VesselAI Partners with US PadoAI... AI Computation Automatically Adjusted to Power Grid Conditions

Power Challenges Become a Key Issue in AI Data Center Operations
Optimizing and Allocating AI Computational Resources Based on Energy Conditions
"Reducing Costs While Maintaining AI Service Quality"

VesselAI, an artificial intelligence (AI) infrastructure company, announced on January 20 that it has entered into a partnership with PadoAI, a US-based data center energy orchestration platform provider. The two companies plan to jointly develop the industry's first "grid-aware MLOps (Machine Learning Operations) solution," which will operate AI workloads by reflecting real-time power grid and energy conditions.

This collaboration aims to efficiently manage the rapidly increasing demand for AI computation by combining PadoAI's energy intelligence with VesselAI's AI infrastructure orchestration technology.

Last year, global investment in data centers reached a record high of $61 billion (approximately 80 trillion won), and the adoption rate of AI by enterprises surged by 55%. As a result, the costs associated with power consumption and the burden on the power grid have emerged as key challenges in data center operations.

The "grid-aware MLOps solution" introduced by the two companies is characterized by optimizing and allocating AI computational resources according to energy conditions. PadoAI provides technology that determines when and where AI computation is most efficient, based on power grid signals, energy prices, and the availability of renewable energy.

VesselAI is responsible for executing, migrating, and resuming actual AI tasks across GPUs and cloud environments, based on these determinations. Through this, the companies plan to implement an operational system in which AI computation is automatically adjusted to match power and energy conditions.

Once this technology is applied, AI computation can be automatically distributed to times and locations where electricity costs are lower or renewable energy is abundant. Data centers and enterprises can maintain the quality of their AI services while reducing operational costs. In particular, data center operators can maximize infrastructure efficiency by utilizing idle resources as power grid assets, thereby creating new revenue streams.

Jae-man Ahn, CEO of VesselAI, stated, "AI has now moved beyond the experimental phase to become an always-on infrastructure. By combining PadoAI's power grid technology with VesselAI's orchestration technology, data centers and enterprises will be able to operate their AI infrastructure much more efficiently and reliably."