Updated
Updated · MIT News · Jun 29
Murakkab AI System Dramatically Improves Efficiency of Agentic Workflows in Cloud Environments
Updated
Updated · MIT News · Jun 29

Murakkab AI System Dramatically Improves Efficiency of Agentic Workflows in Cloud Environments

1 articles · Updated · MIT News · Jun 29

Summary

  • MIT and Microsoft researchers developed Murakkab, an AI system to optimize agentic workflows for speed and energy efficiency.
  • Murakkab dynamically configures AI models, tools, and hardware based on user needs, reducing manual setup and resource waste.
  • Tests showed Murakkab cut computation, energy use, and costs by over 65 percent without sacrificing performance in cloud deployments.

Murakkab: Transforming Agentic AI with 4.3x Cost and 3.7x Energy Reductions in Cloud Workflows

Overview

Announced in early 2026, Murakkab is a groundbreaking system developed by MIT and Microsoft to tackle the major inefficiencies in modern agentic AI workflows. As agentic AI becomes central to many applications, current systems face challenges like fragmentation, poor resource use, and high coordination overhead, leading to increased costs and slower task completion. Murakkab is designed to address these issues, promising to revolutionize how AI agents interact and operate within complex tasks. This innovation aims to help developers and organizations overcome growing challenges in deploying efficient and scalable agentic AI solutions.

...