NVIDIA's stock price has recently exhibited a rare divergence from its fundamentals: on one hand, market expectations for its short-term performance are extremely high, yet on the other hand, the stock price has persistently underperformed the broader market. This weakness does not stem from an earnings miss, but rather from a "wall of worry" confronting investors. The market is no longer simply buying into the story of supply failing to meet demand; instead, it has begun to fret over customers' financing capabilities, the rise of competitors, and the sustainability of AI investment returns.
According to Wind Information, Morgan Stanley analyst Joseph Moore pointed out in a recent research report that while these concerns are weighing on the stock price, they are actually exaggerated. The institution believes that NVIDIA's upcoming next-generation chip, "Vera Rubin," will be a key catalyst to break the deadlock, not only alleviating market anxiety about market share loss but also solidifying the company's dominant position in the next phase of computing architecture. Morgan Stanley maintained its "Overweight" rating on NVIDIA and set a price target of $250.
The current market sentiment is markedly different from the pullbacks seen during the AI wave over the past three years. Previous sell-offs (such as those in late 2024 or 2025) were typically accompanied by specific negative catalysts, for instance, debates over technological roadmaps triggered by DeepSeek, or rumors of inventory buildup in the supply chain. Now, however, despite extremely strong demand feedback from the supply chain—with some investors even discussing the possibility of earnings per share exceeding $9 this year, far above the market consensus of $7.75—the stock price remains lukewarm.
The core of this disconnect lies in investors shifting their focus from "who can produce the chips" to "who can pay for these expensive chips." As the list of AI beneficiaries expands, other segments of the semiconductor ecosystem (such as memory and equipment) are diverting capital, while the market has developed deep-seated doubts about the sustainability of capital expenditures by leading model developers.
The market's concerns about the financing capabilities of NVIDIA's major customers are becoming more concrete. Recently, stocks closely tied to OpenAI (such as Microsoft and Oracle) have significantly underperformed, with the core logic being market skepticism about whether OpenAI's cash burn rate can match its infrastructure construction pace.
Taking Oracle as an example, due to the need to shoulder incremental interest expenses and lease liabilities to support massive IT capital expenditures (largely related to OpenAI), some analysts have significantly lowered their future EPS expectations for the company. If investors doubt Oracle's ability to execute its order backlog, or suspect that OpenAI cannot afford to pay, this sentiment is bound to impact NVIDIA's valuation. More sensitive is the role NVIDIA itself is playing. NVIDIA is now "giving back" to its customers and partners, including OpenAI, CoreWeave, and even Intel, through equity investments. This approach has raised market questions about "circular financing."
In this regard, Jensen Huang's remarks at CES were quite candid; he admitted that if he could turn back time, he would have made a different investment choice regarding Anthropic in the early days—a missed opportunity due to insufficient funds back then, which led Anthropic to initially favor Amazon and Google's ASIC solutions.
The current NVIDIA clearly wants to avoid repeating past mistakes. Morgan Stanley, after verification, indicated that NVIDIA's current investments are mostly equity-based rather than debt. Taking this week's investment in CoreWeave as an example, it was purely an equity injection, involving no credit arrangements. Although the company may retain some credit risk exposure through means like leasing in the future, this is more about maintaining dominance within the ecosystem rather than engaging in "financial engineering" to artificially inflate revenue figures.
Regarding competition, the market widely worries that ASICs and AMD will erode NVIDIA's market share. Indeed, when the entire industry is growing at over 75%, every competitor might feel like they are "winning." Total Cost of Ownership studies often suggest alternative solutions are cheaper, but such calculations based on paper specifications often prove inaccurate in real-world, large-scale deployments.
Morgan Stanley expects NVIDIA to maintain roughly 85% revenue share this year, with its quarterly revenue increments even exceeding the total revenue of its closest competitor. The real game-changer lies in the next-generation "Vera Rubin" architecture. Supply chain checks indicate that the transition from Blackwell to Rubin might be faster than anticipated.
A critical, often overlooked detail is the leap in manufacturing efficiency: information from CES revealed that board-level assembly time for Vera Rubin systems is only about 5 minutes, compared to approximately 2 hours for Blackwell. This massive advantage in manufacturing processes, combined with full-rack-level connection speeds and silicon-level optimizations, means Rubin's production ramp-up in 2026 will be far smoother than that of its competitors. For developers urgently needing large-scale clusters, this is not just a performance issue but a matter of delivery certainty.
By 2026, the AI semiconductor ecosystem will indeed have changed. As capacity bottlenecks resolve, AI红利 begins to spill over, attracting some capital seeking higher leverage to segments like memory and wafer foundries, which possess greater operating leverage. Morgan Stanley's previous shift of its top pick from NVIDIA to companies like Micron was based on this very logic.
But this does not mean NVIDIA has lost its appeal. NVIDIA's current P/E ratio is already lower than most growth or cyclical stocks within its coverage universe. The market has become desensitized to the recent strong earnings expectations. Only when more details about Vera Rubin and the product pipeline become clearer at the mid-March GTC conference, and particularly as the capital market mechanisms surrounding AI spending become more robust, will the valuation shackles suppressing the stock price be broken. Until then, those seemingly aggressive customer spending plans actually form the cornerstone of NVIDIA's bull case.
Comments