Log In to unlock this content.
This content falls outside of your subscription, but you may view up to five pieces of premium content outside of your subscription each month
You have x unlocks remaining.
Artificial Intelligence for Everyone |
NEWS |
Awash in a sea of Artificial Intelligence (AI) announcements across hardware, software, and services, NVIDIA also highlighted the latest for its Omniverse platform, in the form of new Application Program Interfaces (APIs) available for customers. A handful of customer wins were announced to showcase those customers: Ansys, Cadence, Dassault Systèmes, Hexagon, Microsoft, Rockwell Automation, Siemens, and Trimble.
Directly from NVIDIA, the new APIs are as follows:
NVIDIA used Apple Vision Pro as an example of the pipeline from creation to consumption, with Cloud APIs enabling creation and upload to NVIDIA’s Graphics Delivery Network (GDN) that handles distribution through full fidelity streaming to the end device.
While not a primary focus for Omniverse and these announcements, there is, of course, still a tie to AI, found in the NVIDIA generative AI APIs made available in the platform as well. Generative AI is more experimental than productized today, especially for Extended Reality (XR), but paths are being created to add generative AI features when customers are ready.
An Understated, but Impactful, Omniverse Update |
IMPACT |
This marks a more “realistic” positioning of Omniverse, targeting applications and customer types that will find the most value in a cross-platform content creation, management, and distribution platform. This is mostly found in industrial and manufacturing, where the above customer announcements sit. For these companies, digital twin presents an ideal use case for Omniverse.
In the early days of Omniverse, NVIDIA appeared to struggle to identify a focus area, instead promising broad capability. This industrial focus is a clear indicator of NVIDIA identifying the strongest opportunity now, rather than being everything to everyone eventually. Omniverse’s potential is certainly there for use cases outside of industry, but there is a greater deal of education and use case alignment for these areas. Education, healthcare, and retail are all high-potential verticals for Omniverse, with varying limiters like budget and content availability impacting adoption.
Cloud APIs open significant possibilities for platform users, and the interconnected nature of Omniverse means the impact of these updates is as far-reaching as customers can reach with them. Four of the five APIs focus on adding capability for Universal Scene Description (USD), which has become the norm for Three-Dimensional (3D), XR, and broader immersive content creation standards. Standards are incredibly important for interoperability, so adding broader support for USD and the ability to edit, interact, track, and render USD data through an API is the best case scenario for Omniverse users. At least in theory, a little development time to integrate APIs adds significant spatial compute and data visualization capability for a customer—3D content creation and distribution are two of the primary barriers to XR adoption and scaling up implementations, no matter the industry.
Showcasing Apple Vision Pro and Omniverse streaming was an intelligent way to both confirm support for Vision Pro and highlight the level of fidelity for streamed content. Two main concerns come with streaming content—latency and fidelity—so a showcase of Vision Pro streaming at least promises low latency and high fidelity.
Omniverse as a Data Visualization Foundation |
RECOMMENDATIONS |
Despite AI being the expected focus for the GPU Technology Conference (GTC), Omniverse still managed to grab some of the spotlight, which speaks to NVIDIA’s view of the solution and the broader XR ecosystem—AI is obviously the star child for the company, but tools to leverage existing data center and Graphics Processing Unit (GPU) strengths, both with and without AI, will remain key for NVIDIA and competitors in similar markets.
Speaking of competitors, there are not many of note specific to Omniverse (with similar feature sets and capabilities), but there are an increasing number in the broader AI and data visualization market. NVIDIA has several advantages in the GPU compute space, especially with CUDA, a collection of frameworks, programming libraries, and development toolkits for NVIDIA GPUs. There is no competitor to CUDA at scale today, meaning NVIDIA remains the best option for most customers. However, a handful of companies have organized the Unified Acceleration Foundation (UXL), including Intel and Qualcomm, aiming to open source and update oneAPI. In theory, this opens up platform-agnostic AI and GPU acceleration capabilities—compared to NVIDIA’s entrenched position, it is early days for the effort, but notable that the competitive landscape is shifting at both the chipset and software levels.
Watch the industrial and manufacturing space for successes and difficulties with Omniverse and, more broadly, digital twin and enterprise data visualization. Interest comes in waves for most technologies, and enterprise visualization is no different. With NVIDIA pushing Omniverse, Apple entering the market with Vision Pro, and a slew of new AI-driven smart glasses on the horizon, we are seeing another wave of increasing interest stemming from not only AI, but the confluence of technologies with XR and data/content visualization included.
As case studies become more prevalent, specific performance and success data become clearer and should drive greater adoption of and engagement with Omniverse and similar platforms. Questions around the true Return on Investment (ROI) and time to value still exist, despite years of ongoing adoption and data availability, simply because each company has unique variables that are too impactful to generalize, including budget, size of organization (both total and hands-on with data visualization), types of roles, existing infrastructure and implemented technologies, etc. But customer wins and case study examples, as NVIDIA led with at GTC, is the first step to proving that those variables will have less impact going forward.
