[Seminar] Risks of AI Data Center Design and Transformation of Business Models
[Speaker]
Kota Sakatsu
Representative of Sakatsu Research Room
Visiting Professor at Ritsumeikan University Graduate School of Management
[Key Lecture Content]
Traditional hyperscalers can be likened to a "dinosaur" business model that enhances air conditioning efficiency and profitability through increased floor space. However, currently, rising GPU heat output, soaring equipment costs, and energy procurement are intensifying the uncertainty surrounding large-scale investments, raising concerns about the risks of expanding AI data centers with the same mindset as hyperscalers.
In particular, in cooling technology, the water cooling method, which has garnered attention, is unable to withstand GPU requirements within 5 to 10 years, leading to liquid immersion methods being viewed as a strong candidate. This necessitates a review of overall architecture and infrastructure. As a result, an approach that balances short-term demand increases with mid-term structural design revisions is required, and the operation of small-scale units such as containers that can control heat, energy, and weight is emerging as a practical solution for the time being.
[Lecture Items]
1. Current Status of the AI Data Center Market
2. Business Models and Risks of AI Data Centers
3. Evaluation of Unit-Type Data Center Models
4. Q&A / Business Card Exchange

| Date and time | Tuesday, Aug 18, 2026 10:00 AM ~ 12:00 PM |
|---|---|
| Capital | On-site attendance / Live streaming / Archive streaming (available for viewing multiple times for 2 weeks) SSK Seminar Room, 2-6-2 Nishishinbashi, Minato-ku, Tokyo, 4th Floor, Zaimax Nishishinbashi Building |
| Entry fee | Charge |
More Details & Registration
Details & Registration




