Shares of Nvidia Inc., the undisputed king of cutting-edge chips powering artificial intelligence (AI) applications, rose the most in nearly nine months after the American chipmaker. has released even more impressive earnings forecasts.
Nasdaq-listed Nvidia's soaring stock price adds new momentum to what is being described as a “FOMO” AI stock rally, already propelling the Santa Clara-based chipmaker to the title of the world's most valuable chipmaker and surpassing Intel. It overtook well-known competitors such as Intel and Intel. A.M.D.
NVIDIA's revenue for the fourth quarter of 2023 was more than $22 billion, up more than 250% year over year and beating analyst consensus estimates of about $20 billion. Gross profit margin, a key measure of profitability, rose to well over 700% during the quarter.
“Accelerated computing and generative AI have reached a tipping point,” CEO Jensen Huang said in a statement. “Demand is surging around the world, across companies, industries and nations.”
according to bloomberg According to the data, NVIDIA's market capitalization has increased by more than $600 billion this year, bringing its valuation to nearly $1.9 trillion. Prior to the earnings release, Goldman Sachs analysts had labeled Nvidia the “most important stock on the planet,” meaning the company currently controls more than 80% of the AI chip market and will continue to do so. This reflects the fact that it is likely to maintain its dominant position. To the future.
Nvidia's rise appears to have lifted many boats, with shares of US “voice AI” startup SoundHound and British chip design company Arm, in which Nvidia recently announced an investment, in trading on Thursday (February 22). Both soared nearly 5%.
Rival AMD also rose more than 10%, and AI stock Palantir rose more than 2.5%. In addition to AMD, Broadcom, and Marvell Tech, other chipmakers in the AI race have also soared amid what analysts describe as AI FOMO (“fear of missing out”).
GPU wave
Nvidia controls expensive chips called graphics processing units (GPUs) that process data for AI models. There are two reasons for the shortage. One is that the generative AI boom has given GPUs the computational power and operational efficiency to perform calculations that allow AI companies to work on their LLMs (or large-scale systems). , the demand for these specialized chips is rapidly increasing. Language models such as ChatGPT and Bard) to truncate large amounts of data.
Second, NVIDIA's virtual monopoly on GPUs means the chipmaker is currently inundated with orders that it is struggling to deliver.
In its just-released 2024 AI Outlook, Moody's Investors Service predicts that “increased AI spending, improved models, and edge computing” will accelerate AI adoption and move companies “from exploration to adoption.” He said that investment in AI will increase as the industry progresses. “The shortage of high-performance graphics processing units (GPUs), essential for most AI computing, will continue into 2024, but supply will gradually improve,” Moody's noted.
Traditionally, the CPU (central processing unit) has been the most important component of a computer or server, with Intel and AMD dominating the market. GPUs are a relatively new addition to the computer hardware market, initially sold as cards that plug into personal computer motherboards to add computing power to AMD or Intel CPUs.
Nvidia's main claim for years has been that its graphics chips can handle the kind of bursts of computational workloads required by high-end graphics for gaming and animation applications much better than standard processors. . AI applications also require vast amounts of computing power, and backend hardware is increasingly GPU-intensive.
State-of-the-art systems used to train generative AI tools now deploy as many as six GPUs for every single CPU used, completely changing the equation where GPUs were seen as an add-on to the CPU. I did. Nvidia dominates the global market for GPUs and is likely to maintain this lead for the foreseeable future.
AI wave, data center demand
While Taiwan-based foundry specialist TSMC is the most important back-end player in the semiconductor chip business, Nvidia (in partnership with Intel, AMD, Samsung, and Qualcomm) is on the front end. Since 1999, when Nvidia first popularized the term GPU with its GeForce 256 processor, the company's chips have been coveted as shaping what's possible with graphics. His Nvidia GPU chips, such as the new “RTX” series, are currently at the forefront of the LLM-based generative AI boom.
Nvidia's data center business grew more than 10% for most of the last calendar year, versus flat growth for AMD's data center division and a decline in Intel's data center business. In addition to the GPU application, the company's chips are relatively more expensive per unit than most CPUs, which improves profit margins.
Perplexity AI founder and CEO Aravind Srinivas says Nvidia is leading the way in the AI chip race thanks to its proprietary software that easily leverages all GPU hardware capabilities for AI applications. … apparently …
Nvidia also has the systems to back up the processors and the software to run it all, making it a full-stack solutions company. The company is seen as one of the leading companies (and a potential rival to OpenAI and Google).
In addition to manufacturing GPUs, Nvidia provides an application program interface (API) called CUDA, a defined set of instructions that allows different applications to communicate with each other. This enables the creation of parallel programs using GPUs and is deployed at supercomputing sites around the world. world. He also has a presence in the mobile computing market with his Tegra mobile processors for smartphones and tablets, as well as in-vehicle navigation and entertainment systems.
Barrier to entry
Nvidia's resilience is a case study in a field that has very high barriers to entry and offers an incredible premium for specialization. The global semiconductor chip industry is dominated by a few countries and a small number of companies. Taiwan and South Korea account for about 80% of the world's chip foundry bases.
Only one company, Netherlands-based ASML, produces EUV (extreme ultraviolet lithography) devices, and advanced chips cannot be manufactured without them. Arm, a Cambridge, England-based chip design company in which Nvidia is a shareholder, is the world's largest supplier of chip design elements used in products from smartphones to gaming consoles (Nvidia wanted to acquire it).
This has very high barriers to entry, as China's SMIC, the domestic semiconductor champion, is currently reportedly struggling to procure advanced chip manufacturing equipment due to the US-led economic blockade. It turned out to be a nearly closed manufacturing ecosystem. In this market, his Nvidia, which has comprehensive control over chips used in high-end graphics-based applications, has come to dominate multiple end-use sectors including gaming, crypto mining, and now AI. became.
Diversification and business risk
Nvidia is releasing a new tool that will allow owners of its latest series of graphics cards to run AI-powered chatbots offline on Windows PCs, aiming to spread risk and further vertical integration. Announced. The tool, called “Chat with RTX,” allows users to customize generated AI models in the lines of OpenAI's ChatGPT or Google's Bard by linking to files, documents, and notes that can be queried.
“Rather than searching through notes or saved content, users can simply enter a query,” NVIDIA said in a blog post last Tuesday.
Nvidia's chatbot push comes as OpenAI CEO Sam Altman is seeking trillions of dollars in investment to revamp the global semiconductor industry. wall street journal It was reported earlier this month. Altman, who is behind the startup that launched ChatGPT, the fastest-growing consumer software application in history, has repeatedly expressed concerns about supply and demand issues for AI chips, and that's why OpenAI's growth stated that it is restricted.
The newspaper reported that Altman is currently in talks with several investors for a project to increase global chip manufacturing capacity. WSJ report. Mr. Altman is not alone. SoftBank Group CEO Masayoshi Son is also seeking to raise up to $100 billion for a competing semiconductor venture. bloomberg news This was reported on Friday, citing a person familiar with the matter. The project, codenamed “Izanagi,” is likely to supply semiconductors essential to AI, the report said.
These announcements are seen as indicative of the direction in which the generative AI discussion may go in the future. This also means that the lines between the two most important players in the AI story (Nvidia and OpenAI) are gradually blurring, and those currently standing on opposite sides of the hardware-software divide. indicates that their interests are likely to intersect.
