Nvidia Shares Tumble Amid AI Bubble Concerns, Despite Strong Earnings
Nvidia Corporation, the undisputed titan of artificial intelligence (AI) chip manufacturing, has experienced its most significant stock market slump in ten months. The sharp decline, which saw shares drop 5.5% to US$184.89 in New York, was triggered by the company’s latest financial forecast, which, despite appearing robust on the surface, failed to entirely alleviate investor anxieties about a potential AI spending bubble. This marks the largest single-day drop for the company since April 16.
The market’s reaction underscores the growing skepticism surrounding Nvidia’s meteoric rise. Following a period of explosive revenue growth that propelled the chipmaker to become the world’s most valuable company, investors are now demanding more concrete assurances that the current surge in AI-related expenditure is sustainable. Lingering questions persist regarding the long-term viability of this AI spending wave beyond the next few years and whether Nvidia can maintain its dominant market position as the AI landscape evolves from model training to everyday operational tasks.
During a recent conference call, Chief Executive Officer Jensen Huang sought to address these concerns directly. He asserted that Nvidia’s clients are already realising tangible financial benefits from their newly acquired computing power, a factor he believes will drive continued investment at elevated levels. “You need compute capacity, and that translates directly to growth, and that translates directly to revenues,” Huang explained. “I’m confident their cash flows are growing.”
Adding to the chorus of caution, investor Michael Burry, famously known for his prescient insights in “The Big Short,” highlighted Nvidia’s substantial purchase obligations, which have ballooned to US$95.2 billion from US$16.1 billion a year prior. He suggested this could pose a significant risk should demand falter.
Nvidia’s Chief Financial Officer, Colette Kress, also worked to assuage other analyst concerns, particularly those surrounding potential supply chain constraints. She assured stakeholders that the company has secured sufficient components to meet anticipated demand. While acknowledging the inherent challenges in producing Nvidia’s most advanced chips, Kress indicated that the current Blackwell lineup and its upcoming successor, codenamed Rubin, are projected to surpass earlier sales expectations. Nvidia had previously forecast that these chips would generate US$500 billion by the end of 2026. “We believe we have inventory and supply commitments in place to address future demand, including shipments extending into calendar 2027,” Kress stated.
Navigating the Complexities of the Chinese Market and Global Supply Chains
Despite these assurances, uncertainties remain, particularly concerning the critical Chinese market, which represents a significant segment for chip sales. While the US government has granted limited licenses for the shipment of H200 processors to China, Nvidia has yet to secure approval from the Chinese government for these sales. Consequently, the company will continue to exclude data centre revenue from China in its official forecasts for the time being. The specific license requires US inspection of the chips before they can be dispatched to customers, and these processors are subject to a substantial 25% tariff upon re-entry into the United States.
Nvidia’s core business revolves around the production of accelerator chips – processors specifically designed to handle the immense data volumes required for developing AI models. These semiconductors are also crucial for the inference stage, where AI models execute tasks in response to real-world inputs. Beyond its leadership in AI accelerators, the Santa Clara, California-based company has strategically expanded its offerings to include general-purpose processors, networking solutions, and complete computer systems, thereby deepening its integration with its customer base.
For the fiscal first quarter, Nvidia projected revenue to reach approximately US$78 billion, exceeding the consensus analyst estimate of US$72.8 billion, with some forecasts even approaching US$80 billion. In the preceding fiscal fourth quarter, which concluded on January 25, the company reported a robust 73% surge in revenue, reaching US$68.1 billion. Earnings per share, excluding certain items, stood at US$1.62, surpassing analyst predictions of US$65.9 billion in sales and US$1.53 per share in earnings. The adjusted gross margin, a key indicator of profitability, also edged past expectations at 75.2%.
“We aren’t sure what else investors want to hear at this point. But we like what we heard,” commented Bernstein analyst Stacy Rasgon on the company’s results.
Nvidia’s data centre division, the powerhouse behind its AI accelerators and networking products, was a significant contributor, generating US$62.3 billion in revenue for the quarter, outperforming the average analyst estimate of US$60.4 billion.
However, not all segments performed as strongly. The gaming division, once the primary revenue driver for Nvidia, generated US$3.73 billion in sales, falling short of the average estimate of US$4.01 billion. Similarly, automotive-related sales amounted to US$604 million, with Wall Street anticipating US$643 million.
A prevailing challenge across the broader tech industry, and one that impacts Nvidia, is the ongoing shortage of memory chips. These components are vital for short-term data storage in a wide array of electronic devices, from smartphones to high-performance computing systems. The scarcity has driven up memory prices and hampered the production volumes of many electronic goods. This memory crunch has notably affected Nvidia’s gaming division, with Kress expressing uncertainty about whether the situation will ease sufficiently this year to enable growth in that segment.
Nonetheless, the focus remains firmly on data centre chips for AI applications. Earlier this month, Nvidia announced a landmark agreement with Meta Platforms Inc. to deploy “millions” of Nvidia processors over the coming years, solidifying an already strong partnership between two leading AI players. This development comes shortly after Nvidia’s main competitor, Advanced Micro Devices Inc. (AMD), revealed a similar long-term deal with Meta, valued at tens of billions of dollars.
These substantial, long-term commitments are being presented by chip manufacturers as compelling evidence of the AI economy’s robust health. However, the close-knit nature of these transactions, which sometimes involve reciprocal financial investments between suppliers and customers, has also drawn scrutiny regarding the potential for circular deals to artificially inflate demand.
