Agentic AI: Korea's State-Backed Industrial Strategy Needed

The era of agentic AI demands a state-backed industrial strategy. This new phase of AI means agents can independently create plans, make decisions, and complete complex tasks with minimal human intervention. This is a shift from earlier AI that mostly just answered human questions. Wall Street recently saw a company called Cerebras achieve a $56.4 billion valuation on its public debut, with its stock jumping 68 percent. This company, an AI inference chip startup, was valued at $8.1 billion just eight months prior. Investors see this as a sign of the next phase of AI. Agentic AI systems, like coding agents, are already being used in industrial settings. These systems consume tens of thousands of tokens, the basic units AI uses to process language, without pause. This means processing speed is critical. Cerebras has partnered with OpenAI to build large-scale inference infrastructure to handle these demands. The era of a single general-purpose chip for all AI tasks is fading. Instead, specialized architectures optimized for specific workloads are becoming key. The "token" is also emerging as a new economic unit. The competitiveness of an AI service now depends on how many tokens it can process, how quickly, and at what cost. This highlights a fundamental change in the AI economy.