我的征尘是星辰大海。。。
The dirt and dust from my pilgrimage forms oceans of stars...
-------当记忆的篇章变得零碎，当追忆的图片变得模糊，我们只能求助于数字存储的永恒的回忆
作者:黄教授
手机视频列表
华为韬定律不是弯道超车
视频
音频
原始脚本
华为涛定律不是弯道超车，是被逼出来的半导体新路。 最近华为提出的涛涛，希腊字母涛，定律刷屏了。 很多人觉得是黑科技，但说白了，这不是华为独有的聪明，是西方有捷径不走，华为没路可走只能硬闯的必然结果。 今天用大白话把这事说透，人人都能懂。 一、先懂统治芯片60年的摩尔定律，彻底走不动了。 我们手机、电脑越来越快。 核心靠摩尔定律，把晶体管越做越小，几何缩微，每18~24个月数量翻倍，性能翻倍。 以前工艺越先进。 芯片越便宜越快，躺着赚钱。 现在2026年，卡在3纳米、2纳米，物理极限到了，晶体管小到原子级别。 电子会乱跑，量子隧穿，漏电，散热炸，更要命，成本爆炸！设计一颗顶尖芯片超10亿美元，建3纳米工厂要200亿美元。 越小越贵，性价比彻底反转。 简单说，摩尔定律的几何缩微路走到头了。 二、华为涛定律，不拼尺寸小。 打拼时间快。 2026年5月25日，华为正式提出涛定律，核心就一句话，用时间缩微替代几何缩微。 涛。 涛，电路里的时间常数，就是信号在芯片里传输的延迟。 目标，不把晶体管做更小，而是让信号跑的距离更短，延迟更低，照样提性能。 通俗比喻，一看就懂。 摩尔定律，把房子里的家具晶体管做极小，塞更多家具，靠多提效率。 涛定律，把平面房子折成3D高楼。 逻辑折叠，原本隔10米的家具，现在上下叠在一起，信号不用绕路，走一米就到。 靠快提效率，核心手段。 四层全栈挖潜，不是只有三 d 一、器件层，优化材料，减少信号阻力。 二、电路层，核心。 逻辑折叠加三 d 堆叠，关键路径缩短60%，延迟大降。 三、芯片层，软硬件协同，指令调度更高效。 四、系统层，新总线协议，芯片间通信更快。 三、关键真相，这技术谁都能做，但只有华为必须做。 很多人以为涛定律是华为独创，其实原理全行业都懂，技术西方早就有，如3D 堆叠、先进封装，但西方大厂不做。 不想做，做不了。 一、西方有捷径，制程走，懒得多绕路。 过去几十年靠先进制程，7纳米、5纳米、3纳米就能躺赢。 闺女有花戴，何必找红头绳？惯性加懒惰，全产业链。 EDA、设备、软件都围绕平面，缩微造改，3D等于推翻整个生态。 投入大，风险高，资本市场不答应。 二、英特尔等历史包袱太重，改不起。 X86芯片40年向下兼容。 老程序必须能跑，架构大改等于兼容性崩溃，windows 工业软件全受影响，商业代价天文数字，不是不能改，是改不起。 不敢改。 三、华为被逼上梁山，没路可走，只能硬闯。 华为被断先进制程，进一 uv 光刻机。 等于闺女没花戴了。 不走套定律，时间压缩，3D 折叠，就没芯片可用，没手机可卖，没业务可活，反而没历史包袱。 不用兼容老架构，敢推倒重来，把可做可不做的事做成了必须做、优先做的核心战略。 四、一句话总结两种路线。 两种命运。 摩尔定律，西方拼制成，拼尺寸，拼 EUV 有捷径，有包袱，躺习惯了。 涛定律。 华为拼架构、拼时间、拼挖钱，没捷径、没包袱，被逼出来。 本质不是华为更聪明，是绝境逼出了新赛道。 不是西方做不到，是他们不想革自己的命。 五、对我们普通人意味着什么？涛定律不是替代摩尔定律，是互补。 西方继续啃先进制成的硬骨头，我们靠架构挖潜，另起炉灶。 未来不用最先进制成，也能做出高性能芯片，手机、国产设备性能持续提升，不再被卡脖子。 最后。 技术从来不是谁的专属，谁被逼到墙角，谁敢打破惯性，谁就能走出新路。 华为的涛定律就是最好的例子。
修正脚本
华为陶定律不是弯道超车，是被逼出来的半导体新路。 最近华为提出的陶，希腊字母陶，定律刷屏了。 很多人觉得是黑科技，但说白了，这不是华为独有的聪明，是西方有捷径不走，华为没路可走只能硬闯的必然结果。 今天用大白话把这事说透，人人都能懂。 一、先懂统治芯片60年的摩尔定律，彻底走不动了。 我们手机、电脑越来越快。 核心靠摩尔定律，把晶体管越做越小，几何缩微，每18~24个月数量翻倍，性能翻倍。 以前工艺越先进，芯片越便宜越快，躺着赚钱。 现在2026年，卡在3纳米、2纳米，物理极限到了，晶体管小到原子级别。 电子会乱跑，量子隧穿，漏电，散热炸，更要命，成本爆炸！设计一颗顶尖芯片超10亿美元，建3纳米工厂要200亿美元。 越小越贵，性价比彻底反转。 简单说，摩尔定律的几何缩微路走到头了。 二、华为陶定律，不拼尺寸小，只拼时间快。 2026年5月25日，华为正式提出陶定律，核心就一句话，用时间缩微替代几何缩微。 陶是电路里的时间常数，就是信号在芯片里传输的延迟。 目标，不把晶体管做更小，而是让信号跑得距离更短，延迟更低，照样提性能。 通俗比喻，一看就懂。 摩尔定律，把房子里的家具晶体管做更小，塞更多家具，靠数量提效率。 陶定律，把平面房子折成3D高楼。 逻辑折叠，原本隔10米的家具，现在上下叠在一起，信号不用绕路，走一米就到。 靠快提效率，核心手段，四层全栈挖潜，不只是靠三 d：一、器件层，优化材料，减少信号阻力。 二、电路层，核心。 逻辑折叠加三 d 堆叠，关键路径缩短60%，延迟大降。 三、芯片层，软硬件协同，指令调度更高效。 四、系统层，新总线协议，芯片间通信更快。 三、关键真相，这技术谁都能做，但只有华为必须做。 很多人以为陶定律是华为独创，其实原理全行业都懂，技术西方早就有，如3D 堆叠、先进封装，但西方大厂不做。 不想做，做不了。 一、西方有捷径，一直走，懒得多绕路。 过去几十年靠先进制程，7纳米、5纳米、3纳米就能躺赢。 闺女有花戴，何必找红头绳？惯性加懒惰，全产业链。 EDA、设备、软件都围绕平面缩微打造，3D 等于推翻整个生态。 投入大，风险高，资本市场不答应。 二、英特尔等历史包袱太重，改不起。 X86芯片40年向下兼容。 老程序必须能跑，架构大改等于兼容性崩溃，windows 工业软件全受影响，商业代价天文数字，不是不能改，是改不起。 不敢改。 三、华为被逼上梁山，没路可走，只能硬闯。 华为被断先进制程，缺了 EUV 光刻机。 等于闺女没花戴了。 不走陶定律，时间缩微，3D 折叠，就没芯片可用，没手机可卖，没业务可活，反而没历史包袱。 不用兼容老架构，敢推倒重来，把可做可不做的事做成了必须做、优先做的核心战略。 四、一句话总结两种路线。 两种命运。 摩尔定律，西方拼制程，拼尺寸，拼 EUV 有捷径，有包袱，躺习惯了。 陶定律。 华为拼架构、拼时间、拼挖潜，没捷径、没包袱，被逼出来。 本质不是华为更聪明，是绝境逼出了新赛道。 不是西方做不到，是他们不想革自己的命。 五、对我们普通人意味着什么？陶定律不是替代摩尔定律，是互补。 西方继续啃先进制程的硬骨头，我们靠架构挖潜，另起炉灶。 未来不用最先进制程，也能做出高性能芯片，手机、国产设备性能持续提升，不再被卡脖子。 最后。 技术从来不是谁的专属，谁被逼到墙角，谁敢打破惯性，谁就能走出新路。 华为的陶定律就是最好的例子。
