Jensen-Huang-Will-Nvidias-moat-persist-zh
Jensen Huang – Will Nvidia's moat persist?
核心论点：电力转Token
论点：Nvidia 的本质是"把电力转化为有价值的 token"，这个转化过程难以被商品化。
"In the end, something has to transform electrons to tokens. The transformation of electrons to tokens and making those tokens more valuable over time is hard to completely commoditize."
"最终，总有什么东西必须把电力转换成 token。把电力转换成 token，并让这些 token 越来越有价值，这很难被完全商品化。"
商业模式原则："尽可能多做必要的事，尽可能少做"——能合作就不自己做。
"Our job is to do as much as necessary and as little as possible to enable that transformation to be done at incredible capabilities."
"我们的工作是做尽可能多必要的事，同时尽可能少做，以实现这种难以置信的高效转换。"
护城河一：供应链控制（采购承诺）
论点：$1000-2500 亿采购承诺锁定了稀缺组件，上游供应商愿意投资是因为 Nvidia 庞大的下游需求。
"It's one of the things that we can do that is hard for someone else to do. We've made enormous commitments upstream."
"这是我们能做而别人难以做到的事情之一。我们在上游做出了大量承诺。"
供应商投资的原因：
"The reason for that is because they know that I have the capacity to buy their supply and sell it through my downstream. The fact is that Nvidia's downstream supply chain and our downstream demand is so large, they're willing to make the investment upstream."
"原因是他们知道我有能力购买他们的供应并通过我的下游销售。事实是，英伟达的下游供应链和我们的下游需求如此之大，他们才愿意在上游投资。"
护城河二：GTC 生态系统连接器
论点：GTC 大会将上下游聚集在一起，让供应商看见真实需求，从而愿意提前投资。
"I bring them together so that the downstream can see the upstream, the upstream can see the downstream, and all of them can see the advances in AI."
"我把他们聚在一起，以便下游能看到上游，上游能看到下游，他们都能看到 AI 的所有进步。"
护城河三：瓶颈是临时的（2-3年）
论点：所有瓶颈都会在 2-3 年内被解决，因为需求信号会驱动供应链扩张。
"My point is that none of the bottlenecks last longer than a couple of years, two, three years, none of them."
"我的意思是，没有一个瓶颈会持续超过几年，两三年，没有一个。"
最难解决的瓶颈是"水管工"（技术工人）：
"Plumbers. Plumbers and electricians. This is one of the concerns that I have about the doomers describing the end of work and killing of jobs."
"水管工。管道工和电工。我担心的事情之一是末日论者描述的工作结束和岗位消失。"
护城河四：CUDA 生态系统的丰富性
论点：CUDA 生态系统是 Nvidia 最宝贵的资产，包括数亿 GPU 装机量、每个云、所有框架。
"Nvidia's CUDA ecosystem is ultimately its great treasure. We have several hundred million GPUs out there now. Every cloud has it."
"英伟达的 CUDA 生态系统最终是其伟大的宝藏。我们现在有几亿 GPU 在外面。每个云都有。"
安装基础的重要性：
"The second thing is, if you're a developer building anything at all, the single most important thing you want is an install base. You want the software you write to run on a whole bunch of other computers."
"第二：如果你是一个开发者构建任何东西，你最想要的是安装基础。你希望你的软件在其他大量电脑上运行。"
可编程性允许新算法发明：
"The ability to invent new algorithms is really what makes AI advance so quickly."
"发明新算法的能力真的让 AI 如此快速进步。"
护城河五：TCO（总拥有成本）最优
论点：Nvidia 提供世界上最佳的每 TCO 性能，无出其右。
"Nvidia's computing stack is the best performance per TCO in the world, bar none. Nobody can demonstrate to me that any single platform in the world today has a better performance-TCO ratio."
"英伟达的計算堆栈是世界上每 TCO 性能最好的，无出其右。没有人能向我证明世界上任何单一平台有更好的性能 TCO 比率。"
Benchmark 挑战：
"I would welcome Trainium to demonstrate their 40% that they claim all the time. I would love to hear them demonstrate the cost advantage of TPUs. It makes no sense in my mind."
"我很想听他们展示 TPUs 的成本优势。在我看来这毫无意义。"
即使在固定模型推理场景下，Nvidia 仍然最优：
"Even if you had to run inference on a fixed model, you'd want to run it on the most efficient hardware. Our TCO is the best, bar none. So even in that scenario, you'd want to use Nvidia."
"即使你必须在固定模型上运行推理，你也会想在最有效的硬件上运行。我们的 TCO 是最好的，无出其右。所以即使在那种情况下，你也会想用英伟达。"
护城河六：ASIC 无法取代 GPU
论点：ASIC 为今天优化，但 AI 领域变化太快，固定功能芯片跟不上。
"If you're an ASIC, you're optimized for today. If the domain changes, you're stuck. If the domain changes, you have to build a new ASIC."
