In a climactic moment for the tech giant NVIDIA, as its market value hovers closely around the threshold of becoming the top player in the industry, CEO Jensen Huang took center stage at CES 2025, often dubbed the "Tech Gala," to unveil a suite of groundbreaking productsThis moment has sparked questions about whether NVIDIA's stock will experience another surge in value.

During his keynote address at CES, which is renowned for showcasing cutting-edge technology, Huang introduced an array of new hardware, software, and services that underscore NVIDIA's commitment to advancing AI and gaming technologyAmong the highlights were the new consumer-grade GPUs based on the Blackwell architecture, alongside an innovative foundational model known as NVIDIA Cosmos, aimed at enhancing the comprehension of the physical worldFurthermore, a collaboration between Toyota and NVIDIA was announced, focusing on the development of next-generation autonomous vehicles.

Huang emphasized the significance of AI technologies in refining the training of robots and self-driving cars while also enhancing the capabilities of the latest gaming chips

Advertisements

He proudly projected that the volume of data generated by humanity in the next few years will surpass all data created throughout historyThis data is becoming increasingly multimodal, encompassing video, images, and sound, emphasizing the need for advanced computing technologies to process it effectively.

Among the most anticipated products was the new line of consumer GPUs: GeForce RTX 5090, RTX 5080, 5070 Ti, and 5070, which are priced at $1999, $999, $749, and $549 respectivelySet to launch later this month, these GPUs boast an astounding 92 billion transistors and deliver performance metrics like 4000 TOPS and four AI processing units, significantly outperforming the previous Ada architectureThe performance of the RTX 5090, for instance, will be twice that of its predecessor, the RTX 4090, illustrating NVIDIA's monumental leap in GPU capabilities from the 5070 to the 5090 series.

These advanced gaming chips leverage NVIDIA's proprietary Blackwell AI technologies to produce visually stunning graphics that resemble cinematic quality

Advertisements

One of the key innovations is in the realm of 'shading,' where imperfections and fingerprints can be added to surfaces, making representations of everyday objects like ceramic teapots astonishingly lifelikeAdditionally, these new chips enable game developers to create markedly more realistic facial animations.

The new laptops featuring the GeForce RTX 5090, RTX 5080, 5070 Ti, and 5070 are set to hit the market with prices starting at $2899 going down to $1299, with launches expected in MarchHuang noted that machine learning has fundamentally transformed application development and computational methodologies, unlocking new possibilities that extend far beyond traditional approaches, with GeForce GPUs playing a crucial role in harnessing AI for public use.

NVIDIA also introduced the Cosmos foundational model, which is designed to enhance the training speed of robotic and other physical AI systems

Advertisements

Cosmos distinguishes itself as the first model of its kind, trained on a massive dataset of 20 million hours of video, concentrating on physical dynamics and human interactionsThe model is versatile, able to accept prompts in the form of text, images, or video, and generate virtual states of the world tailored for autonomous driving and robotic applications.

By employing "synthetic" training data, the Cosmos model can significantly reduce the costs associated with training robots and self-driving cars, paralleling how large language models enable chatbots to generate responsesUsers need only provide a textual description to produce video outputs that abide by the laws of physicsDevelopers can use Cosmos to create AI feedback for reinforcement learning, enhancing strategy models and testing performance across varying scenarios.

This novel methodology promises to be much more cost-effective than current data collection practices

• Japan's Cautious Stance on U.S. Rate Cuts
• UK GDP Grows Only 0.1% in November
• The Global Wave of Interest Rate Cuts
• Anticipating the Q4 Earnings Season for U.S. Stocks
• The Great Global Collapse: Reasons Uncovered!

Currently, training self-driving cars requires fleets to roam city streets in search of data, while humanoid robots are typically trained by having humans repeat tasksHuang further indicated that despite the potential of the Cosmos model, achieving what he termed a "ChatGPT moment" would require more data acquisition upfrontThe Cosmos model will be available with an “open license,” similar to Meta's Llama3 language model"We truly aspire to bring to the world of robotics and industrial AI the kinds of contributions that Llama3 has made to enterprise AI," Huang stated.

Moreover, Huang emphasized that the future of robotics hinges on creating an effective world model that transcends mere linguistic understanding and incorporates physical dynamics, such as gravity and friction.

A pivotal announcement during the event was the collaboration between Toyota and NVIDIA to develop the next generation of self-driving vehicles

Toyota will adopt NVIDIA's automotive chips and operating system for several models, although specifics regarding these models remain under wrapsThis partnership also entails collaboration with other significant players like AURORA and Continental, joining NVIDIA's growing roster of industrial partnerships, including notable Chinese automotive companies such as BYD, Li Auto, Xiaomi, and Zeekr.

Huang projected that revenues from automotive hardware and software could surge from $4 billion this year to $5 billion by FY2026. He also introduced the next-generation automotive processor named "Thor," a revolutionary robotics computer designed to handle substantial amounts of sensor dataNotably, NVIDIA's Drive OS recently became the first certified programmable AI computing functional safety software, meeting the highest automotive safety standards, ASIL-D.

Another highlight was the unveiling of a desktop computer dubbed Project DIGITS, which is set to feature the same chips found in enterprise data center products

These chips will be housed in a smaller format, allowing individual software developers to swiftly test their AI systemsHowever, the initial version of Project DIGITS is not directly consumer-facing, as it operates on a Linux-based NVIDIA operating system primarily utilized by computer programmers, with a price tag of $3000.

With NVIDIA's stock closing at an all-time high of $149.43, resulting in a market capitalization of $3.66 trillion, the company currently stands as the second-largest publicly traded firm globally, following AppleInvestors are keenly watching whether Huang's latest address will catalyze another breakthrough in stock performanceAttention is particularly focused on updates regarding the Blackwell chips, regarded as NVIDIA's next significant growth driverEven though strong demand exists, Blackwell’s production has faced constraints that have delayed its rollout

Matt Chopa, a portfolio manager at Franklin Templeton Equity Group, remarked, "Demand for Blackwell is expected to remain extremely strongThis may refocus the market on NVIDIA's long-term opportunities."

Optimism among investors is well-foundedOver the last six months, Huang’s comments on chip demand have consistently propelled stock prices upwardsIn October, he characterized the demand for Blackwell as "crazy," and in November stated that demand was "very strong," with shipments expected in the current quarter.

Despite experiencing a downturn in December, NVIDIA's stock remains up 171% year-to-date, making it one of the major drivers of the S&P 500 indexThe stock has increased by 11% this year aloneWith a market capitalization of $3.66 trillion, NVIDIA is on the verge of surpassing Apple once again to reclaim the title of the largest company by market value, as Apple’s valuation stands at $3.7 trillion.

Currently, NVIDIA is still favored by Morgan Stanley analysts, who believe that this chip manufacturer will capture market share this year