英文翻译
Huawei's Tao Law is not about taking a curve overtaking; it's a forced new path in semiconductors. Recently, Huawei proposed the Tao Law, named after the Greek letter Tau, and it went viral online. Many see it as black technology, but frankly, it's not a unique stroke of genius from Huawei. It's the inevitable result of the West having a shortcut they refuse to take, while Huawei had no road left and had to break through. Let me explain this in plain language so everyone can understand. 1. First, understand that Moore's Law, which ruled chips for 60 years, has completely stalled. Our phones and computers keep getting faster. The core relies on Moore's Law: making transistors smaller and smaller through geometric scaling, doubling their number every 18–24 months, and doubling performance. In the past, the more advanced the process, the cheaper and faster the chips were—easy money. Now in 2026, we're stuck at 3nm and 2nm. Physical limits have been reached, transistors are down to atomic scale. Electrons start to wander, quantum tunneling occurs, leakage current, heat dissipation issues blow up, and worse—costs explode! Designing a top-tier chip costs over $1 billion, and building a 3nm fab costs $20 billion. Smaller means more expensive. The cost-performance ratio has completely flipped. In short, the geometric scaling path of Moore's Law has hit a dead end. 2. Huawei's Tao Law: instead of fighting for smaller size, it fights for faster time. On May 25, 2026, Huawei formally proposed Tao Law. The core idea is one sentence: replace geometric scaling with time scaling. Tau is the time constant in circuits—the delay of signal transmission in a chip. The goal: not to make transistors smaller, but to make signals travel shorter distances with lower latency, thereby improving performance. A simple analogy to understand: Moore's Law is like making the furniture (transistors) in a house smaller so you can cram more furniture in—efficiency through quantity. Tao Law is like turning a flat house into a 3D building. Logic folding: furniture that was 10 meters apart is now stacked vertically. Signals don't need to take detours; they travel just one meter. Efficiency through speed. The core approach: a four-layer full-stack optimization, not just 3D: - Device layer: optimize materials to reduce signal resistance. - Circuit layer: the core. Logic folding + 3D stacking shortens critical paths by 60%, greatly reducing latency. - Chip layer: hardware-software coordination for more efficient instruction scheduling. - System layer: new bus protocols for faster inter-chip communication. 