NVIDIA Intros Faster Chips, Cloud Services


By: Mary Jander

At its virtual GTC conference today, NVIDIA (Nasdaq: NVDA) unveiled a host of components and services geared to streamlining 3D applications in the so-called metaverse, with an emphasis on processors for advanced design and development of products in vertical markets such as automotive, medical, manufacturing, and logistics.

The stress on specific vertical applications for NVIDIA wares signals effort toward what the vendor sees as its best way forward -- AI and automation. Following are some products highlighted during NVIDIA CEO Jensen Huang’s keynote today:

--- GeForce RTX 40 Series GPUs. Designed for gaming applications, these new components are based on the vendor’s Ada Lovelace architecture, which emphasizes ultra low latency, high performance, and advanced graphics powered by artificial intelligence (AI). The series is led by the GeForce RTX 4090, priced at $1,599 and available October 12. Two alternate GPUs priced at $899 and $1,199 are set for November release.

--- NVIDIA Omniverse Cloud. This is NVIDIA’s first software-as-a-service (SaaS) offering that allows artists, developers, and design teams to collaborate over a network of 100 geographic regions powered by machines running NVIDIA’s RTX graphics processors and OVX servers. It’s available on AWS as well as directly from NVIDIA and can be accessed on remote devices as well as on-premises machines.

--- Thor. This is a system-on-a-chip (SoC) designed for automotive applications. Interestingly, it replaces Atlan, a SoC announced in 2021. Thor packs a whopping 1,000 trillion operations per second (TOPS) and goes into production in 2025. It is also designed to consolidate robotics design and implementation, Huang said.

--- Jetson Orin Nano. This is a family of “system-on-modules” and developer kit designed for edge AI used in the deployment of real-time intelligent machines such as drones, smart cameras, lidar gear, and other types of Internet of Things (IoT) apps. The series is 80 times more powerful than its predecessor Jetson Orin, Huang said, and it is in use across a range of companies, including Amazon, Boeing, Canon, Microsoft, Shopify, and Teradyne.

--- IGX Orin. The first of a series of edge AI developer platforms for creating secure, safe autonomous systems, such as those used in healthcare, factories, distribution centers, and warehouses. The platform allows developers to create apps that recognize threats to safety – such as pedestrians crossing in front of an autonomous vehicle – and react by stopping.

--- OVX Systems L40 GPU. A GPU for a computing system announced earlier in 2022, this chip is based on the Ada Lovelace architecture and is designed for building applications with AI, real-time graphics, and digital twin simulation. BMW Group and Jaguar Land Rover are using the OVX systems, NVIDIA announced. Also, the L40 will be used in datacenter systems based on RTX processors.

Automotive and Robotics Theme

In contrast to some past GTC conferences during which Huang introduced the concept of the omniverse, sometimes with cute avatars, this talk focused on the realities of what NVIDIA components can achieve, with particular attention paid to automotive and robotics applications.

This could be important for NVIDIA, which disappointed investors by falling short of financial expectations last quarter. By hammering home the specific use cases for each of its new components, the vendor is staking a claim to verticals it sees as vital to its future and more lucrative in the short term than gaming or PCs (though the powerful new GeForce RTX 40 series could spark some action in that sector as well).

Late this afternoon, NVIDIA shares were trading at $131.70, down 2.12 (1.56%).