NVIDIA at CES 2025

"The ChatGPT moment for general robotics is just around the corner"

Nvidia kicked off CES 2025 with a keynote from the CEO, Jensen Huang, and unleashed a storm of new products and efforts that would redefine gaming, AI, and autonomous vehicles. With the official announcement on 7th January 2025, it has surely brought a new tide of enthusiasm among the tech-savvy as well as the analysis communities. So here's all about everything Nvidia revealed in CES 2025:

GeForce RTX 50-Series GPUs

The star of Nvidia's CES 2025 presentation was the launch of its highly anticipated GeForce RTX 50-series GPUs. These graphics cards, built on the groundbreaking Blackwell architecture, promise a significant leap in performance and visual fidelity.

Performance and Specifications

The lineup includes the following models:

Model	CUDA Cores	Boost Clock (GHz)	Memory	Price (USD)
RTX 5090	21760	2.41	32 GB GDDR7	$1,999
RTX 5080	10752	2.62	16 GB GDDR7	$999
RTX 5070 Ti	8960	2.45	16 GB GDDR7	$749
RTX 5070	6144	2.51	12 GB GDDR7	$549

• RTX 5090: The flagship model, priced at $1,999, boasts 21,760 CUDA cores, 32GB of GDDR7 memory, and a boost clock of 2.41GHz. It delivers twice the performance of its predecessor, the RTX 4090, and can handle demanding games like Cyberpunk 2077 at 238fps with ray tracing enabled. It requires a 1,000-watt PSU and can consume up to 575 watts.
• RTX 5080: Priced at $999, the RTX 5080 features 10,752 CUDA cores, 16GB of GDDR7 memory, and a boost clock of 2.62GHz. It offers double the speed of the RTX 4080 and requires an 850-watt PSU.
• RTX 5070 Ti: This card comes with 8,960 CUDA cores, 16GB of GDDR7 memory, and a boost clock of 2.45GHz, priced at $749.
• RTX 5070: The most affordable option in the lineup, priced at $549, features 6,144 CUDA cores, 12GB of GDDR7 memory, and a boost clock of 2.51GHz. It is expected to provide performance comparable to the RTX 4090.

Technological Advancements

The RTX 50-series GPUs also introduce several key technological advancements:

• Blackwell Architecture: This architecture now supports GDDR7 memory, enhanced PCIe Gen 5, and DisplayPort 2.1b UHBR20 ports. The Blackwell architecture also features neural shaders capable of compressing game texture data massively to help boost performance while reducing memory use.
• Neural Rendering: Nvidia is ushering in a new era of computer graphics with neural rendering. By integrating neural networks right into the rendering pipeline, the RTX 50 series can realize dramatic leaps in performance, image quality, and interactivity.
• RTX Neural Shaders: These will be special shaders that are bound to revolutionize the next decade in graphics innovations. Capable of compressing textures up to 7X to save a ton of graphics memory and create movie-level quality textures, and sophisticated game lighting with just a click.
• DLSS 4: This AI-powered technology introduces multi-frame generation that analyzes consecutive frames, creates a completely new frame through the power of AI, and effectively multiplies frame rates up to 8X over traditional rendering with image quality higher than native rendering.
• Nvidia Reflex 2: Powered by Frame Warp technology, Reflex 2 decreases latency as high as up to 75%, delivering gamers even greater responsiveness.
• FP4 Computing - RTX 50 series GPUs are the first consumer GPUs to support FP4 computing. It provides up to 2x increase in AI inference and will enable generative AI models to run locally on the device with a much-reduced memory footprint.

Pricing and Availability

The RTX 5090 and RTX 5080 will be available on January 30th, 2025. The RTX 5070 Ti and RTX 5070 are expected to be released in February 2025. It's worth noting that Nvidia may introduce more affordable options in the future, such as an RTX 5060, to cater to a wider range of users.

The release of the new RTX 5000 series marks the end of the long run of the RTX 4000 series, which had a longer lifespan than usual due to limited competition from AMD and Intel.

Project DIGITS

Nvidia also announced Project DIGITS, a personal AI supercomputer designed to empower AI researchers, data scientists, and students. This compact device, priced at $3,000, combines the power of the NVIDIA Grace Blackwell Superchip with 128GB of unified memory and 4TB of storage.

Key features of Project DIGITS include:

• GB10 Grace Blackwell Superchip: This system-on-a-chip (SoC) delivers a petaflop of AI performance at FP4 precision, enabling rapid prototyping, fine-tuning, and execution of large AI models.
• High-Capacity Memory and Storage: With 128GB of unified memory and 4TB of NVMe storage, Project DIGITS can handle demanding AI workloads.
• Desktop Form Factor: Despite its powerful capabilities, Project DIGITS is compact enough to fit on a desk, making it accessible to a wider range of user.
• NVIDIA Nemotron Models: Project DIGITS will support a new family of NVIDIA Nemotron models based on Meta Platforms Llama models. These models are designed to run locally on RTX AI PCs and are offered as NVIDIA NIM microservices.
• AI Blueprints: NVIDIA also introduced AI Blueprints for crafting digital humans, podcasts, images, and videos. These blueprints provide easy-to-use, preconfigured reference workflows for various AI applications.
• New Tools for Creators: Project DIGITS will be compatible with a new wave of low-code and no-code tools, such as AnythingLLM, ComfyUI, Langflow, and LM Studio. These tools enable enthusiasts to use AI models in complex workflows via simple graphical user interfaces.

