Aolani, a Singapore-founded neocloud powering AI growth, has announced a partnership to supply GPU cloud infrastructure to FriendliAI, the San Francisco-headquartered inference cloud for frontier AI. The move comes as demand for AI inference services surges globally, with organizations shifting from experimentation to full-scale deployment. Inference—the process of running trained AI models to generate outputs—has become a bottleneck for many companies, and access to reliable, high-performance compute is now a strategic differentiator.
FriendliAI, founded by researchers who invented continuous batching, a technique now standard across AI inference serving, has built its inference stack end to end, from optimized GPU kernels to global distribution. The company consistently ranks as one of the fastest inference providers on OpenRouter, with enterprise clients including LG, Kilo Code, and Liner running production workloads on its platform. However, exponential growth in demand has created a pressing need for additional compute capacity to keep services responsive.
Aolani, one of the leading neoclouds offering purpose-built next-generation AI infrastructure, will provide that capacity. Its capabilities across orchestration, automation, and lifecycle management are designed to help AI natives scale efficiently. The partnership equips FriendliAI with the compute to serve rapid customer demand both globally and increasingly in Asia.
"We're seeing inference needs grow faster than companies can find compute to support and service their customers," said Nicholas Chia, Chief Executive Officer at Aolani. "To narrow the supply and demand gap, we actively partner with companies like FriendliAI to deliver compute capacity on time, at scale, and to rigorous standards."
Byung-Gon Chun, Founder and CEO of FriendliAI, added: "We are seeing exponential growth in demand for our frontier AI inference services. Businesses need the freedom to choose the AI models that best suit their applications and the ability to run them efficiently in production. Aolani stood out as a trusted infrastructure partner that can help us scale at the pace our customers need."
The partnership underscores a broader trend: as AI adoption accelerates, reliable compute infrastructure is becoming as critical as the models themselves. More AI natives are turning to Asia for high-performance compute capacity, attracted by expanding digital infrastructure, strategic connectivity, and a growing AI ecosystem. Aolani's Singapore base positions it at the center of this shift, while FriendliAI's San Francisco and Seoul presence bridges Western and Asian markets.
For enterprises deploying AI at scale, the implications are significant. Dependable inference capacity directly affects user experience, operational costs, and the ability to innovate rapidly. By securing dedicated GPU cloud resources, FriendliAI can continue to offer the speed and reliability that agentic AI workloads demand, from real-time chatbots to complex autonomous agents.
The collaboration also highlights the growing importance of specialized neoclouds that cater specifically to AI workloads, rather than general-purpose cloud providers. As the inference market expands, such partnerships may become essential for keeping pace with demand and ensuring that developers can focus on building applications rather than managing infrastructure.
For more information, visit https://www.aolanicloud.com/ or https://friendli.ai/.
