News

(※1) Total theoretical computing performance refers to the maximum theoretical computational performance based on GPU specifications.
(※2) Theoretical value based on FP16 Tensor Cores (without sparsity)

【Background of the Strategic Partnership Agreement】

～Turing’s Initiatives～
Turing is a startup working on the development of fully autonomous driving. It is simultaneously developing E2E (※3) (End-to-End) autonomous driving AI, which uses a single AI to handle everything from environmental perception to path planning and driving control, as well as large-scale foundation models that acquire an understanding of common sense, background knowledge, and context in human society.
By integrating these technologies, Turing aims to realize “fully autonomous driving,” in which vehicles can perform driving operations in place of humans under all conditions, and is concentrating its management resources on the development of autonomous driving AI systems.

(※3) End-to-End (E2E): Unlike conventional systems that process tasks by separating them into multiple functions such as “perception → decision-making → control,” this refers to a unified processing approach in which a single AI model handles everything from input data obtained from sensors such as cameras to control commands for steering, acceleration, and braking.

～GMO Internet’s Initiatives～

Since November 2024, GMO Internet, Inc. has been providing “GMO GPU Cloud,” a GPU cloud service that offers strong computing performance and a highly user-friendly development environment, in response to the growing demand for advanced computational processing driven by the evolution of AI and robotics. Following the deployment of NVIDIA H200 Tensor Core GPUs, “GMO GPU Cloud” began offering its latest GPU cloud service equipped with “NVIDIA Blackwell Ultra GPU (B300)” in December 2025, among the fastest in Japan. By stably supplying high-performance computing resources from domestic data centers, we contribute to strengthening Japan’s AI development capabilities and fostering innovation that transcends the boundaries between industry, government, and academia.

～Technical Strengths of GMO GPU Cloud～

(1) Service Features
Services equipped with NVIDIA H200 Tensor Core GPUs provide, as standard, Japan’s first high-speed NVIDIA Spectrum-X networking, a high-performance storage environment, and a simple, fully managed job scheduler system.
For cloud services featuring the “NVIDIA Blackwell Ultra GPU,” GMO Internet has introduced the “NVIDIA HGX B300 AI Infrastructure,” optimized for AI inference models. This delivers a high-performance environment ideally suited for cutting-edge AI workloads, including high-speed inference of large language models and the development of agent-based AI.

(2) Support for the Latest Technologies
At GMO Internet, Inc., we rapidly adopt NVIDIA’s latest generations and continuously evolve our network, storage, and operational infrastructure in an integrated manner to consistently provide the optimal execution environment for cutting-edge AI workloads. In addition, we continuously expand our capabilities by implementing features such as Multi-Instance GPU (MIG), the web portal software “Open OnDemand,” and monitoring dashboards (Grafana), thereby enhancing resource efficiency and accelerating development cycles.

(3) Significance as a Domestic Provider
With a 24-hour support system provided by our Japan-based engineering team, we significantly reduce the burden on AI developers. Furthermore, as a “sovereign cloud” that stores and manages all data within domestic data centers, we ensure that critical technical information and data sovereignty remain in Japan. Turing’s vision of challenging the world with technology originating in Japan aligns with the mission of GMO Internet, Inc. to support Japan’s domestic AI industry, leading to the realization of this partnership.

～Significance of the Strategic Partnership～

Turing has already adopted “GMO GPU Cloud” for its fully autonomous driving AI development. With this initiative, GMO Internet, Inc. will invest 3.2 billion yen in Turing and conclude a four-year long-term contract for “GMO GPU Cloud,” thereby providing the company with stable funding and computing resources. Through this partnership, we will expand our stable supply and support framework to meet growing demand for computing resources and further accelerate development toward the realization of fully autonomous driving. In addition, by leveraging the expertise gained through environment design and operations, as well as feedback from Turing’s cutting-edge AI development sites, we will establish a framework that continuously delivers the best possible research and training environment.

