Microsoft unveils AI advancements & new cloud technologies

As artificial intelligence (AI) continues to reshape various sectors, Microsoft is broadening its global cloud infrastructure to cater to both developers and customers worldwide. During the recently held Microsoft Build 2024 event, the company revealed advancements in tools and services designed to optimise AI solutions. These upgrades promise increased deployment speeds and reduced operating costs, emphasising customisation and flexibility in AI infrastructure.

Microsoft's AI infrastructure, known for its adaptive platform, is being enhanced to support intelligent AI applications with improved performance and resiliency, the company states. This is part of Microsoft's ongoing effort to deliver cost-effective computing and advanced generative AI capabilities. The company leverages its expertise to run AI supercomputers for prominent services such as Microsoft Azure OpenAI Service, ChatGPT, and Bing.

The recent introduction of Microsoft's in-house designed cloud compute processor, the Azure Cobalt 100, marks a notable milestone. The processor is engineered for general-purpose workloads on the Microsoft Cloud, offering up to 40% better performance compared to previous Arm-based virtual machines (VMs). Azure customers such as Elastic, MongoDB, Siemens, Snowflake, and Teradata are expected to benefit significantly from these new Cobalt 100-based VMs. Furthermore, Microsoft's customer conversation platform, IC3, is adopting Cobalt 100, achieving up to 45% enhanced performance.

As part of its AI infrastructure, Microsoft is integrating technologies along with its custom processors. The ND MI300X VM series, featuring AMD's MI300X Instinct Accelerators, is now generally available on Azure. This enables cost-efficient AI inferencing for advanced models like GPT-4, supporting various AI supercomputing scenarios from building large models to fine-tuning pre-trained ones.

Microsoft's comprehensive AI infrastructure blends multiple virtual machines and GPUs with diverse silicon options from AMD, NVIDIA, and Microsoft's Maia AI accelerators. This setup ensures a cost-effective and efficient platform for AI workloads. Users can leverage Microsoft Copilot or develop custom copilot applications on this optimised platform.

The introduction of the Azure Compute Fleet service further extends Microsoft's cloud infrastructure. This new service simplifies the provisioning of Azure compute capacity across different VM types, availability zones, and pricing models. This aims to optimise operational efficiency and enhance compute flexibility and reliability for both AI and general-purpose workloads.

Managing expanding computing estates can be challenging, and Microsoft is addressing this with AI-enhanced central management and security. The company is introducing Microsoft Copilot in Azure, designed to help manage operations across cloud and edge environments efficiently. This AI companion uses natural language processing to assist users with operational queries and recommendations.

Copilot in Azure, now in preview, allows customers to control access across their organisation, aligning with operational standards and security protocols. The tool simplifies database and analytics management by incorporating AI-enhanced features, and it now supports Azure Kubernetes Service (AKS) through straightforward management tasks.

Additionally, Copilot now provides natural language to SQL conversion and self-help for database administration, simplifying tasks for developers and database administrators alike. The introduction of new security enhancements, such as integration with Microsoft Defender for Cloud, helps streamline risk management and code fixes.