Headwaters Group's DATA IMPACT JOINT STOCK COMPANY Wins Grand Prize in Corporate Division at "Microsoft Agent Hackathon"

DATA IMPACT JOINT STOCK COMPANY (Headquarters: Vietnam, Representative: Dang Quang Duy, hereinafter "Data Impact"), a group company of Headwaters Inc. (Headquarters: Shinjuku-ku, Tokyo, Representative Director: Yosuke Shinoda, hereinafter "Headwaters"), which operates an AI platform business, has won the Grand Prize in the corporate division at the "Microsoft Agent Hackathon powered by Tokyo Electron Device" held on June 18, 2026.

A total of 240 teams participated in this hackathon from across the country, and Data Impact advanced to the final screening and won the Grand Prize in the corporate division.

The solution that received this award focused on the challenges in semiconductor factories, where highly advanced predictive maintenance is particularly required within the manufacturing industry. In semiconductor factories, the stoppage of a single piece of equipment can lead to losses of tens of thousands to hundreds of thousands of dollars per minute, making the prevention of sudden failures through predictive maintenance a critical management issue. In recent years, there has been a shift from conventional threshold monitoring, which detects anomalies when a certain standard value is exceeded, to AI-driven predictive maintenance, increasing the demand for advanced maintenance through data and AI.

Against this backdrop, the AI agent-based predictive maintenance solution "Etsunami Agent" developed by Data Impact presented a practical approach to industrial needs, leading to this award. This solution enables predictive maintenance by utilizing a data infrastructure on Microsoft Azure and an analysis environment by Databricks, with an AI agent continuously interpreting and making judgments on the data.

Event Overview Event Name: Microsoft Agent Hackathon powered by Tokyo Electron Device

Organizer: Classmethod, Inc.

Sponsor: Tokyo Electron Device Limited

Cooperation: Microsoft Japan Co., Ltd.

Event Period: April 7, 2026 - June 18, 2026

Number of Participating Teams: 240 teams

Award: Corporate Division Grand Prize

About the Award-Winning Solution "Etsunami Agent" In semiconductor factories, the losses due to equipment downtime are extremely high, and avoiding sudden downtime is a critical management challenge. Industry reports also indicate the significant ratio of downtime caused by sudden failures and the substantial cost savings achievable through predictive maintenance, strongly indicating a shift from "threshold monitoring to AI-driven predictive maintenance." The solution "Etsunami Agent," developed and proposed by Data Impact, is an autonomous AI agent for detecting equipment abnormalities in semiconductor manufacturing.

While conventional predictive maintenance could predict the probability of abnormalities using machine learning models, the interpretation of these results and the operational decisions regarding when to respond often relied heavily on human judgment.

This solution's distinguishing feature is its integration of cause analysis and response, going beyond simple anomaly detection. By leveraging a data infrastructure built on Microsoft Azure and a large-scale data analysis environment by Databricks, the AI agent continuously analyzes time-series data to detect early signs of equipment degradation trends and abnormalities.

Furthermore, after detecting an abnormality, it automates the execution of maintenance actions such as analyzing the background and impact of the abnormality, prioritizing responses, generating work instructions, and issuing notifications, allowing engineers to respond simply by reviewing and approving the content.

This achieves a transition from the conventional fragmented process of "detection → human judgment → response" to an Agentic AI-type maintenance operation that includes judgment and execution.

As a result, in addition to reducing the burden of monitoring and analysis tasks at the site and standardizing maintenance know-how, the factory as a whole can expect management impacts such as improved equipment utilization rates by reducing sudden stops, decreased production losses, and optimized maintenance costs.

This solution, based on Microsoft Azure and Databricks, serves as a reference model for AI agent-based predictive maintenance for the manufacturing industry, realizing everything from data collection and analysis to AI agent execution in an integrated manner.

Future Outlook We will further accelerate the deployment of AI agent solutions for the manufacturing industry. Based on the knowledge gained, particularly in the semiconductor manufacturing industry, we will strengthen development and proposal activities towards the practical application of Agentic AI, where AI autonomously solves business challenges, while also considering its expansion to other manufacturing sectors.

End

Reference Information ・Microsoft Agent Hackathon powered by Tokyo Electron Device (Official Event Page) https://zenn.dev/hackathons/microsoft-agent-hackathon-2026

・Etsunami Agent - Autonomous AI Maintenance Solution for Semiconductor Equipment Anomaly Detection (Technical Article) https://zenn.dev/headwaters/articles/f3a87f5704f166

About Trademarks Microsoft Azure is a registered trademark or trademark of Microsoft Corporation in the United States and/or other countries. Other company names and product names mentioned are trademarks or registered trademarks of their respective companies.

Company Information Company Name: Headwaters Inc. Location: 4F, Shinjuku Island Tower, 6-5-1 Nishi-Shinjuku, Shinjuku-ku, Tokyo 163-1304 Representative Director: Yosuke Shinoda Established: November 2005 URL: https://www.headwaters.co.jp/

Company Name: DATA IMPACT JOINT STOCK COMPANY

Location: Hanoi, Vietnam

Representative: Dang Quang Duy

Established: June 2023

URL: https://www.dataimpact.jp/