NVIDIA's Q2 Earnings Expected to Surge 53% on AI Demand

Global investors are eagerly awaiting the release of NVIDIA's latest financial report, scheduled for tomorrow morning (Beijing time, Thursday morning). This report is not only a performance report for NVIDIA but also a benchmark for assessing the global artificial intelligence (AI) trend. Its influence can sway the sentiment of the tech sector and the entire market. Market expectations are high, and investors are keen to see if NVIDIA can once again surpass expectations. Key data points will be crucial in determining the outcome. Investors are advised to stay tuned for updates on the financial report.

NVIDIA is expected to release its second-quarter financial results this Wednesday afternoon, which are likely to include impressive figures. However, the question remains whether these figures will be impressive enough. Here are the expected financial metrics for the current quarter: Total revenue is projected to be 460 billion, a 53% increase from the previous year's 300 billion, and higher than the previous quarter's forecast of 450 billion. Earnings per share are expected to be 1.01 dollars.

Data center revenue is projected to be 413 billion, a 57% year-over-year increase. While this growth rate is robust, it is lower than the 73% increase reported in NVIDIA's previous financial report. Adjusted operating profit is expected to be 291 billion, with net profit projected to be 247 billion, higher than the previous year's 166 billion. Revenue forecast for the third quarter, ending in October, is 534 billion.

For the second quarter ending in July, the chip manufacturer's revenue and adjusted operating profit are expected to reach new highs, driven by the continued surge in AI system spending. For NVIDIA's shareholders, this marks another prosperous period, as the company's stock price has recovered from recent sell-offs affecting other AI-related tech stocks.

NVIDIA's stock price has risen nearly 39% in the past three months, outperforming other major tech stocks. The company's market capitalization of 4.4 trillion is currently 20% higher than Microsoft's. During the last earnings release, the company excluded its China business from the current quarter's forecast. Many analysts expect the company to do the same this time. However, data from FactSet shows that since the H20-related news was released last month, market expectations for NVIDIA's data center revenue for the fiscal quarter ending in October have increased by 5%. This could make it more challenging for the company to exceed Wall Street's targets, as it has done unexpectedly in the past three years.

Investors will be closely watching NVIDIA's performance expectations regarding its China business. On August 21, tech media reported that NVIDIA had instructed key component suppliers, including Samsung Electronics and Anker Technology, to halt production related to the H20 chip. Additionally, it was reported that NVIDIA had asked Foxconn to suspend H20 chip production. In response, NVIDIA stated that it continuously manages its supply chain to adapt to market conditions but declined to provide further details. However, analysts still believe that H20 chip sales have upside potential. A Morgan Stanley analyst reported that NVIDIA's existing H20 inventory could generate up to 6 billion in incremental revenue. It is expected that the China business could achieve double-digit growth in the coming quarters.

For several quarters, NVIDIA chips have been in high demand, and Wall Street will be looking for signs of sustained strong demand to gauge the company's potential for exceeding earnings expectations and the growth of the data center industry. It has been observed that the number of data center projects announced for 2026 has increased, which could bring NVIDIA up to 100 billion in incremental revenue over the next year. Demand is increasingly coming from sovereign clouds and new cloud developments outside the United States, indicating that NVIDIA's business scope is expanding.

The company's Blackwell chip has seen significant production increases, and the next-generation Rubin chip is scheduled for release in the second half of next year. Analysts will be paying close attention to more specific demand forecasts. It is expected that NVIDIA's CEO will have a clear outlook on data center demand for the 2026 calendar year, including the expected ramp-up of Rubin chip production in the latter part of the 2026 calendar year.

Will large-scale cloud service providers continue to invest? Large-scale cloud service providers, which are building AI infrastructure, have been continuously increasing their capital expenditure forecasts and purchasing NVIDIA's GPUs in large quantities. These companies' budgets, totaling tens of billions of dollars, are a significant benefit for NVIDIA. However, last week's significant drop in the U.S. stock market, led by large tech companies, has made some investors concerned about the sustainability of these large-scale investments.

Additionally, in NVIDIA's first-quarter report, the company revealed that 30% of its revenue came from just two customers, indicating a high concentration of customers. Analysts remain optimistic about the spending of large-scale cloud service providers, believing that these companies are engaged in a race to develop general AI (AGI) services and have plans to maintain this level of spending for several more years.

On August 25, NVIDIA announced the release of Jetson Thor, a supercomputer for robots based on the Blackwell architecture. This product is designed to provide computing power for millions of robots in industries such as manufacturing, logistics, transportation, healthcare, agriculture, and retail. Compared to the previous generation Jetson Orin, Jetson Thor's AI computing power has increased by 7.5 times, and its energy efficiency has improved by 3.5 times. It can run various generative AI models, including general models like Cosmos Reason, DeepSeek, Llama, Gemini, and Qwen, as well as robot-specific models like Isaac GR00T N1.5. The Jetson Thor developer kit is now available for purchase at 3499 dollars, and the Jetson T5000 module is priced at 2999 dollars for orders of 1000 units or more.

From computers and data centers to robots, NVIDIA's GPU computing platform has reached more endpoints and is expanding rapidly. In recent years, robots have become an important business line for NVIDIA, often featured as the highlight of their product releases. Jetson Thor is considered the most powerful "brain" for current robots, integrating "large models + real-time sensing + control" at the edge. Based on the Blackwell GPU and 128GB of memory, Jetson Thor delivers up to 2070 FP4 TFLOPS of AI computing power, enabling robots to run multiple modalities such as vision, language, and action (VLA) in parallel, significantly reducing cloud dependency and latency.

This means that robots equipped with Jetson Thor will have an even higher level of intelligence, capable of efficiently handling complex applications such as proxy AI, high-speed sensor data processing, and general robot tasks.