One of the most anticipated developer conferences, Nvidia GTC is upon us. This year’s GTC, scheduled for March 20-23, will focus on Nvidia’s latest advances in AI computing systems, generative AI, industrial metaverse, and robotics.
The Nvidia GTC would also feature a keynote address from Nvidia CEO Jensen Huang and host over 650 sessions with researchers, developers and industry leaders in almost every computing field. Huang’s keynote announcement will be live streamed on Tuesday, March 21 at 8:30 p.m. IST (8:00 a.m. PT).
In addition to presentations by DeepMind’s Demis Hassabis, Stability AI’s Emad Mostaque, and many others, the event will include a special fireside chat with Huang and OpenAI co-founder Ilya Sutskever.
Click here to register.
Nvidia has become almost synonymous with AI in recent months. That’s according to the company’s most recent earnings report, where AI accounts for a chunk of revenue. AI for Enterprise, the company’s new business model, leaves the air open for what Nvidia will bring to this segment.
Download our mobile app
Generative AI
Earlier, CEO Jensen Huang noted how generative AI will be the focus of his GTC event. Nvidia has already given some hints of what to expect – with NeMo and BioNeMo. These pre-trained layers of generative AI models, currently in early access with customers, are designed for enterprise customers looking to develop proprietary generative models and services to improve their business.
Huang doubled the strong sales figures from hyperscale customers like AWS, Azure, Oracle and GCP, and also announced Nvidia DGX Cloud, “the fastest and easiest way to have your own DGX AI supercomputer”, which will help in providing generative AI Applications will help businesses in the cloud.
GTC 2023 is likely to have an exhibition of generative AI used in various industries like healthcare, biology, etc.
This will be part of Nvidia’s AI cloud service offerings, enabling enterprise customers to have full access to AI computing across their private to any public cloud. The AI-as-a-Service model for AI computing, training, and software layers aligns with the Generative AI for Enterprise business model for enterprises.
There’s also 3D rendering, which is central to Nvidia’s overall plans for Metaverse. Last year, Nvidia released models like get3D and Magic3D that could render 3D models from 2D images better than existing benchmarks using a pre-trained text-to-image diffusion model. The company also released 3D MoMa, which allows developers to edit 3D representations.
Alongside this, Nvidia has conducted research in areas that include training AI agents in video game environments, diffusion models for generating high-quality images, 3D construction using approaches such as Latent Point Diffusion Modeling and PeRFception, a variation of NERF that can convey 3D information better . Research has also been done into creating animations from a set of source images.
AIM Forecast: AI shows no signs of slowing down. There are constant updates from OpenAI, Meta, Google and the like about the work being done in text-to-video generation, multimodal AI, consistency models, Indic language dataset training approaches and much more. With that in mind, Nvidia is likely to make some big announcements not only on 3D modeling but other generative AI applications as well.
At CES 2023, Nvidia also announced the Omniverse Avatar Cloud Engine (ACE), a suite of cloud-native AI microservices — encompassing text-to-speech and 3D animation — for creating and delivering interactive avatars. We’ll likely see many more additions to the services given how Nvidia is keeping up.
AI computing
Earlier, at the 2021 GTC event, Nvidia detailed the successor to the Bluefield-3, a 400-gigabit computing unit that is expected to launch by 2023. Bluefield-4 with a traffic throughput of 800 gigabits per second will have AI acceleration hardware added. Gazettabyte reports that this AI-assisted DPU will support tasks like video analytics, 5G and robotics. We can assume that Nvidia will give an update to Bluefield-4.
Meanwhile, the Nvidia Hopper H100 announced at GTC 2022 is now in full production. As the successor to the very successful A100 accelerator, H100 has already surpassed the sales of the A100.
AIM Prediction: Nvidia will likely provide use cases for how the H100s will be used in training and inference of transformer-based large language models, and how enterprises will use language models while keeping the data secure.
Additionally, according to Liftr Insights, the first sighting of the H100s being used in the public cloud is yet to be seen. The A100 also took up to a year to appear in the public cloud after initial production. We can expect Nvidia to accelerate the deployment of H100s in the public cloud as they focus on enterprise.
omniverse
No conversation about Nvidia goes without mentioning the Omniverse – such big are its plans. Earlier this year, Nvidia announced the Omniverse Avatar Cloud Engine (ACE), which will provide generative AI services to create digital avatars in the virtual world.
The Nvidia Omniverse platform is already being used for several digital twin applications such as Mercedes-Benz and BMW in their manufacturing and assembly facilities. The company also announced that it will work with Fujitsu to develop a new AI-on-5G solution that combines 5G vRAN, edge AI and digital twin workloads in a hyper-converged and GPU-accelerated all-in -One system combined.
We can expect Nvidia to announce more industry collaborations at GTC 2023 to enable digital twins and AI on 5G for metaverse applications and computer vision.
Additionally, at the recent CES event, Nvidia announced that Isaac Sim, its robot simulation application, will be available in the cloud, while simultaneously updating Isaac Gym for reinforcement learning and Isaac Cortex for collaborative robot programming. It also introduced Isaac ORBIT for simulation operating environments and benchmarks for robot learning and motion planning.
We’re also likely to see announcements related to the deployment of AI robotics at scale, as well as the development of AI models with faster learning from human behavior.
hardware, games and others
Nvidia has been making waves in the PC GPU market with the RTX 4090, 4080, 4070, 4060, and 4050 variants and has started shipping and powering some of the best gaming laptops around.
It is very likely that Nvidia will announce the GeForce RTX 4070 as rumors have it that it will be released in April. The GPU will be a mix of optimal price and performance, likely offering a mid-range option for end users. We can expect more updates to arrive on the release of 4060 and 4050.
Last year, Nvidia also announced the release of CUDA’s 12.0 software toolkit update, which focuses on new programming models and accelerated processing capabilities.
The company has also made the leap into building a platform for hybrid classical quantum computing. The QODA architecture powering this platform will be crucial to achieve any future quantum benefits. There are other reasons to believe that Nvidia will announce potential collaborations and use cases with QODA.
On the autonomous technology front, Nvidia Drive, an end-to-end modular development platform announced at GTC 2021, is being used by the likes of Arrow, Foxconn and Polestar to develop autonomous vehicles. With increasing demand from automotive companies for driver assistance systems for future development, GTC will also include more such partnerships and customer case studies in this area.
