Moreh Unveils Distributed Inference System on AMD at AI Infra Summit 2025
Moreh and SGLang Collaborate at AI Infra Summit 2025
In a significant showcase at the AI Infra Summit 2025, Moreh, a leading AI infrastructure software company, introduced its latest distributed inference system optimized for AMD architecture. This unveiling occurred in Santa Clara from September 9 to 11, where the event brought together over 3,500 attendees and more than 100 partners from around the globe, highlighting the critical advancements in AI technology.
The AI Infra Summit, known for being one of the most comprehensive conferences on AI infrastructure, has evolved from its inception as the AI Hardware Summit in 2018. This year’s agenda was tailored for hardware manufacturers, hyperscalers, and enterprise IT professionals seeking cost-effective and efficient AI implementations.
During the Enterprise AI session on September 10, Moreh's CEO, Gangwon Jo, delivered an impressive presentation, revealing benchmark results that underscored the system's enhanced efficiency in optimizing leading-edge deep learning models, including DeepSeek, compared to offerings from established competitors like NVIDIA. The newly introduced system signifies a major leap in performance and affordability for companies looking to implement AI solutions.
Moreover, the partnership with Tenstorrent was spotlighted, showcasing a next-generation AI semiconductor system that merges Moreh’s software prowess with Tenstorrent’s innovative hardware. This collaboration aims to provide businesses with viable alternatives to NVIDIA, renowned for its stronghold in the AI chip market.
As part of the event, Moreh also co-hosted a presentation with SGLang, a prominent player within the deep learning inference software ecosystem. The alliance is strategic, focusing on tapping into the burgeoning North American market for AI technology. By organizing a booth and networking sessions together, Moreh and SGLang aim to leverage their combined strengths to foster deeper relationships within the global AI community.
Looking forward, both companies revealed plans to jointly develop an AMD-based distributed inference system, reinforcing their commitment to expanding their reach in the rapidly growing market for deep learning inference.
Gangwon Jo expressed his confidence in Moreh’s capabilities, stating, “Moreh possesses the strongest technical expertise among AMD’s global software partners and is currently conducting proof-of-concept projects with several leading LLM companies.” This statement reflects Moreh’s ambition to establish itself as a key global player, providing customers with diverse options in AI computing.
Additionally, Moreh is in the process of developing a core AI infrastructure engine. Through its subsidiary, Motif Technologies, the company aims to secure comprehensive technological capabilities that span various domains, positioning itself as a competitive force in the global market. Collaborations with major entities like AMD, Tenstorrent, and SGLang are crucial in achieving these goals and expanding Moreh’s innovative offerings.
This summit not only marks a pivotal moment for Moreh but also represents a significant milestone for the AI infrastructure industry as a whole. With a focus on building partnerships and advancing technology, the future looks promising for the integration of advanced AI solutions in business logistics, enhancing operational efficiency, and staying competitive in an increasingly data-driven world.
Overall, the AI Infra Summit 2025 served as an incredible platform for Moreh and SGLang to reveal their groundbreaking advancements in distributed inference systems, setting a roadmap for future initiatives and collaborations in the AI landscape.