· 供应链分析

OpenAI优势减弱,AI行业多极化,基本面未受损公司现买入机会

涉及标的:

中文翻译

OpenAI 此前是明确的领导者。现在,Gemini 3.0 在图像生成方面超越了它,Claude Opus 4.5 在编码方面超越了它。以此类推。看起来在获得先发优势后,ChatGPT 似乎无法被颠覆。但现在从 Similarweb 数据我们可以看到,Gemini 的受欢迎程度相对于 ChatGPT 正在增长。而且我相信由于使用内部张量处理单元(TPU)进行推理,它们的成本效益也更高。OpenAI 试图通过要求 Mag7 和政府为其合同承诺提供担保,使整个行业依赖它,以便大家同船沉没,但这最终失败了。整个 AI 行业不仅仅是 OpenAI,随着我们看到更多的代理(Agentic)和机器人应用,我们很可能会看到 Anthropic、Google、XAI 和其他供应商的广泛大型语言模型(LLM)和日益增长的算力使用。但目前我们看到这种传染效应从 OpenAI 蔓延到该领域的几家公司 $ORCL、$CRWV,可能还有 $AMD 以及其他几跳远的公司,它们正在为承诺的资本支出构建产能。但这正在拉垮那些与 OpenAI 隔离(就股价而言)的其他公司,因此如果基本面没有受到实质性影响,这就是一个买入机会。

英文原文

OpenAI was the clear leader before. Now, Gemini 3.0 surpasses it in image generation. Claude Opus 4.5 surpasses it in coding. Can go down the line. It looked like ChatGPT couldn’t be disrupted after they got first movers advantage. Now from similarweb data we can see popularity with Gemini growing vs ChatGPT. And im sure they’re a lot more cost effective as well using in house TPUs for inference. OpenAI tried making the whole industry dependent on it, by asking Mag7 + government to backstop their contract promises so they all go down the ship together but that ended up flopping. AI industry as a whole is not OpenAI, we’ll likely see widespread LLM + growing compute usage from Anthropic, Google, XAi and other vendors as we see more agentic + robotics applications next. But right now we’re seeing contagion spread from OpenAI to several companies in the sector $ORCL, $CRWV, maybe $AMD and others few hops away that were building out capacity for their promise capex spend. But it’s taking down other companies isolated from OpenAI (in terms of stock price), which is why it’s a buying opportunity if fundamentals aren’t too materially impacted.

在 X 上查看原推 ↗