With strong performance and doubled net profit, NVIDIA's market value evaporated by 300 billion yuan overnight

Text/Yang Jianyong

Nvidia suffered a major setback in the capital market after the earnings report, with its market value falling by $50 billion (about 317 billion yuan) overnight to its current market value of $612.6 billion. Unfortunately, despite the staggering operating indicators in fiscal 2022 and the huge profit of nearly 10 billion US dollars, it still cannot stop the sharp decline of 7.56% in the capital market. According to the financial report, the full year revenue of fiscal 2022 was $26.9 billion, an increase of $10.3 billion compared with fiscal year 2021, an increase of 61% year-on-year; and the net profit was $9.75 billion, an increase of 122% year-on-year. Both revenue and net profit are record.

Nvidia is lower in the capital market, both the current tense situation in Russia and Ukraine, but also the reasons for the excessive increase in the early stage, the market value continues to brush the record high, and there are hidden dangers hidden behind the vigilance of the surge. In fact, of course, Nvidia began a shock downturn after reaching a record high of $866 billion last November. As of February 17, 2022, the cumulative decline in the capital market this year was 17%, and the market value was reduced by US$122.6 billion (about 776 billion yuan) this year.

At the same time, more centrally, the acquisition of Arm failed, for which Nvidia needed to pay a break-up fee of $1.36 billion and write it down in the next quarter, thus affecting profitability in the first quarter of fiscal 2023. In my opinion, the termination of the acquisition of Arm means that NVIDIA is facing major challenges in developing high-performance CPUs. Prior to this, Nvidia extended to the CPU territory with the help of Arm, building three new processors based on Arm IP. In the PC field, create a new series of laptops. On edge computing, we partnered with Marvell to combine its Arm-based OCTEON processor with NVIDIA GPUs.

In particular, breakthrough Grace CPUs for large AI and high-performance computing workloads will take AI, the cloud and high-performance computing to new heights, including the Swiss National Supercomputing Center, which will build an Alps supercomputer, which will use the next generation of Grace and NVIDIA GPUs. Overall, the Grace-based system is tightly integrated with NVIDIA GPUs, and this innovative design increases aggregate bandwidth by a factor of 30 compared to today's ultra-fast servers, and 10 times more performance for applications running trillions of bytes of data.

However, Nvidia ultimately failed to make the acquisition of Arm, and its future development of high-performance CPUs was a heavy blow. For Nvidia, due to the acquisition of Arm, the focus on high-performance CPUs can not only improve the PC market, but also improve its competitiveness in the cloud data center market. At present, data centers, as an important segment of NVIDIA, are expected to replace games as its largest business segment.

Revenue from data centers was $3.26 billion in the fourth quarter of fiscal 2022, up 71% from the same period last fiscal year and 11% sequentially, and full-year revenue of $10.6 billion for the fiscal year showed strong growth. Revenue from the Games segment for the quarter was $3.42 billion, up 37% year-over-year and 6% sequentially, and full-year revenue was $12.46 billion. Through year-on-year and month-on-month growth, in the future, data centers will surpass NVIDIA's traditional game business.

You know, any technology giant in the world (including Google, Microsoft) relies on NVIDIA GPUs to train AI systems, riding the dust in the global artificial intelligence market. As a result, NVIDIA is even more hot in the capital market, becoming the world's largest chip designer, and its market value is equivalent to 3 Intel. NVIDIA is $418.9 billion (about 2.65 trillion yuan) higher than Intel's market capitalization. Behind it shows that Nvidia has shown strong competitiveness in cloud computing, enterprise and edge data centers, supercomputing, PCs and other markets. For example, Meta and Nvidia have teamed up to build large-scale AI research supercomputers.

Meta leverages the NVIDIA DGX A100 system to build its AI research super cluster. Meta hopes to build entirely new AI systems that, for example, can provide real-time speech translations for teams in multiple different languages, allowing them to collaborate seamlessly on research projects and even play AR games together. Meta's AI supercomputer is by far the largest NVIDIA DGX A100 customer system. Meta's AI research super cluster has hundreds of NVIDIA DGX systems connected via the NVIDIA Quantum InfiniBand network, helping the Meta AI research team increase the working rate. In the second phase later this year, Meta's AI research super cluster will expand to 16,000 GPUs.

This is the second time Meta has chosen NVIDIA technology as the basis for its research platform. In 2017, Meta built the first generation of AI research infrastructure, equipped with 22,000 NVIDIA V100 Tensor Core GPUs, which can handle 35,000 AI training jobs per day. Nvidia and several partners, including icrosoft Azure, set a record of eight hot workloads in the latest MLPerf training test, further elevating their leadership in AI. It can be seen that the global technology giants are relying on NVIDIA to develop AI technology, driving the market demand for NVIDIA AI surge, driven by super-scale and cloud expansion, making the data center sector revenue continue to set records.

In response, Nvidia CEO Jen-Hsun Huang said: "We see a strong demand for NVIDIA's computing platform. Nvidia is driving advances in artificial intelligence, digital biology, climate science, gaming, creative design, self-driving cars and robotics in many of today's most influential areas. In addition, NVIDIA also released NVIDIA Quantum-2 - the next generation of InfiniBand network platform, providing cloud providers and supercomputing centers with extreme performance, extensive access capabilities and strong security.

In NVIDIA's view, today's supercomputing centers and public clouds are converging, and they must provide the highest possible performance for the next generation of high-performance computing (HPC), AI and data analytics applications, while also safely isolating applications and responding to different user needs for traffic. With NVIDIA's Quantum-2 InfiniBand platform, modern data centers can already make this vision a reality.

Finally, the transformation of information technology to artificial intelligence, the active development of artificial intelligence technology around the world, and the injection of innovative vitality into industrial intelligence, artificial intelligence technology has been widely used in various industries, which will further promote the vigorous development of global artificial intelligence. According to IDC's latest forecasts, the global artificial intelligence market (including software, hardware and services) revenue is expected to reach $432.8 billion in 2022, an increase of 19.6% year-on-year, and it is expected that by 2023, this market size will exceed the $500 billion mark.

According to the report, it shows that the revenue of the global artificial intelligence market will show a high-speed growth pattern, and players on the track will be expected to continue to reap the dividends of industrial development.

Jianyong Yang, a contributor to Forbes China, is committed to in-depth interpretation of cutting-edge technologies such as the Internet of Things, cloud services and artificial intelligence.