## Huawei's Big Step Forward in Agentic AI Technology

At the Mobile World Congress in Barcelona, Spain, the Chinese company Huawei made headlines by revealing its Agentic Core and AI Data Platform. These are not regular computer programs—they are intelligent systems that can think and act on their own. The most impressive feature is that Huawei cut the time it takes for AI to give its first response by 90%, which means AI agents can now answer questions almost instantly. This is a huge deal because waiting time was one of the biggest complaints people had about AI systems.

The company also introduced something called the Agentic Network, which is a special kind of phone network designed just for AI agents and robots. Imagine a road system designed not for cars driven by people, but for self-driving cars that talk to each other—that is what an agentic network does for AI agents. This technology could help robots work better and faster in factories, hospitals, and even at home.

Besides speed improvements, Huawei's system also got much better at finding the right information. The new system can find information accurately more than 95% of the time, which helps fix a big problem called hallucination—when AI makes up false information. This matters because if doctors, scientists, or business leaders use AI to help them work, they need to be able to trust the information it provides.

## New Operating Systems Built for AI Agents

On March 5th, a company called Amdocs showed off something completely new: an Agentic Operating System or AOS. Think of a regular operating system like Windows or macOS on your computer—it runs many programs at the same time. The AOS does something similar, but instead of managing different programs, it manages multiple AI agents that each specialize in different jobs. One agent might be good at finding data, another at making decisions, and a third at completing tasks like scheduling meetings or processing payments.

What makes this system special is that these AI agents remember what happened before. Scientists call this Episodic Memory—the ability to remember past events and learn from mistakes without having to retrain the entire AI system. This is similar to how humans learn: if you burn your hand touching a hot stove, you remember it and do not do it again.

## Scientific Research Gets a Boost in the United Kingdom

On March 4th, the United Kingdom government announced a major investment: £40 million (about $50 million) for a new Fundamental AI Research Lab. This lab will focus on solving one of the hardest problems in AI right now—something called the Mastery Gap. This is when an AI system stops getting better even when you give it more information to learn from.

The research will concentrate on Structured Language Models—special kinds of AI that work like scientists solving a problem step by step. Instead of just guessing the next word (the way most AI chat systems work), these models follow logical paths to reach correct answers. This is more like how humans think when we solve math problems or science puzzles.

## MIT Develops AI for Cancer Medicine

Scientists at MIT, a famous university in the United States, used agentic AI to create new medicine for cancer. The AI agents helped design and test drug combinations much faster than humans could do alone. This is important because cancer treatments are complicated, and finding the right medicine takes many years in normal conditions. By using agentic AI, scientists might help cancer patients get better medicine much sooner and at lower costs.

## How Agentic AI is Changing Software Development

Companies are now thinking differently about how to build software and websites. Instead of taking old programs and adding AI to them, they are designing brand new systems that use AI agents from the very beginning—this is called the agentic-first approach. These AI agents can review code (checking if programs are written correctly), test software, and make updates with very little help from human workers.

## The Good News and the Worries

Agentic AI is helping scientists and researchers work faster and smarter than ever before. Students who want to become researchers can now have AI agents help them write programs, look for information, and even organize their research—tasks that used to take weeks now take hours. However, scientists worry about fairness. These powerful AI tools cost money, and rich universities and companies might be able to use them much more than poor ones. This could create a skills gap where some researchers have amazing tools while others do not.

Also, there are security risks. If AI agents become powerful at doing things on the internet, hackers might trick them into doing bad things like stealing money or deleting important files. Scientists are working on ways to keep agentic AI safe and under human control.