Jensen Huang 在 Nvidia 的季度财报电话会议上，再次有力地阐述了 AI 的未来，将自己塑造成公司在那个未来中角色的教育者与传道者。投资人整体上仍然看涨，但市场已经习惯 Huang 一再发表关于 AI 的讲解，尤其是在资金涌入晶片制造商、而 Nvidia 自家股价也大幅上涨的一年之后。尽管如此，该公司再次超出预期，且给出的营收展望高于华尔街预估，进一步巩固了 Nvidia 作为 AI 基础设施支出最明显受益者的地位。

直接结果显示出强劲与广度兼具。Nvidia 表示，下一季营收将比分析师预期高出约 $4 billion，这对几乎任何其他公司而言都会是惊人的数字，但对 Nvidia 的季度财报来说已成常态。消息传出后，亚洲的科技与供应链类股上涨，而 LG Electronics 及相关公司在 Huang 谈到机器人、类人机器人、自驾车以及所谓的 physical AI 时，股价飙升超过 20%。Nvidia 也将股息从 1 cent 提高到每股 25 cents，并授权再回购 $80 billion 股票，尽管其市值仍高于 $5 trillion。

Huang 的主要论点是，Nvidia 并不会过度依赖 Microsoft、Meta、Amazon 和 Google 这类 hyperscalers，因为他所称的 Category 2 这第二类客户群，随著时间推移将带来更大的需求。预计这些 hyperscalers 今年将在 AI 晶片与设备上花费 $725 billion，Nvidia 仍预期从中拿下很大一部分支出，但 Huang 表示，较小的工业公司与 AI 云端供应商在采用这项技术时，成长速度会更快。最新一季中，hyperscaler 营收年增 115% 至超过 $37 billion，而 Category 2 销售也年增 74% 至超过 $37 billion；若按季度比较，CFO Colette Kress 表示 Category 2 成长更快，支持了 Huang 关于 Nvidia 客户基础正从最大买家之外持续扩大的说法。

Jensen Huang used Nvidia’s quarterly earnings call to deliver another forceful case for the future of AI, presenting himself as an educator and evangelist for the company’s role in that future. Investors remain broadly bullish, but the market has become accustomed to Huang’s repeated AI lectures, especially after a year in which money has rushed into chipmakers and Nvidia’s own stock has risen sharply. Even so, the company again beat expectations and guided revenue above Wall Street estimates, reinforcing Nvidia’s status as the clearest beneficiary of AI infrastructure spending.

The immediate results showed both strength and breadth. Nvidia said the next quarter’s revenue would be about $4 billion above what analysts expected, a figure that would be extraordinary for almost any other company but has become routine for Nvidia’s quarterly reports. In Asia, tech and supply-chain shares rallied on the news, while LG Electronics and related companies jumped more than 20% as Huang talked about robotics, humanoids, autonomous vehicles, and so-called physical AI. Nvidia also raised its dividend from 1 cent to 25 cents a share and authorized another $80 billion for buybacks, even as its market value remained above $5 trillion.

Huang’s main argument was that Nvidia is not overly dependent on hyperscalers such as Microsoft, Meta, Amazon, and Google because a second customer group, which he called Category 2, will drive even larger demand over time. The hyperscalers are expected to spend $725 billion this year on AI chips and equipment, and Nvidia still expects to capture a large share of that spending, but Huang said smaller industrial firms and AI cloud providers will expand even faster as they adopt the technology. In the latest quarter, hyperscaler revenue rose 115% year over year to more than $37 billion, while Category 2 sales rose 74% to more than $37 billion as well; on a quarter-over-quarter basis, CFO Colette Kress said Category 2 grew faster, supporting Huang’s claim that Nvidia’s customer base is widening beyond the biggest buyers.