【Key Terms of the MOU】

【Comment from Turing CEO Kazunari Yamamoto】

Turing keeps development as simple as possible and achieves high development speed by quickly incorporating technological trends from AI development around the world. GMO GPU Cloud, with its high-speed network and storage, aligns extremely well with our development philosophy, and its 24/7 support system has significantly reduced the burden on our engineers.
Through this strategic partnership, we will further strengthen our AI computing infrastructure and accelerate development toward the realization of fully autonomous driving.

【Future Outlook】

Advances in AI technology will accelerate not only autonomous driving but also the development of a wide range of industries, including healthcare, manufacturing, and agriculture, and will directly contribute to addressing Japan’s structural challenges such as an aging population and labor shortages. Among these, fully autonomous driving is a cutting-edge field in which companies around the world are fiercely competing in technological development. For Japanese startups to create global standard technologies, it is essential to combine access to state-of-the-art computing resources with a stable, long-term development framework.
Through this strategic partnership, GMO Internet will continue to evolve “GMO GPU Cloud” and strongly support Turing’s development of fully autonomous driving systems. As an infrastructure provider delivering computing resources, GMO Internet will continue to take on the challenge of building a foundation that enhances industrial competitiveness and the sustainability of society through the advancement of domestic AI technologies.

【About GMO Internet】

GMO Internet, Inc. is a comprehensive Internet company engaged in Internet Infrastructure—including domains, cloud and rental servers, and Internet connectivity—as well as Online Advertising and Media. Under the corporate slogan “Internet for Everyone,” the company supports social infrastructure by helping realize a safe and secure Internet society, while also contributing to value creation that builds a new future with AI. GMO Internet remains committed to delivering “smiles” and “inspiration” to everyone it works with.

【About Turing】

Turing is a startup dedicated to the development of fully autonomous driving. It is developing an E2E (End-to-End) autonomous driving system in which AI makes all driving decisions—such as steering, acceleration, and braking—using only data obtained from cameras.
Through innovations in the autonomous driving domain, including the multimodal generative AI “Heron,” which performs advanced decision-making using multiple types of data; “Terra,” a generative world model for autonomous driving capable of producing realistic driving scenes as video; and the autonomous driving VLA model dataset “CoVLA Dataset,” which can describe driving environments derived from images in natural language and generate appropriate route planning, Turing aims to realize fully autonomous driving.

・Heron - https://tur.ing/posts/n7UT9tmK
・Terra - https://tur.ing/posts/V0YV1eyS
・CoVLA - Dataset https://tur.ing/posts/ACnNV8lk

【About “GMO GPU Cloud”】
（URL：https://gpucloud.gmo/）

“GMO GPU Cloud” is equipped with NVIDIA H200 Tensor Core GPUs and implements Japan’s first high-speed NVIDIA Spectrum-X network along with high-performance storage.
In the global supercomputer performance ranking “TOP500” (※4) announced in November 2024, it ranked 37th worldwide and 6th in Japan, providing one of the fastest computing infrastructures among commercial cloud services in Japan. Furthermore, in June 2025, it achieved 34th place worldwide and 1st place in Japan in the global power-efficiency ranking “Green500” (※5), earning international recognition for combining high performance with energy efficiency. In addition, in December 2025, GMO Internet began offering a cloud service featuring NVIDIA’s next-generation GPU, the “NVIDIA HGX B300” equipped with the “NVIDIA Blackwell Ultra GPU” (※6).
Through this service, GMO Internet provides companies and research institutions working in the generative AI field with optimized infrastructure and flexible, customizable computing environments tailored to customer workloads. By contributing to shorter development cycles and lower costs, the company promotes the growth of the domestic AI industry.


(※4) “GMO GPU Cloud” ranked 37th in the global TOP500 supercomputer ranking
(as of November 2024: https://group.gmo/news/article/9266/ )
(※5) “GMO GPU Cloud” ranked 34th worldwide and 1st in Japan in the global power-efficiency ranking “Green500”
(https://internet.gmo/en/news/article/50/ )
(※6) “GMO GPU Cloud” began offering a cloud service featuring “NVIDIA HGX B300” among the fastest in Japan
(https://internet.gmo/en/news/article/122/ )