Microsoft Upgrades Azure AI Infrastructure with Nvidia H100 v5 GPU Virtual Machines

Microsoft has introduced a major upgrade to the artificial intelligence capabilities of its Azure cloud platform – the new ND H100 v5 series of virtual machines powered by Nvidia’s latest H100 GPUs. With significantly faster acceleration tailored for AI workloads, the new offering future-proofs Azure’s leadership in cloud-based AI development.

Key features of the H100 v5 VMs include up to 8 of Nvidia flagship H100 Tensor Core GPUs, which drive dramatic leaps in AI model training and inference speed versus previous gen hardware. Combined with 4th-gen Intel Xeon CPUs, vast memory, and bleeding-edge Nvidia Quantum-2 interconnect networking, the H100 virtual machines aim to handle even the most demanding AI tasks.

According to Microsoft, the ND H100 v5 series VMs allow matrix math operations to run up to 6 times quicker by utilizing Nvidia’s new FP8 truncated floating point format. This allows researchers to iterate faster.

Large state-of-the-art AI language models also see inference speed gains up to 2x when leveraging hardware FP8 support in the H100 GPUs. As model complexity expands exponentially, the optimized hardware pays dividends.

To future-proof AI development capabilities even further on Azure, Microsoft is deploying H100 v5 VMs at massive scale, with plans to add hundreds of thousands of the new GPUs to its infrastructure in 2024.

The substantial investment highlights Microsoft’s commitment to providing elite access to bleeding-edge AI accelerators via the cloud. Azure continues firmly establishing itself as the foremost cloud platform for creating and deploying next-generation AI innovations.

By integrating Nvidia’s most advanced data center technology like the H100 GPUs and Quantum-2 networking into new virtual machine configurations, Microsoft makes the power of AI supercomputing accessible to every Azure customer according to their needs and budget.

Access to nimble yet powerful H100 resources on-demand removes barriers to exploring new frontiers in AI like multimodal models and digital twin simulation. Azure marries flexibility with world-class performance.

Related Posts: Microsoft’s Push into Custom Azure AI Chips Upends Silicon Landscape

Latest

Nothing Ear 3 Earbuds Crank Up the Sound, Keep the Iconic Look

Nothing has launched its latest true wireless earbuds, simply...

Tony Fadell Warns Apple Has ‘Zero’ Data Center Muscle for AI Arms Race

In a scathing critique, former senior Apple executive Tony...

What to Watch in Qualcomm’s Q2 2024 Earnings Amid AI and Auto Ambitions

As Qualcomm prepares to drop its fiscal Q2 2024...

Qualcomm Expands Snapdragon X Family With New Entry-Level Snapdragon X Plus Chip

Qualcomm, mobile chipmaker doubled down on its processor roadmap...

Newsletter

Don't miss

Nothing Ear 3 Earbuds Crank Up the Sound, Keep the Iconic Look

Nothing has launched its latest true wireless earbuds, simply...

Tony Fadell Warns Apple Has ‘Zero’ Data Center Muscle for AI Arms Race

In a scathing critique, former senior Apple executive Tony...

What to Watch in Qualcomm’s Q2 2024 Earnings Amid AI and Auto Ambitions

As Qualcomm prepares to drop its fiscal Q2 2024...

Qualcomm Expands Snapdragon X Family With New Entry-Level Snapdragon X Plus Chip

Qualcomm, mobile chipmaker doubled down on its processor roadmap...

Meta’s Multimodal AI Unleashed on Ray-Ban Meta Glasses, Bringing Vision More AI Interactions

Meta enabled multimodal AI assistant on the Ray-Ban Meta...
jerrywanint
jerrywanint
Work through curiosity, & the internet.

Nothing Ear 3 Earbuds Crank Up the Sound, Keep the Iconic Look

Nothing has launched its latest true wireless earbuds, simply called the Nothing Ear, despite the confusing name, this is the 3-gen model Ear(3) following...

Tony Fadell Warns Apple Has ‘Zero’ Data Center Muscle for AI Arms Race

In a scathing critique, former senior Apple executive Tony Fadell warned the iPhone maker is massively ill-equipped to take on the AI superpowers when...

What to Watch in Qualcomm’s Q2 2024 Earnings Amid AI and Auto Ambitions

As Qualcomm prepares to drop its fiscal Q2 2024 earnings report, there are a few key areas investors will be laser-focused on to gauge...