Nvidia GTC: What’s Next for the Global Technology Sector?
Founded in 1993 as a graphics processor company, Nvidia has undergone a remarkable transformation over three decades to become the central hardware provider for the AI revolution across the world.
It has accelerated companies across the world as businesses seek to harness Gen AI for competitive advantage and the demand for more powerful and efficient hardware has intensified, placing significant pressure on technology providers.
Now, at its annual GPU Technology Conference (GTC), the company has made several announcements that are sure to transform the global AI and technology market again, from new platforms to accelerate Gen AI, to innovations to address the challenges faced by multiple sectors implementing AI.
Technology Magazine looks at some of Nvidia’s top announcements that are set to transform the technology industry across the world.
AI Data Platform: To build infrastructure that supports AI reasoning workloads
The AI Data Platform enables Nvidia-Certified Storage providers to develop systems with specialised AI query agents – aiming to generate insights from corporate data in near real time, leveraging Nvidia AI Enterprise software including NIM microservices for the new Nvidia Llama Nemotron models.
“Data is the raw material powering industries in the age of AI,” says Jensen Huang, Founder and CEO of Nvidia.
“With the world’s storage leaders, we’re building a new class of enterprise infrastructure that companies need to deploy and scale agentic AI across hybrid data centres.”
Blackwell architecture: For Gen AI applications
The Blackwell architecture is Nvidia's first AI chip built specifically for accelerated computing and Gen AI applications.
The platform features multiple GPU models including the GB200 NVL72, which combines 72 Blackwell GPUs in a single system.
“AI has made a giant leap — reasoning and agentic AI demand orders of magnitude more computing performance,” says Jensen.
The new architecture offers significant performance improvements compared to its predecessor, the Hopper platform.
Nvidia claims Blackwell can deliver up to 4x faster training and 30x faster inference on large language models while using the same amount of energy.
Engineers have also integrated two GPU dies using Nvidia's chip-on-wafer-on-substrate technology, enabling the chips to function as a single GPU despite physically being two chips connected together.
Vera Rubin: The next Gen AI powerhouse
Nvidia also unveiled its upcoming Vera Rubin Computing System that is set to launch in late 2026.
This next-generation AI platform combines the custom-designed Vera CPU with the powerful Rubin GPU, offering significant performance gains over its predecessor.
The Rubin GPU delivers 50 petaflops of FP4 precision, 2.5 times faster than Blackwell Ultra, while the Vera CPU features 88 custom Arm cores.
When deployed in the NVL144 rack configuration, the system achieves 3.6 exaflops of FP4 inference computing power.
Spectrum-X networking platform: For AI workloads
Spectrum-X is Nvidia's new networking platform designed to optimise AI workloads across data centres.
The platform includes the Spectrum-X800 Ethernet switch, ConnectX-8 Ethernet NIC and BlueField-3 DPU.
“The massive amount of computation required for trillion-parameter models needs to be distributed across tens of thousands of GPUs,” says Jensen.
“Spectrum-X connects these GPUs with uncompromising bandwidth, enabling them to work as one.”
Expanded DGX Cloud: For new regions with Blackwell support
Nvidia has also expanded its DGX Cloud enterprise AI platform to include new regions and features.
The platform allows companies to access Nvidia's AI computing capabilities through major cloud providers including Oracle Cloud Infrastructure, Microsoft Azure and Google Cloud.
Now, the DGX Cloud includes support for the new Blackwell architecture, enabling enterprises to train and deploy larger AI models more efficiently.
- Blackwell Ultra GPU
- Vera Rubin computing system
- Spectrum-X networking platform
- Expanded DGX Cloud
- NIM inference microservice
- Enhanced Omniverse platform
- Upgrades to the Isaac robotics platform
Nvidia claims the improved capabilities will reduce the time and cost involved in developing custom AI applications.
NIM inference microservice: For AI model deployment
The company also unveiled NIM, a new inference microservice that simplifies the deployment of AI models.
NIM allows developers to deploy models with consistent APIs regardless of the underlying hardware.
“Established enterprise platforms are sitting on a goldmine of data that can be transformed into Gen AI copilots,” says Jensen.
“Created with our partner ecosystem, these containerised AI microservices are the building blocks for enterprises in every industry to become AI companies.”
Enhanced Omniverse platform: For industrial applications
Nvidia's Omniverse platform, a virtual world simulation and collaboration tool, received updates focused on industrial applications at GTC.
The platform now includes improved physics simulations and digital twin capabilities, which allow companies to create virtual replicas of physical systems for testing and optimisation.
Upgrades to the Isaac robotics platform: With new capabilities
Nvidia's robotics platform, Isaac, has been enhanced with new simulation capabilities and pre-trained models.
The improvements aim to accelerate the development of autonomous machines for manufacturing, logistics and service applications.
The financial impact of these announcements was immediate, with Nvidia's stock price responding positively.
“The age of AI has started. A new computing era that will impact every industry and every field of science,” says Jensen.
Explore the latest edition of Technology Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.
Discover all our upcoming events and secure your tickets today.
Technology Magazine is a BizClik brand
