NVIDIA’s New AI Capabilities Put Amazon, Google and Microsoft on Notice
By Jim Lundy
The shutdown that we all experienced during the COVID-19 lockdown created a war for innovation. Some providers stood still, while others raced ahead with real innovation. One of the providers who has been hard at work during COVID-19 is NVIDIA, which announced some major new AI offerings at its conference last week. This blog discusses a few of these new moves.
Omniverse Will Challenge Facebook and Microsoft
While both Facebook and Microsoft are making noise about their metaverse, NVIDIA is also not standing still as it announced updates to its new omniverse platform. The NVIDIA omniverse avatar will allow enterprises to generate interactive AI avatars which in turn can connect to the companies speech AI computer vision and natural language understanding engines as well as simulation technologies.
Analysis: This is far beyond an avatar that humans control and represent in advance in digital labor since these avatars are essentially virtual agents that can do tasks and perform jobs in an intelligent way.
On Omniverse itself, NVIDIA is positioning it as more of a business focused environment so that enterprises like BMW can design factories and other businesses can create unique digital twins. We contrast this with approaches such as Facebook, which are focused more on entertainment and engagement at the consumer level but not focused on the enterprise.
NVIDIA: AI or Bust
Many know that NVIDIA has been positioned more in the semiconductor space as the maker of GPUs that are used by AI based applications. The truth is that NVIDIA is now an AI Software provider and that makes them both a supplier and a competitor of Cloud AI providers Amazon, Google and Microsoft.
The new products include 65 new and updated software development kits. Some of the highlights of what was announced:
NVIDIA Modulus – allows development of Physics AI Models
NVIDIA ReOpt – allows for vehicle route planning and logistics
NVIDIA cuNumeric – enables accelerated computing for the Python NumPy ecosystem
NVIDIA Triton Inference Server – updates to its open source offering that works with all AI models and frameworks
NVIDIA TensorRT – updated and also updates AI Models
NVIDIA Zero-trust Cyber Security Platform – for defending data centers in real-time
NVIDIA Riva and NVIDIA Custom Voice – Custom Voice is a subset of NVIDIA Riva and allows for custom text to speech and will be offered for virtual assistants and video conferencing
The Trend Towards AI Everywhere
NVIDIA is really highlighting a bigger trend, which is AI everywhere. Cloud Data Centers are becoming AI enabled, and due to its GPUs, NVIDIA is a supplier to Cloud providers and to enterprises who want to do this level of processing in their own data centers.
While NVIDIA is known in gaming and in Data Centers, and as a GPU supplier, they now are entering the applications arena and thus are a provider and a competitor to existing cloud tech titans.
We have entered the era of Intelligence and NVIDIA is as much of an AI company as they are a provider of GPUs. This represents a trend towards the 4th industrial revolution that cannot be ignored.
NVIDIA should be looked at as more than a GPU and a Gaming platform provider. The bottom line is that intelligence is here to stay. The question is, how will your enterprise take advantage of it?