AI bubble concerns rise! Nvidia's earnings beat expectations but its stock plunged 5.6%, marking the biggest intraday drop in three months.

Nvidia, the leader in the artificial intelligence (AI) chip market, suffered its worst drop in three months on Thursday after releasing impressive performance guidance. Analysts say the company's outlook failed to dispel market concerns about an AI bubble.

On Thursday, Nvidia's stock fell as much as 5.6% in New York to $184.58, marking the biggest intraday decline since November 25. The drop came after the company released its first-quarter sales forecast, in which Nvidia not only easily beat analysts’ average expectations, but also reported a 73% year-on-year increase in fourth-quarter revenue.

However, analysts believe the market reaction highlights investors' skepticism towards Nvidia. Previously, explosive sales growth propelled Nvidia to become the world’s most valuable company; now investors want stronger assurances that the AI spending boom can continue.

Analysts at Hargreaves Lansdown stated in a post-earnings report, shareholders are still worried “whether this current wave of AI spending can continue to support growth in the coming years and whether Nvidia can maintain its leading position as AI shifts from training models to routine task execution.”

CEO Jensen Huang refuted these concerns during a conference call on Wednesday. He said that customers have already started making money from newly acquired computing power and will therefore continue to invest at high levels.

Huang said:

“You need computing power, which will directly translate to growth and revenue. I am confident their cash flow is growing.

I am very confident in their cash flow growth—and for a very simple reason. We are now seeing a turning point for agentic AI and the practicality of intelligent agents. The industry has reached another turning point.

Investor Michael Burry, famous for “The Big Short,” further exacerbated market concerns on Thursday. He pointed out that Nvidia's current purchase commitments total $95.2 billion, compared to only $16.1 billion a year ago. If demand fluctuates, this could pose risks.

Nvidia CFO Colette Kress attempted to ease other concerns raised by analysts, including supply constraints. She said the company has secured enough components to meet growing demand.

She told analysts that producing Nvidia's most advanced chips remains challenging. But the current Blackwell product line and the upcoming Rubin successor will outperform previous sales expectations. Nvidia previously stated that these chips would generate $500 billion in revenue by the end of 2026.

She said:

“We believe we have built up inventory and supply commitments to meet future demand, including shipments extending into 2027.”

Memory Chip Shortage Raises ‘Circular Transaction’ Concerns

Nvidia expects first-quarter revenue of about $78 billion, higher than the analysts' average forecast of $72.8 billion, though some analysts previously projected nearly $80 billion. Additionally, Nvidia’s fourth-quarter revenue, earnings per share, and adjusted gross margin all exceeded expectations.

Nvidia’s data center business brought in $62.3 billion this quarter, higher than the analysts' average forecast of $60.4 billion. Other business segments performed weaker. Gaming, once Nvidia's main source of income, posted revenue of $3.73 billion, below the analysts' forecast of $4.01 billion. Automotive-related sales were $604 million, while Wall Street expected $643 million.

Analysts believe the tech industry currently faces a hidden concern: memory chip shortages. Like most companies in the electronics industry, Nvidia products depend on stable memory chip supply. These chips provide short-term storage for devices ranging from smartphones to supercomputers. Tight supply has driven up memory chip prices and made it harder to ship more devices this year.

This tension has weighed on the gaming segment. Kress said she was unsure whether the problem would ease enough this year for the segment to resume growth.

Regardless, AI datacenter chips have become a bigger focus. Earlier this month, Nvidia announced that Meta will deploy “millions” of Nvidia processors in the coming years, further deepening an already close partnership between the two companies.

Nvidia’s main competitor AMD also announced a similar long-term deal with Meta this week. The chipmaker said the transaction would reach a scale of tens of billions of dollars.

Chipmakers lock in future computing power demand with a series of large long-term agreements to demonstrate the continued strength of the AI economy. But such close transactional relationships—sometimes with suppliers and customers holding stakes in each other—invite criticism. Some argue that these “circular transactions” may artificially inflate demand.

Risk Warning and DisclaimerMarkets are risky, and investments should be cautious. This article does not constitute personal investment advice and does not take into account special investment goals, financial situation or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular situation. Investment is at your own risk.