Highreso Co., Ltd. (Headquarters: Shinjuku-ku, Tokyo; Representative Director: Kiyoyuki Shikura; hereinafter "the Company"), which operates the GPU cloud service "GPUSOROBAN," will co-host a webinar with APTO Co., Ltd. on the theme of "Successful Equation for LLM Data Development."
Webinar Overview
It greatly influences. In this seminar, we will organize everything from the premise of social implementation, which is the safety design of LLMs, to the concept of high-quality data that determines performance, and further to the potential for deployment in Physical AI.
Furthermore, we will explain the requirements for the "computing infrastructure" essential for putting these design philosophies into practice, and share design perspectives for GPU environments that balance development speed and cost optimization.
・Date & Time: Thursday, April 23, 2026, 14:00-15:00
・Viewing Method: Online (Zoom)
・Capacity: 1000 people
・Participation Fee: Free
・Application here: https://soroban.highreso.jp/423webinar
Target Audience for Webinar
・Those who want to systematically understand LLM safety design and the concept of high-quality data
・Those working to improve LLM accuracy in fields such as medicine, instruction following, and mathematics
・Those who want to understand dataset design necessary for Physical AI and the latest domestic and international trends
・Those interested in designing GPU infrastructure and optimizing costs to accelerate AI development
Speakers
・Yoichi Kano, APTO Co., Ltd.
After working at a general publishing company and a cybersecurity company, he joined APTO in 2022 and assumed his current position in 2024. He continues to support companies in solving AI challenges by leveraging data. Currently, in addition to overseeing business and corporate sides, he is actively developing and presenting research data with domestic and international research institutes, universities, and operating companies.
・Takuya Omi, APTO Co., Ltd.
From his time in graduate school, he was involved in planning operations for news media, working on business efficiency improvements using natural language processing libraries, AI system development, and directing product improvements based on user behavior data. Later, as a UX/UI designer for a video streaming service, he was responsible for user experience design and data-driven problem analysis.
Currently, as a project manager overseeing AI-powered service planning and development, he designs and promotes new products applying natural language processing (NLP) and large language models (LLM). Based on his experience in LLM training, he is also responsible for creating synthetic data and designing training methods that lead to accuracy improvements.
・Shunsaku Endo, APTO Co., Ltd.
He began seriously studying programming while enrolled in a specialized game programming school, where he received numerous awards in internal school competitions. As he honed his technical skills, he developed a strong interest in the AI field and co-founded APTO in 2020. He focused on the problem of inefficiency in data collection, preparation, and utilization in AI development and is working to solve it. Currently, he manages both development and business at APTO.
・Jun Fukamachi, Highreso Co., Ltd.
Responsible for sales of "GPUSOROBAN," a domestic cloud service utilizing cutting-edge GPUs. Providing high-performance GPUs at low prices from regional data centers in Japan, he supports AI research, scientific and technological computing (HPC), and DX utilization, primarily for universities, research institutions, and manufacturing industries.
He is committed to providing GPU clouds utilizing regional data centers, aiming to achieve both cost-effectiveness and reliability.
・Facilitator: Masato Yamaguchi, Highreso Co., Ltd.
After experience in sales and business development in advertising, publishing, and digital domains, he is now responsible for alliance strategy at Highreso. Centered on the GPU cloud business, he promotes collaborations and builds business partnerships with companies, research institutions, and partner companies. Through cross-industry collaboration, he is working to create new collaborative models utilizing GPUs.
About Highreso Co., Ltd.
Since 2019, Highreso has operated GPU data centers specializing in AI development and provides the GPU cloud service "GPUSOROBAN."
June 2022: Awarded NVIDIA "Best CSP Partner of the Year"
April 2024: Certified by the Ministry of Economy, Trade and Industry for the "Cloud Program" supply security plan
February 2026: A small-scale demonstration project for building a Green AI Data Center in the Kingdom of Bhutan was adopted for the Ministry of Economy, Trade and Industry's FY2024 supplementary budget for Global South Future-Oriented Co-creation etc. Project Subsidy (Small-scale Demonstration/FS Project).
About GPU Data Centers Specializing in AI Development
Highreso has established data centers in five locations in Ishikawa, Kagawa, and Saga prefectures, utilizing idle facilities such as former schools in rural areas. We promote GX by operating with a low environmental impact using unique construction know-how that leverages regional characteristics. We also contribute to the government's "Watt-Bit Linkage" concept, which aims to promote domestic economic growth by developing digital infrastructure and revitalize regional areas by decentralizing data centers.
• August 2019: Shiga Town No. 1 Data Center opened in Ishikawa Prefecture as a GPU-exclusive data center.
• August 2022: Shiga Town No. 2 Data Center, one of the largest in Japan, opened in Ishikawa Prefecture.
• December 2024: Takamatsu City Data Center, the first in the Chugoku-Shikoku region, opened in Kagawa Prefecture.
• August 2025: Genkai Town Data Center, the first in Japan to utilize a former school building, opened in Saga Prefecture.
• March 2026: Ayagawa Town Data Center, the second base in the Chugoku-Shikoku region, opened in Kagawa Prefecture.
About GPUSOROBAN
"GPUSOROBAN" is a GPU cloud service that accelerates massive computational processing for image generation AI, LLMs (Large Language Models), and more. By reducing data center construction and operating costs, it provides high-performance NVIDIA GPU servers at a low cost. GPUSOROBAN has been used over 2,000 times and is widely utilized across various industries, from IT to manufacturing, construction, and university research institutions.
Company Profile
Company Name: Highreso Co., Ltd.
Head Office: 106-1 Moroura, Genkai-cho, Higashimatsuura-gun, Saga
Tokyo Head Office: 3-24-1 Ichigaya-Tamachi, Shinjuku-ku, Tokyo
Representative: Kiyoyuki Shikura, Representative Director
Business Activities: Operation of GPU-exclusive data centers, provision of GPU cloud service GPUSOROBAN
Corporate Website: https://highreso.jp/
GPUSOROBAN: https://soroban.highreso.jp/
Contact for service inquiries: ln-gpu-sales@highreso.co.jp
Contact for release inquiries: press@highreso.co.jp
FACT BOX
- Source: PR TIMES
- Category: News