Microsoft Azure has revealed their latest Virtual Machines with NVIDIA’s high-end H100 GPUs.
These virtual machines are designed to boost the performance of generative AI applications like ChatGPT.
This development is a significant advancement in AI and highlights the use of sophisticated technology to improve the abilities of AI-powered programs, especially those that generate content.
Integrating H100 GPUs into Azure Virtual Machines will enhance the speed and efficiency of processing complex generative AI tasks, leading to faster and better-performing applications.
To fulfill the increasing need for cutting-edge AI solutions, it is essential to have a robust computing infrastructure, specialized services, and expert knowledge to tackle the ever-increasing complexity of the latest AI models.
The size and intricacy of these models are increasing exponentially, which requires powerful supercomputing resources to effectively develop and deploy AI applications that meet the needs of our customers. Therefore, it is essential to have the necessary infrastructure, tools, and expertise to fulfill our commitment to delivering advanced AI technologies to our clients.
Microsoft is drawing on its ten years of experience in supercomputing and managing large-scale AI training workloads to address the difficulty of developing AI infrastructure that can provide exceptional performance on a large scale.
The company is utilizing its expertise to develop AI infrastructure that can handle the significant complexity and size of the latest AI models. Microsoft relies on its graphics processing unit (GPU) accelerated virtual machines (VMs) in the Azure cloud to achieve this. These VMs are the foundation for various generative AI advancements from Microsoft and its customers.
Microsoft is committed to delivering advanced AI applications that meet its customers’ needs by using its supercomputing expertise to develop AI infrastructure that can support massive performance at scale.
Greg Brockman, the CEO of OpenAI, has highlighted the importance of partnering with Azure to design supercomputers that meet the rigorous demands of OpenAI’s AI training requirements.
This collaboration has enabled OpenAI to conduct research and alignment work on complex systems such as ChatGPT that would have otherwise been impractical. The partnership has enabled OpenAI to develop AI infrastructure capable of supporting massive AI training workloads and delivering optimal performance.
The field of generative AI is rapidly evolving and holds immense value across various industries. Microsoft is at the forefront of this change, incorporating AI into its products like Microsoft Bing, Edge, and Dynamics 365, making it an indispensable component of its software.
Microsoft is delivering exascale AI supercomputers to the cloud to support this advancement. Their AI infrastructure can handle the constantly increasing size and complexity of the latest models, making it possible to scale AI capabilities to new heights.
Drawing on our vast experience of providing multiple-ExaOP supercomputers to Azure customers worldwide, we offer a reliable infrastructure that can deliver genuine supercomputer performance.
This powerful solution is designed for large-scale deployments, making it ideal for organizations like Inflection, NVIDIA, and OpenAI, who have made substantial investments in this field. With this advanced technology, a new generation of large-scale AI models can be developed, which will undoubtedly transform the industry.
Scaling AI has always been a top priority at Azure. Our past initiatives, such as investing in comprehensive language models research, such as Turing, and engineering breakthroughs, like building the initial cloud-based AI supercomputer, were all aimed at preparing us for the era of generative AI. Our unwavering focus on these areas has put us in a prime position to leverage the latest advancements in the field and continue driving innovation in AI at scale.
Azure has deeply ingrained AI at scale in its values, drawing on its early investments in language model research and pioneering work like creating the first AI supercomputer in the cloud. With services like Azure Machine Learning and Azure OpenAI Service, Azure brings its AI supercomputer to customers to train and access large-scale generative AI models, respectively.
By making supercomputing capabilities more accessible, Azure helps startups and companies leverage AI at scale without needing expensive hardware or software investments.
NVIDIA and Microsoft Azure have a long-standing partnership to bring innovative AI solutions to businesses worldwide.
In this context, Ian Buck, NVIDIA’s VP of hyperscale and high-performance computing, highlighted the collaboration in developing the NDv5 H100 virtual machines. These machines promise to power a new wave of generative AI applications and services, further expanding the possibilities of AI at scale.
Microsoft has announced that ND H100 v5 is available for preview and will soon become a standard offering in the Azure portfolio, allowing anyone to take advantage of the potential of AI at scale in the cloud. Those interested can request access to the new VMs.