SEOUL, South Korea / AMSTERDAM, Netherlands  – FriendlyAI, an AI inference service specialist, announced a strategic partnership with Nebius, a global AI cloud infrastructure company and a major Microsoft partner. This collaboration is set to significantly accelerate FriendlyAI’s global expansion by integrating its cutting-edge inference acceleration technology into Nebius's large-scale AI cloud infrastructure.

The core of the partnership involves supplying FriendlyAI’s proprietary optimization technology to Nebius’s AI cloud platform. This integration promises to deliver a superior inference environment for Nebius's clients—which include companies operating customer support chatbots, coding assistants, and AI agent services—by offering faster speeds, greater cost-efficiency, and enhanced stability through FriendlyAI's API.

FriendlyAI’s distinct technology is lauded for providing world-class AI inference acceleration and optimization, capable of reducing the most significant burden on companies—GPU costs—by up to 90%. Furthermore, its unique techniques, such as model infrastructure-level optimization and Continuous Batching, enable inference speeds that are more than two times faster. A guaranteed uptime (SLA) of 99.99% is another key selling point, highly valued in the current market where stable operation of hyper-scale AI infrastructure is paramount. FriendlyAI also supports over 460,000 Hugging Face models, streamlining the entire product lifecycle from prototype to large-scale deployment.

Nebius, headquartered in Amsterdam and listed on Nasdaq (NBIS), is a neo-cloud company providing a full-stack infrastructure for the AI industry across Europe, North America, and Israel. The company recently solidified its position as a key global AI infrastructure provider by securing an AI computing partnership with Microsoft valued up to $19.4 billion.

"Our goal is to make world-class AI inference technology accessible to every company," said Byung-gon Jeon, CEO of FriendlyAI. "The combination of FriendlyAI’s inference optimization technology with Nebius's AI Cloud means all customers can now deploy AI models with the highest levels of latency, stability, and cost-efficiency."

The cooperation marks a significant milestone in FriendlyAI's global strategy, which has been underway since the beginning of the year. Prior achievements include being the first Korean company to supply an AI model deployment option to Hugging Face in January. Following the appointment of COO Sang-won Lee—who has a Silicon Valley exit under his belt—in May, the company has actively promoted its technological superiority at global AI conferences.

FriendlyAI’s growing prominence has attracted major domestic and international partnerships and investments. It is an official deployment partner for LG AI Research's latest model, Exaone 4.0, and a key participant in the Korean government’s K-AI Model project. Most notably, in August, the company secured a rare $20 million seed extension investment round with participation from prominent Silicon Valley VCs, reaffirming its growth potential in the global market.