After announcing the GeForce RTX 40 Super series of graphics processors primarily for gaming enthusiasts, Nvidia Inc. shares 64% to close at 522$53, the highest price ever**. NVIDIA is seen as the leading provider of processors for AI computing, and its stock price more than tripled in 2023.
Traders exchanged more than $32 billion worth of NVIDIA ** in Monday's trading, making it the most traded company on Wall Street, according to LSEG. NVIDIA's ** market cap is currently close to 1$3 trillion.
NVIDIA's latest help drive the Philadelphia Semiconductor Index 33%。
Nvidia's stock price action over the past five years.
Such an achievement is related to the fact that NVIDIA is consolidating its position as a data center chip while also broadening the application of AI in more fields.
Compete with the chip triumvirate
In the past, Nvidia's main AI chips have revolved around data centers, as its expensive server GPUs, including the H100, are critical to training and deploying generative AI, such as OpenAI's ChatGPT. Now, Nvidia is using its strengths in consumer-grade GPUs to enable so-called "native" AI that can run on a personal computer or laptop at home or office.
NVIDIA announced three new graphics cards on Monday — the RTX 4060 Super, RTX 4070 Ti Super, and RTX 4080 Super — *between $599 and $999. These cards have additional "tensor cores" designed to run generative AI applications. Nvidia will also supply graphics cards for laptops from companies such as Acer, Dell, and Lenovo.
The demand for NVIDIA enterprise GPUs has led to a surge in NVIDIA's overall sales, with a market value of more than $1 trillion, costing tens of thousands of dollars per GPU, and often equipped with systems with 8 GPUs working together. PC GPUs have long been Nvidia's staple products designed to run ** games, but the company says improvements have been made to graphics cards this year with an eye toward running AI models without having to send information back to the cloud.
The company said the new consumer-grade graphics chip will be mainly used for gaming, but it can still break through artificial intelligence applications. For example, NVIDIA says the RTX 4080 Super generates AI ** 150% faster than the previous generation model. Nvidia says other software improvements recently announced by the company will result in a five-fold increase in the processing speed of large language models.
Justin Walker, NVIDIA's senior director of product management, told reporters at a press release, "With 100 million RTX GPUs shipped, it provides a huge installed base for powerful PCs for AI applications. ”
Nvidia expects new AI applications to emerge next year to take advantage of the increased horsepower. Microsoft expects to release a new version of Windows, Windows 12, later this year, that can further leverage AI chips.
According to Walker, the new chip can be used to generate images on Adobe Photoshop's Firefly generator, or to remove backgrounds from calls. Nvidia is also creating tools that allow game developers to integrate generative AI into their games, for example, to generate dialogue with non-player characters.
NVIDIA's chip announcement this week suggests that while it is the company most associated with large server GPUs, it will compete with Intel, AMD, and Qualcomm, as will local AI. Because all three have released new chips that will power so-called "AI PCs" and be equipped with machine-learning-specific components.
Nvidia's move comes as the tech industry is looking for the best way to deploy generative AI, which requires a lot of computing power and can be incredibly expensive to run on cloud services.
One of the technology solutions introduced by Microsoft and Nvidia's competitors is called "AI PC" and is sometimes referred to as "edge computing." Instead of using powerful supercomputers over the internet, devices have more powerful artificial intelligence chips built into them and can run so-called large language models or image generators, albeit with some trade-offs and drawbacks.
NVIDIA proposes applications that can use cloud models to solve tough problems and on-premise AI models to handle tasks that need to be completed quickly.
NVIDIA GPUs in the cloud can run very large language models and use all the processing power to power very large AI models, while RTX Tensor Cores in PCs will run AI applications that are more sensitive to latency," said NVIDIA's Walker.
The company also said that the new graphics cards will comply with export controls and can be shipped to China, providing an alternative for Chinese researchers and companies who don't have access to Nvidia's most powerful server GPUs.
Become an important player in automotive chips
In addition to the above-mentioned chips, automotive chips are also the most important force of NVIDIA. According to the financial data, the company's automotive business reported revenue of 2$6.1 billion, up 3% sequentially and 4% year-on-year. Among them, China is undoubtedly an important focus of NVIDIA's efforts in automotive chips.