Project DIGITS is expected to be available by May 2025.

NVIDIA Cosmos

Nvidia introduced NVIDIA Cosmos, a platform designed to accelerate the development of physical AI systems, such as robots and autonomous vehicles. Cosmos empowers developers to train AI agents in photorealistic, physics-based simulated environments, reducing the need for extensive real-world data collection and enabling faster and safer AI development.

Key features of NVIDIA Cosmos include:

• World Foundation Models (WFMs): These models allow developers to generate massive amounts of synthetic data for training and evaluating AI systems. They are trained on 20 million hours of video, focusing on physical, dynamic elements like humans walking and hands moving.
• Open Model License: Cosmos models will be available under an open model license, fostering collaboration and innovation within the robotics and AV community.
• Early Adoption: Leading companies like Uber, 1X, Agile Robots, and XPENG are already adopting Cosmos.
• Isaac Robot Simulation Platform: Nvidia has introduced new features in its Isaac robot simulation platform to produce large amounts of synthetic training data for robots.

NVIDIA Cosmos represents a significant step towards democratizing physical AI, making it easier for developers to create more sophisticated and capable robots and autonomous vehicles.

Automotive Partnerships

Nvidia announced a major collaboration with Toyota to integrate its Drive AGX Orin supercomputer and DriveOS operating system into Toyota's upcoming vehicles. This partnership will enable Toyota to offer more advanced driver assistance systems and accelerate the development of autonomous driving technology.

The Future of AI Agents

Nvidia's advancements in AI, particularly with the introduction of AI Blueprints and NVIDIA Nemotron models, are paving the way for a future where AI agents play a more prominent role in various industries. These agents will be able to perform complex tasks, automate workflows, and provide valuable insights.

However, managing and coordinating these agents effectively will be crucial. Nvidia predicts that enterprises will require an "orchestrator" to "conduct a symphony of agents." This orchestrator will help coordinate the numerous AI agents working across the enterprise, manage them, and ensure they all perform as anticipated.

Nvidia is already taking steps to address this need by creating an orchestration layer that supports the development of such solutions. This layer ensures that all agents work effectively together, regardless of which partner helped develop the Blueprint behind the agentic AI application.

Nvidia envisions many enterprises setting up AI factories to gather new insights and automate business-wide flows with AI agents. Accenture is helping businesses start taking these AI agents into production and prepare for that future by announcing its "AI Refinery for Industry with NVIDIA." Built on the NVIDIA AI Enterprise software platform, the AI Refinery will support businesses in deploying AI agents rapidly across various verticals.

Market Impact

The excitement of the new Nvidia CES 2025 announcements was so great among technology enthusiasts that their stock also saw a positive result. Nvidia stock rose 3.4% in the wake of Huang's speech, topping an all-time high from November. The confidence in investors was palpable, seeing continuous innovation and leading by the company in the AI and gaming sectors.

Implications for the Future

Nvidia's CES 2025 announcements have far-reaching implications for the future of gaming, AI, and automotive industries:

• Gaming: The RTX 50-series GPUs will raise the bar for visual fidelity and performance in gaming, enabling more immersive and realistic experiences. They will also empower game developers to create more complex and dynamic worlds with advanced AI-powered characters and environments.
• AI: Project DIGITS and NVIDIA Cosmos will empower researchers and developers to push the boundaries of AI innovation, leading to advancements in robotics, autonomous systems, and other AI-powered applications. This could lead to the development of more sophisticated autonomous robots, personalized AI assistants, and innovative solutions in fields like healthcare and education.
• Automotive: The partnership with Toyota and the development of NVIDIA Cosmos will accelerate the adoption of autonomous driving technology, paving the way for safer and more efficient transportation. This could revolutionize the automotive industry, leading to the development of self-driving cars and trucks that can navigate complex environments and improve road safety.

Public and Media Reactions

Nvidia's CES 2025 announcements generated a wave of excitement and anticipation among the public and media. Gamers expressed enthusiasm for the RTX 50-series GPUs, with many eager to upgrade their systems. Some users, however, expressed concerns about the high price of the flagship RTX 5090.

Industry analysts praised Nvidia's efforts in the development of AI and autonomous driving, noting that the company is a leader in these areas. The announcements also stirred conversations about how AI might disrupt many industries and what the future of technology might hold.