Microsoft Unveils New AI Products and Infrastructure Chips

Microsoft has developed two infrastructure chips for data centers powering artificial intelligence operations. These chips aim to accelerate AI performance and enhance security, according to Reuters.

The corporation has invested significant resources in creating its own silicon for general tasks and artificial intelligence. The company highlighted performance and cost advantages during the development of chips tailored to specific needs.

The first processor is focused on security, with installation set to begin next year. The second is responsible for data handling.

The company is striving to develop microprocessors capable of processing information in data centers at the speed required by artificial intelligence. It also announced a new liquid cooling system for servers.

Alongside the chips, Microsoft announced new tools for cloud service clients, enabling the creation and deployment of AI applications.

Azure AI Foundry facilitates switching between large language models for each task. This allows the corporation’s clients to transition from older versions of ChatGPT to new ones or products from other companies like Mistral and Meta.

Additionally, features have been introduced for launching AI agents capable of performing actions on behalf of users.

Among other announcements:

• Maia AI chip for working with the Azure OpenAI service;
• a feature for automating repetitive tasks in Microsoft 365 Copilot;
• Teams Copilot can understand and respond to questions about slides, web images, and other visual content;
• text translation in Teams and PowerPoint (to be available next year).

Previously, Microsoft announced the launch of AI-based tools for sending emails, managing documentation, and performing other actions on behalf of employees. 

Back in October, the corporation introduced an updated version of the AI assistant Copilot with voice mode and screen image understanding.