Nvidia said four Chinese EV brands, including Li Auto, Great Wall Motor, Zeekr and Xiaomi, will use its DRIVE technology as the brains of autonomous driving systems. Among them, Zengli Auto has selected NVIDIA Drive Thor** on-board computers to power its next-generation fleet. According to reports, NVIDIA Drive Thor is a next-generation centralized automotive computer that integrates a wide range of intelligent functions into a single AI computing platform, providing autonomous driving and parking functions, driver and passenger monitoring, and AI cockpit capabilities.
Nvidia says Li Auto currently uses two Drive Orin processors to power AD Max, the driver assistance system for its L-series models. These processors deliver a total of 508 trillion operations per second (TOPS) that fuse and process sensor information in real time to power all-scenario autonomous driving such as navigation on Advanced Driver Assistance Systems (ADAS), All-Scene Driver Assistance, and Lane Change Control (LCC), Auto-Parking, and Auto-Emergency Braking (AEB) active safety functions.
New ad max 30 upgrade transitions the system to an end-to-end algorithm architecture dominated by large-scale AI models. It leverages occupancy networks, spatiotemporal trajectory planning, and model** control algorithms to provide a safer and more comfortable intelligent driving experience.
In addition, electric vehicle manufacturers GWM (Great Wall Motors), Zeekr, and Xiaomi have also adopted the NVIDIA Drive Orin platform to power their intelligent autonomous driving systems.
According to reports, Great Wall Motor will build a self-developed high-end intelligent driving system, Coffee Pilot, based on the Drive Orin centralized computing platform. Coffee Pilot can support parking, high-speed, and urban scenarios, and realize full-scene intelligent navigation and assisted driving functions without the need for high-precision maps.
Great Wall Motor has partnered with NVIDIA to develop the intelligent driving system, and will launch the first model equipped with the system in the first half of this year. Advanced intelligent driving functions such as Urban N**iGate Autopilot and cross-layer memory parking will be the first to be launched on the Great Wall WEY modelZeekr's flagship car is powered by two Drive Orin system-on-chips, which provide smart parking and automated operation on highways and city roadsXiaomi's first electric car, the Su7 sedan, is based on a dual-Drive Orin configuration that enables highway driving. Built using Xiaomi's leading large language perception and decision-making model, the sedan is able to navigate seamlessly through Chinese cities regardless of geography, domestic administrative divisions, or road types.
Since 2022, Drive Orin has been produced with leading automakers, trucks, robo-taxis, and shuttles to deliver up to 254 TOPS and is scalable to support Level 2+ to Level 5 autonomous driving capabilities. Nvidia emphasized.
In Reuters' view, China's rising electric vehicle brand is a key market for Nvidia's automotive technology business. Chinese automakers are racing to introduce more advanced in-vehicle infotainment displays and autonomous driving features. This creates growth opportunities for Nvidia, Intel, Qualcomm, and other semiconductor manufacturers. But Nvidia and its U.S. rivals face an unprecedented challenge as they meet the needs of Chinese customers for powerful chips while complying with tighter U.S. controls on the export of advanced semiconductors to China.
In addition, NVIDIA partner Cerence has launched CALLM, a large language model tailored for the automotive industry, laying the foundation for its next-generation in-vehicle computing platform powered by NVIDIA Drive.
"The transportation industry is adopting centralized computing to enable highly automated and autonomous driving," said Xinzhou Wu, vice president of automotive at NVIDIA. The AI car computer of choice for today's smart fleets is NVIDIA Drive Orin, and automakers are increasingly looking to their successor, NVIDIA Drive Thor, for the advanced features and AI performance to shape future vehicle roadmaps. ”
Nvidia's worries
For Nvidia, although a lot of business is singing all the way. But as Nvidia founder Jensen Huang said, Nvidia has been in "danger." On the one hand, the escalating chip restriction testing in the United States is limiting Nvidia's play in their largest market, China. For example, yesterday's report pointed out that Nvidia's downgraded version of the chip made Chinese manufacturers unhappy and turned to domestic chips, which is undoubtedly a big challenge. (This can be seen in the report "Downgraded Nvidia Chips, Chinese Manufacturers Don't Pay");On the other hand, Nvidia is now pressing down on its heavy car chips, which may one day become the target of restrictions imposed by the ruling party in the United States, which is undoubtedly another potential blow to them.
Finally, now including Intel, AMD and Qualcomm, as well as the start-up AI chip industry in China and even the world, are also bringing a huge "threat" to NVIDIA. In particular, Intel and AMD, the two chip giants, whether they are investing in software or hardware, are expected to bring new uncertainty to Nvidia one day.
This may not happen in the short term. But in the future, who knows?
I'm a tech creator