In major leap of EDGE AI processing, NTT Corporation has announced a groundbreaking AI inference chip that can process real-time 4K video at 30 frames per second with power of less than 20 watts. This new, large-scale integrated (LSI) chip is the first in the world to achieve such high-performance AI video inference in a power-constrained environment, and is a breakthrough in edge computing applications.
It was revealed during NTT Upgrade 2025 The San Francisco summit is specifically designed for deployment on edge devices that are physically located near sources of data, such as drones, smart cameras, and sensors. Unlike traditional AI systems that rely on cloud computing for inference, this chip brings powerful AI capabilities directly to the edge, eliminating the need to send ultra-high resolution videos to centralized cloud servers for analysis, and eliminates the need to centralized cloud servers for analysis.
Edge Computing vs Cloud Computing: Why is it important?
In traditional cloud computing, data from devices such as drones and cameras is sent to remote data centers hundreds or even thousands of miles away. This approach offers virtually unlimited computing power, but introduces delays due to data transmission. This is problematic with real-time applications such as autonomous navigation, security monitoring, and live decision making.
In contrast, edge computing processes data locally above or near the device itself. This reduces latency, preserves bandwidth, and allows real-time insights even in environments with limited or intermittent internet connections. It also enhances privacy and data security by minimizing the need to transmit sensitive data over public networks.
NTT’s new AI chip fully embraces this edge-first philosophy, which introduces real-time 4K video analytics directly within the device, without relying on the cloud.
A new era of real-time AI for drones and devices
With this chip installed, the drone can detect people or objects up to 150 meters (492 feet) of the legal altitude limit for Japanese drones. This is a dramatic improvement over traditional real-time AI systems, usually limited to the 30m range due to lower resolution and processing speed.
This advancement allows for many new use cases, including:
-
Infrastructure inspection In a hard-to-reach location
-
Disaster response Areas with limited connections
-
Agricultural monitoring Beyond a wide range of fields
-
Security and monitoring No constant cloud uplinks
All of this is achieved with chips that consume less than 20 watts. This is dramatically lower than the hundreds of watts required by a GPU-powered AI server that is unrealistic on a mobile or battery-powered system.
In-chip: NTT’s unique AI reasoning engine
The performance of the LSI depends on NTT’s custom built AI inference engine. This ensures fast and accurate results while minimizing power usage. Important innovations include:
-
Interframe correlation: By comparing sequential video frames, the chip reduces redundant calculations and improves efficiency.
-
Dynamic Bit Fertilization Control: This technique uses fewer bits for simpler tasks to adjust the numerical accuracy required on the fly, conserving energy without compromising accuracy.
-
Native Yolov3 Run: The chip supports direct execution of You only watch v3 onceone of the fastest real-time object detection algorithms in machine learning.
These combined features allow the chip to provide robust AI performance in environments where advanced speculation was previously thought to have limited power or bandwidth.
The road to commercialization and the IOWN vision
NTT plans to commercialize chips through NTT Innovative Devices Corporation, the operating company within the second grade of 2025.
Researchers are already investigating integration into the innovative Optical and Wireless Network (IOWN)-NEXT generation infrastructure vision, aimed at overhauling the digital backbone of modern society. Within Iown’s data-centric infrastructure (DCI), the chip utilizes all-optical networks to complement the local processing power that brings to ultra-low latency, high-speed communications, and edge devices.
In addition, NTT is NTT Data, Inc. It combines CHIP functionality with Attribute-Based Encryption (ABE) technology. These technologies support AI applications that require both speed and security, such as healthcare, smart cities, and autonomous systems.
A legacy of innovation and a vision for the future
This AI reasoning chip is the latest demonstration of NTT’s mission to empower a sustainable and intellectual society through deep technological innovation. As a global leader with revenues of over $92 billion, NTT, with 330,000 employees and $3.6 billion annual R&D, serves more than 75% of Fortune Global 100 companies and millions of consumers in 190 countries.
Whether drones flying across the visual line, cameras that detect events in real time without cloud dependency, or protect data flows with attribute-based encryption, NTT’s new chip sets the stage for the next frontier of AI at the edge where intelligence meets immediacy.
