Nvidia has just made a series of announcements strengthening its position in the AI and ML market with the release of new GeForce RTX SUPER desktop GPUs, AI-centric laptop architecture, and improved Nvidia RTX software and tools for AI, when the market in an intensely competitive landscape around artificial intelligence, end-user hardware platforms (gamers or professional users), and developer tools.
The company’s strategy is based on leveraging its leadership in the PC space with more than 100 million RTX GPUs. In the data center, Nvidia faces competition from Intel and AMD, but has an edge through software and GPUs, and its Cuda software platform contributes to the company’s overall competitive advantage. Google and Amazon develop their own chips, but they also use Nvidia chips, giving them a strong presence in the market. In short, despite growing and fierce competition, Nvidia GPUs still dominate AI training in data centers, and their efficiency in inference promises strong growth. However, Nvidia will need to improve its performance/energy efficiency ratio to maintain its lead.
Nvidia is improving the PC experience by introducing tools like Nvidia TensorRT, which accelerates the Stable Diffusion XL model for text-to-image workflows, and Nvidia RTX Remix with generative AI texture tools.. An important development is the open source Nvidia TensorRT-LLM (TRT-LLM) library, which improves the performance of large language models (LLMs). This library supports a wider range of pre-optimised models for PCs and powers the Chat with RTX tech demo, enabling more interactive AI experiences.
Nvidia’s strategic advantage lies in addressing privacy, latency and cost efficiency requirements when running generative AI natively on PCs. This approach requires a significant installed base of AI-enabled systems and appropriate development tools to improve AI models for PC platforms. Nvidia is responding by innovating its entire suite of technologies to improve over 500 PC applications and games already accelerated by Nvidia RTX technology.
On the hardware side, Nvidia’s new GeForce RTX 40 SUPER series graphics cards, including the GeForce RTX 4080 SUPER and others, promise AI performance that makes video and image creation faster and more efficient. Nvidia believes these GPUs with Tensor cores will “unlock the full potential of generative AI on PCs”.
In addition, Nvidia’s collaboration with major manufacturers such as Acer, ASUS, Dell and others introduces a new wave of RTX AI laptops, dramatically increasing the capabilities of generative AI right out of the box. Mobile workstations with RTX GPUs can also run Nvidia AI Enterprise software for simplified AI and data science development.
On the software side, Nvidia uses its rich portfolio to facilitate the development and deployment of AI-based applications. To that end, AI Workbench is an unified toolkit that enables developers to create, test, and customize generative AI models and LLMs. This simplifies access to popular repositories and extends projects across platforms, from data centers to local RTX systems. In addition, the partnership between Nvidia and HP to integrate AI Foundation Models into HP AI Studio simplifies the development of AI models and makes it easier to discover, import and deploy optimized models.
TensorRT’s recent expansion of text-based applications and the addition of new pre-optimized models demonstrate Nvidia’s commitment to accelerating LLM on PC to improve text-based AI applications across all segments of the AI market. Finally, Nvidia and its partners are expected to showcase new computing applications and services powered by generative AI at CES. This includes Nvidia RTX Remix for game mastering, Nvidia ACE microservices for adding digital avatars to games, improved performance for a reliable broadcast model, and Nvidia DLSS 3 with frame generation for an enhanced gaming experience. To meet the growing demand for high-performance AI applications running on-premises, Nvidia is joining the trend of harnessing the power of advanced GPUs and dedicated software tools to integrate AI more deeply into customer and developer experiences. Although Nvidia faces competition from other chip makers and in-house hypercomputing solutions, Cuda’s strong software and hardware integration and widespread deployment ensure Nvidia’s healthy lead in the professional AI chip market.
