On November 12, Nvidia announced that Jensen Huang resigned as president, CEO, and director, effective immediately. On that day, after the opening of the US stock market, Nvidia's stock fell more than 14% at one point, hitting a low of $148.8, nearly 40% off its peak in October, and at one point, the market value evaporated by more than $600 billion.

Image source: Vision China

The founder's departure is the most direct trigger for this round of sell-off—the market is not only responding to a single risk but is also re-pricing the entire company. Meanwhile, when a leader in AI computing power sees its valuation drop to historical lows while demand remains strong, discussions around whether to buy the dip also began to heat up. If we shift our attention away from this personnel change, we might find that the real questions Nvidia needs to answer are about valuation, its economic moat, and whether demand has undergone irreversible changes.

The sell-off triggered by Jensen Huang's resignation pushed Nvidia's valuation to a rare position: approximately 12 times forward PE, the lowest level for the company's stock price in fifteen years. During the same period, the expected price-to-earnings ratio of the S&P 500 index was about 18 times. In other words, after a governance crisis, Nvidia's valuation has become cheaper than the overall US stock market.

This valuation level corresponds to a still strong business. For fiscal year 2026 (ending January 2026), Nvidia's revenue is projected to be $215.9 billion, a year-on-year increase of 65%, with data center revenue reaching $62.3 billion in a single quarter, maintaining a gross margin of around 75%. In the second quarter of fiscal year 2027, revenue was $96.2 billion, with data center revenue of $89 billion, a year-on-year increase of 117%, an operating profit margin of 66.2%, and free cash flow of $21.3 billion.

Although multiple investment banks lowered their target price for Nvidia after the Jensen Huang incident, most still maintained a buy or overweight rating.

Barclays lowered its target price from $275 to $210, maintaining an overweight rating, clearly stating, "We are reducing the valuation multiple by about 25% to reflect reputational and transition risks over the next few months," with the current stock price implying a potential upside of about 20.9%. Goldman Sachs maintained a buy rating, lowering its target price from $300 to $220; Citigroup lowered from $285 to $225, also maintaining a buy rating; Royal Bank of Canada reduced its target from $330 to $230. Truist is one of the few institutions that directly downgraded its rating from "buy" to "hold," cutting its target price to $174, which is just around the current stock price.

Clearly, the consensus among institutions is: the business is not bad, but the multiple given to the business needs to be reset. And the 12 times forward PE has largely completed this reset. The main support behind this is that the long-term narrative on the demand side has yet to be falsified.

During the GTC conference in 2026, Jensen Huang predicted that by 2027, the demand for AI chips from the Blackwell and Rubin systems would reach at least $1 trillion, doubling the prior forecast of $500 billion for 2026. Anthropic's latest disclosed annual recurring revenue (ARR) is also still growing, and the total forecast for AI capital expenditures has not been downgraded due to short-term fluctuations, only the slope of expenditures is being shifted backward.

However, cheap valuations do not mean that risks have disappeared. The risks Nvidia currently faces may be more fundamental than "who will be the CEO."

Firstly, the lock-in effect of CUDA is being eroded from the inference side.

The reason CUDA acts as an economic moat is that the cost of migrating models from Nvidia GPUs to other chip architectures is relatively high. However, Anthropic has disclosed a landmark progress during its IPO roadshow: its model Claude Mythos Preview completed the migration of the core GPU kernel autonomously within three days, with little human intervention, allowing Claude to run key inference processes on AMD’s flagship GPU.

Comments by Silicon Valley investor Marcelo P. Lima have been widely cited: "Nvidia's CUDA moat no longer exists." This comment may be somewhat exaggerated, but the direction of the trend is clear: as large models begin to have the capability to rewrite system software, the migration costs previously constructed at the expense of human engineering labor are rapidly being flattened by automation—CUDA will not collapse overnight, but the shift from "everyone must use it" to "it is still needed for training, but can be bypassed for inference" is already happening.

Secondly, Nvidia faces supply bottlenecks not in wafers, but in HBM and packaging.

Industry data shows that TSMC can slice more than 400 million GPU chips in a year, but Nvidia’s actual annual delivery is only about 10 million. The real constraint lies in HBM: global annual HBM capacity is about 3.75EB, with Nvidia allocated approximately 70%, with each GPU averaging 260GB, which can only support about 10 million GPUs in total. Meanwhile, packaging capacity is also tight, and adjustments in the supply of HBM4 have already led to potential delays in Rubin's mass production pace. This means that even with unlimited demand, Nvidia's shipping volume in the medium term is constrained by a physical limit, rather than its design capability or customer orders.

Moreover, most importantly, what Jensen Huang left to Nvidia is not just organizational assets like CUDA and architectural accumulation, but three highly personalized elements: strategic judgment on the technological roadmap, negotiation skills with TSMC and SK Hynix in the supply chain, and direct communication channels with clients and the government. Interim CEO Colette Kress is a familiar face in the capital markets, having long served as CFO, but questions about who will decide Nvidia's product roadmap and maintain supply chain relationships have no answers until an official successor is in place. This means that while Nvidia's stock price is unlikely to collapse after "the departure of Jensen Huang," the premium that the market was willing to pay for Jensen Huang will not automatically transfer to the successor.

Therefore, when looking at opportunities and risks together, Nvidia's current position can perhaps be summarized in one sentence: the business is still there, but the "faith premium" attributed to the business is being systematically stripped away.

Indeed, the 12 times forward PE has absorbed a considerable degree of pessimistic expectations. If the long-term direction of AI capital expenditures remains unchanged, the demand for Blackwell and Rubin materializes, and the successor can stabilize the supply chain and customer relations, then the current price provides a decent entry window. Barclays, Goldman Sachs, and Citigroup maintain buy ratings and expect over 20% potential upside, reflecting this judgment.

However, "cheapness" itself still fails to constitute sufficient reasons for buying. The loosening of the CUDA lock-in effect is a structural change occurring slowly over the years but difficult to reverse. The physical bottlenecks on the supply side in HBM and packaging limit Nvidia's revenue ceiling when demand explodes. The uncertainty of the successor means that the shedding of the "founder's premium" may not yet be over.

Bernstein bluntly stated in its latest research report: "If you see two ants in the kitchen, there are definitely more than two ants in the kitchen." It pointed out the potential chain of developments: from personal scandals to governance vacuums and business news, to government investigations; each link could trigger a new round of sell-offs. Each outbreak in the chain pulls Nvidia's valuation closer from "Jensen Huang premium" to "ordinary semiconductor company premium."

Without a doubt, the market has always feared not bad news but rather the uncertainty of when and in what form bad news will appear. Until everything settles, "how much is Nvidia worth without Jensen Huang" remains an open question.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到[email protected]，本平台相关工作人员将会进行核查。