Author | Chen Wenqi
Nvidia CEO Jen-Hsun Huang once imagined: "Combining NVIDIA's AI technology capabilities with the vast ecosystem of ARM CPUs, we can advance computing from the cloud, smartphones, PCs, autonomous driving and robotics to the Edge Internet of Things, and extend AI computing to every corner of the world." ”
But his idea fell short. On February 8, SoftBank and Nvidia officially issued a statement that NVIDIA's $66 billion acquisition of Arm from SoftBank was terminated. This 16-month-long merger and acquisition in the semiconductor field ended in failure. AmD, on the other hand, completed the acquisition of FPGA giant Xilinx on Valentine's Day. The semiconductor industry has escalated competition, turbulence, and hegemony is difficult.
However, although ARM was missed in the arms race, NVIDIA is still the stable NVIDIA. After hours on February 16, local time (early morning Beijing time), NVIDIA released the fourth quarter of fiscal 2022 (as of January 31) and the annual report of fiscal 2022, with revenue exceeding expectations for 12 consecutive quarters to reach $7.4 billion, and the core business data center and game revenue broke records again:
During the reporting quarter, Nvidia's revenue was $7.643 billion, exceeding Wall Street expectations, up 52.8% year-on-year and 8% sequentially; of which data center business, gaming business and professional visualization business revenue were 4.32 billion, 3.263 billion and 643 million, respectively, all record highs; non-GAAP, net profit was $3.350 billion, an increase of 71% compared with $1.957 billion in the same period last year, an increase of 13% sequentially, and gross margin was 67%. Unchanged from the previous quarter; in fiscal 2022, NVIDIA revenue was $26.914 billion, up 61% year-over-year, beating analysts' expectations and non-GAAP net profit of $11.259 billion for the full year. In the next fiscal quarter guidance, revenue is expected to be $8.1 billion, with a 2% up-down fluctuation.
"We're seeing huge demand for NVIDIA's computing platform," Jen-Hsun Huang, founder and CEO of the company, said in its earnings report. "Nvidia is driving progress in some of the most influential areas of the moment, including AI, autonomous driving, robotics, gaming, and more." He said NVIDIA will release more hardware and software products at GTC in March.
In the face of NVIDIA's excellent performance, it fell 2.68% after hours, and by the time of publication, it stopped at $265.11, and the current market value is $662.8 billion. This is related to the conference call, NVIDIA said that many businesses are still affected by semiconductor supply chain issues.
Previously, affected by the rise in game graphics card prices, cryptocurrency mining fever, meta-universe becoming an outlet, cloud computing industry demand growth and other factors, Nvidia's stock price has risen all the way, from $59.02 in early 2020 to the highest point of $333.76 in November last year, becoming one of the hottest technology stocks last year.
Although in 2022, due to the ongoing tight semiconductor supply chain and the Fed's tightening policies, NVIDIA shares have fallen back and fallen by nearly 12%, but analysts are still full of confidence in Nvidia, with 41 of the 43 analysts in the Wall Street Journal giving a positive rating, of which 28 rated it as "buy".
NVIDIA stock price trend chart
Hit the track
As an infrastructure provider in different fields, NVIDIA divides its complex business into five segments in its earnings report: Gaming, Data Center, Professional Visualization, Automotive and Robotics, OEM and others.
Among them, the game business is NVIDIA's largest source of revenue, mainly including the company's traditional graphics card business. In the quarter, gaming revenue accounted for 44.7% of total revenue to $3.42 billion, up 37% from the same period last year and 6% sequentially. Revenue from this business increased 61% year-over-year in fiscal 2022 to an annual close of $12.46 billion.
In the gaming business, Nvidia has introduced the latest GeForce RTX 3050 desktop GPUs based on ampere architecture and GeForce RTX Ti and RXT 3070 Ti GPUs for laptops. More than 30 games were released this fiscal quarter, further optimized in terms of ray tracing (graphics rendering method), Relflex technology (reducing system latency), and more than 65 games were added to its cloud gaming platform GeForce Now, announcing partnerships with AT&T and Samsung.
Source: NVIDIA official website
Cloud computing, AI computing industry demand for high computing power, so that the data center business has become the strongest growth of Nvidia, but also the most certain growth of the business, Alibaba Cloud, Amazon AWS, Microsoft Azure, etc. are NVIDIA customers. In the field of supercomputing accelerators, NVIDIA occupies an absolute dominant position with a 90% market share.
In Q4 of fiscal 2022, data center business revenue reached $3.26 billion, up 71% year-on-year and 11% sequentially, and revenue increased 58% throughout the fiscal year to $10.61 billion, a new high.
In January, Meta and Nvidia officially announced that the former is using the NVIDIA DGX A100 system to build an AI supercomputer AI Research Supercluster used to train AI models, which is also NVIDIA's largest NVIDIA DGX A100 customer system.
Meta's supercomputer AI Research SuperCluster
In terms of AI software, NVIDIA AI Enterprise, NVIDIA's AI enterprise suite for infrastructure released last year, has also made all updates, adding production support for VMware vSphere with Tanzu, enabling developers to work on containers and virtual machines in vSphere environments.
Q4 revenue for the professional visualization business was $643 million, a rapid growth rate of 109% year-on-year, 11% sequentially, and fiscal year revenue growth of 100% to $2.11 billion.
Known as the Metaverse Infrastructure platform, the business's growth engine, has now opened a free version to millions of users worldwide. It is an open platform that integrates core rendering, AI and other technologies, built for virtual collaboration and physically accurate real-time simulation, and is provided by NVIDIA RTX technology to provide real-time collaboration, accelerating 3D workflows and addressing the massive 3D modeling needs of the metaverse.
