Can AI data centers be mass-produced like cars? This startup aims to find out
techcrunch ·
Runware, an AI infrastructure startup, is aiming to mass-produce modular data centers similar to automobile manufacturing. The company plans to scale its Romania-based factory to produce 30 megawatts of capacity daily by 2028, leveraging an assembly line approach to reduce build times from weeks to a single day. Unlike competitors who design infrastructure for others, Runware has built its own production facilities and aims to have pods online within 48 hours of installation. Co-founder Flaviu Fadulescu emphasizes that the company uses standard components and in-house hardware teams to adapt quickly to new GPU generations and avoid high costs. While acknowledging challenges in supply chain logistics and execution, Runware cites $1 billion in committed interest for its Sonic Inference Pods. The modular units utilize closed-loop liquid cooling to minimize water usage and are designed to operate silently on smaller sites with existing power connections. Runware intends to expand this model globally, positioning factory-built pods as the future standard for rapid AI deployment.
techcrunch