AI Firm TEN Launches New Software to Combat Soaring GPU Costs

As the demand for artificial intelligence surges, the cost of GPUs, the essential 'brains' for AI systems, has skyrocketed. This escalating expense presents a significant hurdle for companies eager to develop AI services but hesitant due to high equipment and operational costs. Addressing this challenge, TEN, an AI infrastructure operation specialist led by CEO Oh Se-jin, has unveiled new technologies designed to drastically cut AI operational expenses. At their '2026 Partners Day' event, TEN introduced two new iterations of their flagship AI infrastructure management program, AIPub: 'AIPub Enterprise' and 'AIPub Lite'. These products are built upon the world's first 'GPU dynamic allocation technology'. This innovative system acts as an intelligent navigator, continuously monitoring a company's AI semiconductor resources. It identifies idle hardware and automatically reallocates it to needed tasks, then reclaims it promptly for other projects once completed. This minimizes the time expensive equipment sits unused, bringing wasted resource time close to zero. Furthermore, it enables multiple departments or projects to efficiently share a single piece of hardware without performance degradation. 'AIPub Lite' is specifically tailored for emerging AI agents, offering essential features at a lower cost, making AI development more accessible for small and medium-sized businesses and startups. TEN emphasizes a collaborative approach, viewing AIPub as a platform built with their partners. The company is committed to advancing its technology to provide tangible cost-saving value to businesses grappling with the high cost of AI operations.

“

AIPub is a platform we build together with our partners. Based on a strong cooperative system, we will continue to advance our technology to provide practical cost-saving value to companies struggling with soaring AI operating costs.