Omniverse platform Image source: Nvidia official video
The automotive and robotics business was underperformed, declining for the third consecutive quarter sequentially, with Q4 revenue of $125 million, down 14% year-over-year and 7% sequentially, and fiscal year-over-year revenue up 6% year-over-year to $566 million.
The automotive business has not seen significant growth in the past fiscal year related to the tight automotive supply chain, the business accounts for a small proportion of NVIDIA's total revenue, and has not had a large impact on the company as a whole, but NVIDIA is more optimistic about the future of autonomous driving and continues to invest. Over the past fiscal quarter, Nvidia has partnered with Jaguar Land Rover to announce that the Robotaxi teams of NIO, Xiaopeng and Pony.ai are using NVIDIA DRIVE Orin.
There is strong uncertainty about when the growth of the automotive business will be released, but under the trend of automotive intelligence, Nvidia's bet may be cashed in the future.
OEM (Original equipment manufacturer) and other businesses Q4 revenue of $192 million.
From the perspective of financial indicators, NVIDIA maintains a healthy operating condition.
Benefiting from growth in its data center business and software suite, NVIDIA's gross margin remained high at 67% in Q4, compared to 66.8% in fiscal 2022. Under Non-GAAP, operating costs and operating income for fiscal 2022 were $5.279 billion and $12.690 billion, respectively, accounting for 19.6% and 47.2% of total revenue.
As mentioned in the earnings report, significant regulatory challenges hindered the completion of the ARM acquisition, and the Company and ARM parent SoftBank agreed to terminate the transaction and in the next fiscal quarter, a fee of $1.36 billion will be included in the table to reflect the write-off of the advance payment it made when it first signed the agreement in September 2020.
Overall, the complete layout of popular tracks such as games, cloud computing, meta-universe, and automatic driving has given Nvidia a strong basic disk, in which the growth of the game and data center business has a strong certainty. Last fiscal quarter, Nvidia's cryptocurrency mining processor (CMP) revenue fell to $24 million, down 77% from $105 million in the previous quarter, fluctuating more as cryptocurrency prices changed, but did not have a strong negative impact on NVIDIA's core business.
Infrastructure maniacs
The gold digger did not make money, but the man who sold the gold panning shovel made a lot of money, which may be the story of NVIDIA.
Cloud computing, Internet of Things, AI, blockchain and other technologies developed, a number of emerging areas emerged high demand for computing power, NVIDIA did these "gold mine" industry edge, the person who sold gold shovels.
At the GTC in 2021, Huang Jenxun, the helmsman of NVIDIA, known as "Lao Huang", upgraded the product route to the "three-core" strategy of "GPU + CPU + DPU".
According to the logic of this strategy, it is easy to understand why NVIDIA wants to take arm. Nvidia is stronger than GPUs, which is also the fundamental reason for its long-term growth in the context of the explosion of computing power, ARM has an absolute advantage in cpu architecture, and the acquisition can give Nvidia the ability to have both CPU and GPU.
Previously, Nvidia became the definer in the field of DPU chips with the acquisition of Israeli network equipment manufacturer Mellanox. However, the universality of the ARM architecture has made its customers Microsoft, Apple, Qualcomm, etc. worried that Nvidia's acquisition will damage its neutrality, and even risk leaking trade secrets.
Similarly, AMD, the "second oldest of the millennium", also wants to use the acquisition of Xilinx to have more complete capabilities in the field of data centers. Xilinx's business is mainly concentrated in the fields of FPGA chips, programmable SOCs and ACAP, FPGA and GPU are also computing accelerators, although the market size is not large, but can be programmed later, high flexibility, is an important technology for 5G communications, data centers and other industries.
Although Nvidia failed to achieve this acquisition, the route of continuing to work closely with ARM to enhance CPU capabilities will not change. At gtcc. 2021, Nvidia launched its first self-developed CPU for large-scale AI and high-performance computing applications in data centers, the Grace chip based on the Arm architecture, which is expected to be available next year. Its self-driving car, the SoC NVIDIA DRIVE Atlan, is also based on the ARM architecture.
Nvidia founder and CEO Jen-Hsun Huang Source: NVIDIA official website
The secondary market has been optimistic about Nvidia in the past two years, with the superimposed impact of the epidemic, metacosm, cryptocurrency, and U.S. stock market, since this year, the cryptocurrency market has cooled and the Federal Reserve has tightened, which has made Nvidia's stock price fall in the short term.
The Seeking Alpha article featured KeyBanc Capital Markets analyst John Vinh, who overweighted NVIDIA with a price target of $350, noting that market and macroeconomic conditions could interfere with the company's price target, but fundamentals remain strong.
WestPark Capital analyst Ruben Roy reiterated his buy rating and $365 price target, saying NVIDIA's outlook will be driven by positive demand for cloud data centers and "continued demand for high-end gaming GPU graphics."
Nvidia paid a quarterly cash dividend of $100 million to shareholders in Q4 and a total of $399 million in fiscal year 2022, with a quarterly cash dividend of $0.04 per share paying to shareholders next month.
From the CPU to GPU era, NVIDIA seized the opportunity to grow from a startup to a chip giant, and bet on a number of gold industries along the way, and the investment in research and development was huge, which was recently maintained at about 20% of total revenue. However, in the high-tech field, technology research and development and iteration are uncertain, and it is not easy to maintain a monopoly position in the market and high growth.