3. The key truth: this technology is doable by anyone, but only Huawei had to do it. Many think Tao Law is Huawei's unique creation, but in fact, the principle is understood across the industry. The West had the technology long ago—like 3D stacking and advanced packaging—but Western giants didn't adopt it. They didn't want to, and couldn't. A. The West had a shortcut and kept walking it—too lazy to take a detour. For decades, relying on advanced processes (7nm, 5nm, 3nm) was easy money. If your daughter already has flowers to wear, why bother finding a red headband? Inertia and laziness pervade the entire supply chain: EDA, equipment, software—all built around planar scaling. Going 3D means overturning the entire ecosystem. High investment, high risk—capital markets won't allow it. B. Companies like Intel have too much historical baggage to change. X86 chips have been backward-compatible for 40 years. Old programs must run. A major architectural change would break compatibility, affecting Windows, industrial software—astronomical commercial costs. Not that they can't change, but they can't afford to change. They don't dare. C. Huawei was forced into a corner—no road left, so they had to break through. Huawei was cut off from advanced processes, lacking EUV lithography machines. It's like the daughter has no flowers to wear. Without Tao Law—time scaling and 3D folding—they'd have no chips, no phones to sell, no business to survive. But they also have no historical baggage. No need to maintain old architectures; they dare to start from scratch. They turned what others view as optional into a must-do, top-priority core strategy. 4. One sentence summary of the two paths and two fates. Moore's Law: the West competes on process, size, and EUV. They have a shortcut, but also baggage—addicted to easy living. Tao Law: Huawei competes on architecture, time, and potential. No shortcut, no baggage—forced into existence. Essentially, it's not that Huawei is smarter; it's that desperation forced a new track. It's not that the West can't do it; they just don't want to revolutionize themselves. 5. What does this mean for ordinary people? Tao Law doesn't replace Moore's Law; it complements it. The West continues to tackle the hard nuts of advanced processes. We rely on architectural optimization to forge a new path. In the future, we can make high-performance chips without the most advanced processes. Our phones and domestic equipment will keep improving, and we'll no longer be choked. Finally. Technology has never been anyone's exclusive property. Whoever is backed into a corner, whoever dares to break inertia, can carve a new path. Huawei's Tao Law is the best example.
back to top