"如果你是一个 ASIC，你为今天优化。如果领域变了，你就卡住了。如果领域变了，你必须构建一个新的 ASIC。"
可编程性是关键：
"That's why we invest so heavily in CUDA, in the ecosystem, in the programmability. That's why we can adapt to new algorithms, new architectures, new memory systems, new interconnects, so quickly. ASICs just can't do that."
"这就是为什么我们如此大力投资 CUDA，投资生态系统、投资可编程性。这就是为什么我们能够如此快速地适应新算法、新架构、新内存系统、新互连。ASIC 就是做不到。"
加速计算 vs 专用芯片：
"Accelerated computing is used for all kinds of things... It's also used for fluid dynamics and particle physics. In addition, we use it for AI. Accelerated computing is much more diverse."
"加速计算用于各种事情……它也用于流体力学和粒子物理。此外，我们还用它做 AI。加速计算要多样化得多。"
护城河七：Nvidia 专家支持
论点：Nvidia 工程师深度嵌入客户，帮助优化可获得 2-50 倍性能提升。
"Our expertise helps our AI lab partners to get another 2x out of their stack easily oftentimes. It's not unusual that by the time we're done optimizing their stack or optimizing a particular kernel, their model sped up by 3x, 2x, 50%."
"我们的专业知识帮助我们的 AI 实验室合作伙伴轻松地从他们的堆栈中再获得 2 倍的性能不足为奇。到我们完成优化他们的堆栈或优化某个特定内核时，他们的模型提速 3 倍、2 倍、50% 并不罕见。"
GPU 如 F1 赛车，需要专业驾驶：
"Nvidia's GPUs, accelerators, are like F1 racers. I could imagine everybody's able to drive it at a hundred miles an hour, but it takes quite a bit of expertise to be able to push it to the limit."
"英伟达的 GPU、加速器就像一级方程式赛车。我能想象每个人都能以一百英里一小时的速度驾驶，但需要相当多的专业知识才能将其推到极限。"
护城河八：2006年 CUDA 转型
论点：Nvidia 早在 2006 年就开始投资 CUDA，经过多年坚持才建立今天的生态。
"We did it in 2006. We started CUDA in 2006. That was the pivot that allowed us to be in this position. But it took a long time. It took many years. It took a lot of investment. It took a lot of patience. It took a lot of belief. We had to stick with it even when it was really hard."
"我们在 2006 年做了。我们2006 年启动了 CUDA。那次转型使我们能够处于这个位置。但这需要很长时间。需要很多年。需要大量投资。需要大量耐心。需要大量信念。我们不得不坚持，即使真的很艰难。"
关于出口管制的观点
核心论点：出口管制没有阻止中国 AI 发展，反而迫使中国建立自己的生态，自力更生。
"The export controls have not prevented China from advancing AI. They've actually accelerated it. They've forced China to become self-reliant. They've forced China to innovate. They've forced China to build their own ecosystems."
"出口管制没有阻止中国推进 AI。它们实际上加速了它。它们迫使中国变得自力更生。它们迫使中国创新。它们迫使中国建立自己的生态系统。"
华为的竞争力：
"They have plenty of logic, and they have plenty of HBM2 memory. They're building as fast as they can. They're going to continue to advance."
"他们有大量逻辑芯片，他们也有大量 HBM2 内存。他们在尽快建设。他们将继续推进。"
核心主张：
- "We should always be first and we should always be ahead."
- "We should never give up a market. We should never give up a customer."
- "We should never do anything that helps our competitors."
公司哲学：最大化影响而非利润率
不追求利润最大化的原因：
"We want to make sure that the technology is accessible to as many people as possible. We want to democratize AI. We want to make sure that everyone can benefit from it."
"我们想让 AI 民主化。我们想确保每个人都能从中受益。"
中性平台策略：
"We want to be neutral. We want to be the partner of every AI company. We want to be the partner of every AI researcher. We want to be the partner of every AI developer."
"我们想保持中立。我们想做每个 AI 公司的合作伙伴。我们想做每个 AI 研究人员的合作伙伴。我们想做每个 AI 开发者的合作伙伴。"
关键引用汇总
核心业务模式：
"The input is electrons, the output is tokens. In the middle is Nvidia."
CUDA 护城河：
"CUDA 生态系统最终是英伟达伟大的宝藏。"
摩尔定律 vs 算法创新：
"真正获得 10 倍或 100 倍飞跃的唯一方法是每年从根本上改变算法及其计算方式。"
能源才是真正约束：
"没有能源你就不能创建产业。更多芯片产能，那是两三年的问题。"
中国市场50%开发者：
"50% 的 AI 开发者在中国。美国不应该放弃那个。"