Face. Many variants of its Nemotron models have millions of downloads each, and the Nemotron family recently crossed 100M in total. You cannot build like that from the weights alone.
You may think releasing open models sounds great. But why would a company that sells GPUs give away the AI that runs on them?
Bryan shared two reasons. The first is that NVIDIA needs to understand AI deeply to build the right hardware for it. To design future chips, you have to know where AI is going. The best way to really know is to build models yourself and put them in front of real users. A model kept private is easy to fool yourself about, so releasing it openly keeps the work honest and keeps NVIDIA’s researchers connected to the wider community they learn from. With open models being such a hot topic, the real value is in enabling developers to post-train these models themselves at a fraction of the cost.
The second reason is about the business. NVIDIA grows whenever AI grows. Its goal is to support the ecosystem, not to compete with the companies building on top of it. Every team that adopts open models and builds its own AI can become a future customer for compute. So the company would rather hand the ecosystem the models and let the world’s demand for AI pull its business forward.
There is also a side benefit: safety. Open technology tends to be safer because more people can inspect it and find problems, the way the open internet is hardened by everyone who uses it.
Shipping this many models in public teaches lessons that closed labs never face. Bryan shared several valuable lessons.
In a field moving this fast, every model is obsolete before it ships. So no single release is the point. What earns trust is a steady, sustained program. When the community believes a model line is here to stay, it pays attention to each new release instead of treating it as a one-off.
A clean launch is the result of getting a hundred small things right, from licensing to documentation to ready-to-run examples. The VP admitted this is the unglamorous part. Listening to community feedback and smoothing that first-run experience is what turns a download into lasting use.
NVIDIA moved away from its own homegrown licenses to OpenMDW, a community license from the Linux Foundation built specifically for open AI. Using a shared, purpose-built license removes friction and signals that the models are meant to be used freely.
The direction from here is to open even earlier. Rather than build a model privately and post it when it is done, NVIDIA wants to bring collaborators in sooner, since it plans to open everything anyway. The aim is to let partners shape the models while they are still being built. To make this happen, NVIDIA formed the Nemotron and Cosmos Coalitions.
What stands out most is the conviction that this is a long-term commitment, not a phase. The VP pointed to CUDA, which took more than a decade before the rest of the world understood why it mattered. NVIDIA stuck with it, and it became the foundation of the company.
扫码直达文章详情