Rumors of BYD's Self-Developed Humanoid Robot Codenamed 'Yao Shun Yu', Plan to Use 20,000 Units Internally by Year-End Denied by BYD 网传比亚迪自研人形机器人代号“尧舜禹”、年内自用2万台 比亚迪否认

As soon as rumors about BYD's "Yao Shun Yu" robots surfaced, the company swiftly debunked them, delivering a sharp and decisive reality check. Nowadays, automakers feel almost obligated to dabble in humanoid robots to signal their "futuristic" vision, yet scaling up invites criticism for straying from core business. BYD’s denial reveals a clear-headed pragmatism: rather than chasing concepts, it prefers to maximize the utility of robotic arms on its welding lines. The grand blueprint of "deploying 20,000 units internally within the year" is likely just an "industry imagination" spun by certain analysts—a narrative the capital markets love to hear, but not the real concern for manufacturers at this stage. Sometimes, the absence of news is the best news, indicating that the company isn’t spinning its wheels on illusory tracks.

This pragmatism stands in ironic contrast to the jittery memory chip stocks on the other side of the Pacific, spooked by a single research note. A line from SemiAnalysis suggesting "memory might be downgraded" can send the market caps of storage companies, both domestic and overseas, into a collective sneeze. Expectations for AI have reached a point where the market is on high alert—any slight rustle can be interpreted as a precursor to an industry earthquake. But a closer look at the report reveals that the so-called "downgrade" is limited to system memory on the CPU side, with no impact on GPU compute cores. This is essentially a routine engineering decision based on supply chain, power consumption, and cost considerations—yet it became the final straw that broke the market's confidence. This precisely exposes the core ailment of the current AI hardware market: valuations are overly reliant on a perfect narrative of "expectations," leaving no room for any real-world technical trade-offs.

Thus, we witness a split landscape. On one side, giants like BYD, at the forefront of manufacturing, steer clear of rumors about "building robots," emphasizing the difficulties and costs of implementation. On the other, financial markets pour out panic over a rumor that a single screw might change material. Together, they depict the dual faces of today's AI fervor: the "cold reflection" of the real industry, which proceeds cautiously and manages risks; and the "hot imagination" of capital markets, which spins narratives and amplifies volatility.

Between this "cold" and "hot," there lies a dangerous temperature gap. When the market treats every minor tweak in hardware parameters as a sign of wavering faith, it has effectively detached from the technology itself and entered the realm of collective psychological games. NVIDIA’s strategic choices—perhaps driven by better energy efficiency or cost control—are directly equated with "memory shrinkage" and "innovation fatigue." This simplistic linear logic is an affront to complex engineering. The fragility of memory stocks also underscores how unstable the valuation shift from "cyclical stock to AI core asset" truly is—it rests on an extreme bet on a single giant's product roadmap.

Isn’t the real AI infrastructure revolution built brick by brick through these trivial decisions of "downgrading" and "optimizing," "denial" and "pragmatism"? It requires companies like BYD to first produce world-leading vehicles before even discussing the possibility of humanoid robots. It also needs players like NVIDIA, which, while pushing compute power to its limits, never loses sight of balancing system power consumption and cost. The market needs patience to understand these "unsexy" decisions, rather than merely paying for stunning presentations and astronomical promises.

Rumors and denials, research reports and crashes—this noise will eventually fade. The questions that remain are the crucial ones: Are we investing in a technological revolution, or in a myth about technology that is increasingly detached from reality? When the market cap of "AI concepts" begins to tie only to sentiment-driven news, the foundation of this feast may require more vigilance than a "downgraded" piece of memory.

比亚迪的“尧舜禹”机器人传闻刚起，官方就光速辟谣，一盆冷水泼得干净利落。这年头，车企不造点人形机器人都不好意思说自己有“未来感”，但真正把摊子铺开了，又怕被戴上“不务正业”的帽子。比亚迪的否认里，透着一股子清醒的务实。比起造概念，它更乐意把焊装线上的机械臂用到极致。所谓“年内自用2万台”的宏伟蓝图，大概率只是某些分析师笔下的“产业想象”，是资本市场爱听的故事，却不是制造业此刻关心的真命题。没有消息，有时候才是最好的消息，意味着企业没在虚无缥缈的赛道上空转。

这种务实，与大洋彼岸因一纸研报而风声鹤唳的存储股，形成了绝妙的讽刺对比。SemiAnalysis一句“内存可能降配”，就能让海内外存储公司市值集体打喷嚏。市场对AI的期待已经高到了“风声鹤唳”的程度，任何一点风吹草动，都能被解读为产业地震的前兆。但细看报告，所谓“降配”仅限于CPU侧的系统内存，与GPU算力核心并无关系。这本是一个基于供应链、功耗与成本的常规工程抉择，却成了压垮情绪的最后一根稻草。这恰恰暴露了当前AI硬件市场的核心病症：估值太依赖于“预期”的完美叙事，以至于容不得任何一点现实的技术权衡。

于是，我们看到了一个分裂的图景。一边是比亚迪这样身处制造业一线的巨头，对“造机器人”的传闻避之不及，强调落地的艰辛与成本；另一边是金融市场，为了一颗螺丝钉可能换材质的传闻而倾泻恐慌。两者共同描绘了当下AI热潮的双面：一面是实体产业谨慎求索、控制风险的“冷思考”；一面是资本市场狂热叙事、放大波动的“热想象”。

这“冷”与“热”之间，存在着危险的温差。当市场把每一个硬件参数的微调都视为信仰动摇的信号时，它实际上已经脱离了技术本身，进入了集体心理博弈的领域。英伟达的战略选择（或许是出于更优的能效比或成本控制），被直接等同于“内存缩水”和“创新乏力”，这种简单的线性逻辑，是对复杂工程学的傲慢。而存储股的脆弱，也反衬出所谓“从周期股到AI核心资产”的估值切换，根基何其不稳——它建立在一条对单一巨头产品路线图的极致押注之上。

真正的AI基础设施革命，难道不正是在这些“降配”与“优化”、“否认”与“务实”的琐碎抉择中，一步步垒砌起来的吗？它需要像比亚迪那样，先造出全球第一的车，再谈人形机器人的可能；也需要像英伟达那样，在算力狂飙的同时，不忘系统功耗与成本的平衡。市场需要耐心去理解这些“不性感”的决策，而非仅仅为震撼的PPT和天文数字的承诺买单。

传闻与辟谣，研报与暴跌，这些噪音终将过去。留下的问题才是关键：我们到底是在投资一场技术革命，还是在投资一个关于技术的、越来越脱离地面的神话？当“AI概念”的市值开始只与情绪新闻挂钩时，这场盛宴的根基，恐怕比一块“缩水”的内存更需要警惕。