Nvidia unveils groundbreaking AI innovations at CES 2025 that aim to progress the gaming, content creation, automotive and manufacturing industries.
The global semiconductor industry is experiencing unprecedented demand for AI-capable hardware.
This call is pressuring companies to integrate increasingly sophisticated AI capabilities into their products and services – whilst the worldwide tech industry grapples with the growing demand for innovation and moves away from purely cloud-based solutions.
It’s no secret that this shift has been driven by increasing concerns about data privacy, latency requirements and the rising costs of cloud computing for AI workloads.
However, new offerings have been released that directly address these market dynamics.
In what many consider the company’s most ambitious product launch to date, Nvidia, the company that currently commands over 80% of the enterprise AI chip market, has introduced several groundbreaking technologies that promise to reshape how we interact with AI-powered devices and applications.
At CES 2025, Nvidia’s CEO, Jensen Huang, unveiled a suite of AI innovations in his keynote speech that span consumer graphics, enterprise computing, specialised AI applications, high-end gaming, industrial robotics and autonomous vehicles – all in aim to significantly advance the frontier of consumer and enterprise computing.
Advancing graphics and AI for consumer devices
Firstly, Nvidia introduced its new GeForce RTX 50 Series GPUs, which is based on the Blackwell architecture.
“Now, we’re entering the era of “physical AI, AI that can proceed, reason, plan and act.” – CEO of Nvidia, Jensen Huang.
These graphics cards are designed to harness AI for improved gaming performance and visual quality.
GeForce RTX 5090
The GeForce RTX 5090, described by Jensen as “a beast”, contains 92 billion transistors and can perform 3,352 trillion AI operations per second.
DLSS 4
The company also unveiled DLSS 4, a technology that uses AI to generate additional frames in games, potentially increasing performance by up to eight times.
Jensen explained: “The latest generation of DLSS can generate three additional frames for every frame we calculate.
“As a result, we’re able to render at incredibly high performance, because AI does a lot less computation.”
AI foundation models for RTX PCs
For content creators, Nvidia also introduced AI foundation models for RTX PCs.
These models, offered as Nvidia NIM microservices, are designed to assist in creating digital humans, podcasts, images and videos.
Jensen highlighted the potential of these tools: “We’re creating a whole bunch of blueprints that our ecosystem could take advantage of. All of this is completely open source, so you could take it and modify the blueprints.”
Innovations in robotics and autonomous vehicles
Nvidia’s Cosmos platform was also announced as a significant development for robotics and industrial AI.
Cosmos
Jensen compared the potential impact of Cosmos to that of large language models on Gen AI, suggesting it could lead to a breakthrough moment for general robotics.
SOME KEY FEATURES OF NVIDIA’S LATEST INNOVATIONS:
- Cosmos platform advances physical AI in robots, autonomous vehicles and vision AI
- GeForce RTX 50 Series GPUs with the RTX 5090 features 92 billion transistors and 3,352 trillion AI operations per second
- AI foundation models for RTX PCs feature Nvidia NIM microservices for creating digital humans, podcasts, images and videos
- AI Blueprints for agentic AI integrates with platforms from providers like CrewAI and LangChain
He stated: “The ChatGPT moment for general robotics is just around the corner.”
Cosmos integrates various AI models and processing pipelines to enable physical AI systems like autonomous vehicles and robots to better understand and interact with their environments.
Several companies in the robotics and automotive sectors, including Uber, have already adopted Cosmos.
DRIVE Hyperion AV platform
Meanwhile, in the automotive industry, Nvidia presented the DRIVE Hyperion AV platform, built on the new Nvidia AGX Thor system-on-a-chip.
This platform is designed to support advanced autonomous driving capabilities and functional safety.
Jensen explained the complexity of building autonomous vehicles: “Building autonomous vehicles, like all robots, requires three computers: Nvidia DGX to train AI models, Omniverse to test drive and generate synthetic data and DRIVE AGX, a supercomputer in the car.”
He emphasised the importance of synthetic data in training autonomous vehicles, explaining that Nvidia’s AI data factory can scale “hundreds of drives into billions of effective miles” using Omniverse and Cosmos technologies.
Agentic AI and digital manufacturing
The company also expanded its AI offerings with the launch of AI Blueprints for agentic AI.
AI Blueprints
These blueprints, developed in partnership with leading providers like CrewAI and LangChain, integrate Nvidia AI Enterprise software to enable developers to build and deploy custom AI agents for various enterprise workflows.
Jensen noted the potential of these tools: “Developers can use Nvidia NIM microservices to build AI agents for tasks like customer support, fraud detection and supply chain optimization”.
Isaac GR00T Blueprint
Furthermore, in the manufacturing sector, Nvidia introduced the Isaac GR00T Blueprint for synthetic motion generation.
Jensen emphasised the importance of training robots efficiently in his keynote, using Nvidia’s Omniverse to generate millions of synthetic motions for humanoid training.
These innovations set the stage for Nvidia’s latest project: a personal AI supercomputer called Project DIGITS, further solidifying the company’s position at the forefront of AI technology.
Jensen noted: “All of the enabling technologies that I’ve been talking about are going to make it possible for us in the next several years to see very rapid breakthroughs, surprising breakthroughs, in general robotics.”
Project DIGITS: AI supercomputing for developers
Concluding the presentation, Jensen unveiled Project DIGITS, described as Nvidia’s smallest yet most powerful AI supercomputer.
Project DIGITS
This device, powered by the GB10 Grace Blackwell Superchip, is designed to bring high-performance AI computing capabilities to individual developers and creatives.
“Every software engineer, every engineer, every creative artist – everybody who uses computers today as a tool – will need an AI supercomputer”, Jensen stated, highlighting the potential widespread impact of this technology.
“It runs the entire Nvidia AI stack – all of the Nvidia software runs on this. DGX Cloud runs on this.”
“This is Nvidia’s latest AI supercomputer,” he said, showcasing the device to the audience.
“It started with perception AI – understanding images, words, and sounds. Then Gen AI – creating text, images and sound,” Jensen summarised.
Now, we’re entering the era of “physical AI, AI that can proceed, reason, plan and act.”
