NVIDIA has soared to new financial heights, boasting average daily profits of 2.25 billion yuan. Its latest financial report has sent shockwaves through the market.

The fourth-quarter results for fiscal year 2026, which concluded on January 25, 2026, reveal that NVIDIA's core operational metrics have not only met but significantly surpassed expectations. The company's guidance for the upcoming quarter is exceptionally robust, underscoring the persistent and intense market demand for artificial intelligence (AI) computing chips.

The financial report highlights that NVIDIA achieved a staggering revenue of $68.1 billion in the fourth fiscal quarter, marking a 73% year-over-year increase. This figure outperformed both the company's own projection of $65 billion and the market consensus of $65.68 billion. From a full-year perspective, NVIDIA's total revenue for fiscal year 2026 reached a record $215.9 billion, up 65% from the previous year.

The data center segment, which accounts for over 90% of NVIDIA's revenue, continues to be the primary growth driver. Data center revenue surged to $62.3 billion in the fourth quarter, reflecting a 75% year-over-year jump.

Leveraging the product mix optimization resulting from increased production of Blackwell architecture chips, NVIDIA's gross margin climbed to 75.2% in the fourth fiscal quarter, hitting a new 18-month high.

Profitability soared even higher. The net profit for the fourth fiscal quarter stood at $42.96 billion, with the full-year net profit reaching $120.067 billion. Based on the annual data, this translates to daily earnings of approximately 2.25 billion yuan.

Looking ahead, NVIDIA remains bullish about its financial prospects, projecting that revenue for the first quarter of fiscal year 2027 will once again set a new record, reaching around $78 billion (excluding data center computing revenue from China). This represents a 14.5% sequential increase and roughly a 77% year-over-year surge. For fiscal year 2026, revenue is expected to continue its upward trajectory, with the previously forecasted $500 billion now revised upwards.

During the earnings call, NVIDIA's CEO, Jensen Huang, emphasized that computing demand is experiencing exponential growth. He noted that the turning point for Agentic AI has arrived, with enterprise adoption of this technology skyrocketing.

Meanwhile, concerns about AI, including the "AI bubble theory," are widespread in the market. These worries primarily revolve around the possibility that the substantial capital expenditures of AI giants may not yield a positive return cycle. According to reports, Google, Microsoft, Meta, and Amazon are set to collectively invest nearly $700 billion this year to fuel AI expansion, marking a more than 60% increase from 2025.

The significant capital expenditures of these major players provide a solid foundation for NVIDIA's short-term business. In 2026, the world's top five cloud computing service providers and hyperscale cloud enterprises contributed over 50% of NVIDIA's data center revenue.

Huang pointed out that customers are eagerly investing in AI computing, convinced that it will ultimately translate into profitable ventures. Regarding concerns about the sustainability of capital expenditures, Huang expressed confidence in the cash flow growth of cloud provider customers, stating that "AI computing power equals revenue."

However, amid these impressive results, the market's scrutiny of NVIDIA is evolving from a focus on singular growth metrics to a broader assessment of dimensions such as "earnings quality" and "cash flow health."

Experts recommend paying close attention to the following indicators:

Firstly, the sustainability of gross margins. With rising memory chip prices and supply chain costs, it remains uncertain whether the gross margin improvements driven by the Blackwell architecture can withstand these cost pressures.

Secondly, cash collection ability. Given the influence of hyperscale cloud providers on NVIDIA, investors are keen to see not just order contracts but also actual cash in hand to verify that AI computing investments can genuinely translate into sustainable cash flow returns.

Additionally, some experts view Small Language Models (SLMs) and Application-Specific Integrated Circuits (ASICs) as potential threats to NVIDIA.

Michael Burry, the real-life inspiration behind "The Big Short," stated, "For most AI applications, SLMs and ASICs represent the future direction. If necessary, they will be backward compatible with CUDA. NVIDIA's current solution is extremely power-hungry and inefficient, only temporarily holding its ground."

As AI is deployed on a larger scale, the scale of "model inference" will far surpass that of "model training," making cost sensitivity a critical factor. NVIDIA's general-purpose GPUs, designed to be compatible with the CUDA ecosystem and adapt to various computing tasks, inherently possess some redundancy in their chip architecture.

In contrast, ASICs tailored for specific tasks can eliminate this redundancy, achieving higher energy efficiency in certain scenarios. This is the rationale behind self-developed chips like Google's TPU and Amazon's Trainium, and why Burry believes SLM+ASIC could represent the future direction.

In December 2025, NVIDIA invested $20 billion, under the guise of "non-exclusive technology licensing," to acquire the core technological assets of Groq, a prominent player in AI inference chips. The company also recruited Groq's founder and core team, bolstering its capabilities in low-latency inference scenarios.

From this annual report, it is evident that NVIDIA remains at the forefront of the AI wave. Rapid earnings growth, sustained customer capital investments, and Huang's assessment of the arrival of the Agentic AI era all point in the same direction: demand for computing power has not yet peaked.

However, the market's focus has begun to shift from "how fast can it grow" to "how stable can it remain." The sustainability of gross margins, the quality of cash inflows, and potential technological shifts in the inference era could all influence investor pricing of this chip giant.