# Fixstars offers DeepSeek inference on GPUs in Japan

_Published Saturday, November 7, 2026 at 5:18 PM EST · AI, Infrastructure · Latest · Tier 2 — Notable_

![Fixstars offers DeepSeek inference on GPUs in Japan — Primary](https://assets.st-note.com/production/uploads/images/333476828/rectangle_large_type_2_427842d63666f10845616ebd2bd95a42.png?fit=bounds&quality=85&width=1280)

Fixstars is offering access to DeepSeek-V4.1-Flash-1017 through GPU servers in Japanese data centers, note reported in its November 6 coverage. Fixstars Inference Cloud supports APIs compatible with OpenAI and starts at 48 yen, excluding tax, per 1 million input tokens.

The company says processed prompts and outputs are neither saved nor used for training, and requests are not sent to servers in China. The service runs a model developed in China while processing requests in Japan.

## Sources

- [note](https://note.com/asalab/n/nbd8b7dfd468c?hl=en)

---
Canonical: https://techandbusiness.org/newswire/wNtTbbXTygj5urpoQlE6N3
Published: 2026-11-07T22:18:36.044Z
Story chronology: 2026-11-06T00:00:00.000Z
Retrieved: 2026-11-08T06:10:03.839Z
Publisher: Tech & Business (techandbusiness.org)