NVIDIA has become Naver's third-largest shareholder with a $1 billion investment. This move signifies more than just a supply contract; it highlights the global AI leader's choice of a key domestic IT partner.
Naver plans to use the funds to expedite the establishment of an AI factory and accelerate its entry into the global sovereign AI market, particularly in Asia. On the surface, this appears to be a significant achievement for K-AI on the global stage.
However, it is essential to confront the cold, hard reality. The market views this investment as a typical example of an 'AI bubble.' The structure involves NVIDIA, a global semiconductor giant, funding Naver, which in turn purchases NVIDIA's advanced graphics processing units (GPUs).
This raises concerns about 'circular investment' or 'self-dealing.' In Silicon Valley, warnings have emerged regarding the accounting illusions created when big tech companies invest in AI startups or cloud firms, only for that funding to return to their own AI chip purchases. If GPU demand slows or AI profitability falls short of expectations, it could lead to a chain reaction of risks.
It is unrealistic to paint an overly rosy future. NVIDIA aims to solidify its GPU ecosystem and enjoy long-term lock-in effects. This suggests that domestic companies like Naver may only serve as consumers or data center hubs.
Nonetheless, opportunities coexist. The substantial capital and collaboration with massive infrastructure should be thoroughly leveraged to enhance our independent 'technological strength' and 'capabilities.'
We must use the expanded GPU infrastructure as a foundation to significantly elevate South Korea's unique AI models and service capabilities. By building a sovereign AI ecosystem tailored to various industries such as finance, healthcare, gaming, and public services, we can secure practical 'technological sovereignty' that allows us to manage data independently and provide customized solutions optimized for different countries. Only then will we gain negotiating power with big tech.
To overcome the limitations of AI hardware that heavily relies on foreign sources, we need to strengthen connections with next-generation technologies like domestic AI semiconductors (NPU) and processing-in-memory (PIM). In the medium to long term, we should collaborate with domestic semiconductor startups and major memory semiconductor firms to develop an integrated ecosystem of AI hardware and software that dramatically reduces power consumption and costs.
We must also effectively manage research and development (R&D) collaborations and talent cultivation with global big tech. As NVIDIA plans to establish research hubs in partnership with domestic universities and research institutes, we should seize this opportunity to train a significant number of world-class AI talents and rapidly internalize advanced technological know-how.
The ongoing bubble controversy and concerns over excessive investment in the AI market will persist. However, in the face of the undeniable trend of AI transformation, whether we will leap forward riding the wave of liquidity or be left with only the shell of technological dependence after the bubble bursts ultimately depends on our capabilities.
It is time to clearly understand the dual nature of big tech's strategic investments and to strengthen our technological independence and global competitiveness, turning the risks associated with being at the center of the bubble into opportunities.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.