Nvidia GPU Technology Conference

Background

At this year’s GTC conference, NVIDIA once again emphasized that there is no AI bubble, asserting that the construction of AI infrastructure is progressing vigorously while the company actively builds a U.S.-based manufacturing supply chain. Jensen Huang expressed optimism that over the next five quarters, NVIDIA’s chip revenue could reach as high as USD 500 billion, driven by its Blackwell and Vera Rubin series.

Huang has also been aggressively expanding into AI applications:

First, NVIDIA has partnered with Nokia, announcing a USD 1 billion investment to jointly develop next-generation 6G technologies. The collaboration integrates NVIDIA’s AI chips and infrastructure into Nokia’s data center technology, allowing NVIDIA to extend into the telecommunications sector and strengthen critical national infrastructure.

Second, NVIDIA is collaborating with Foxconn, Stellantis, and Uber to advance the deployment of AI-powered autonomous taxis.

Commentary

We believe this differs from the dot-com bubble of the past in that AI’s growth extends to tangible, real-world applications, rather than being driven purely by speculative investment. According to Jensen Huang, today’s AI models are powerful enough to deliver real value, making users willing to pay for their use. This creates exponential industry growth, as model training continuously improves accuracy and utility—further reinforcing user adoption and fueling demand for computing power in a positive feedback loop.

Cloud service providers (CSPs) continue to show strong demand for AI data centers and computing power. Recently, Microsoft, Google, and Meta all raised their capital expenditure forecasts for 2026, signaling ongoing robust growth. Given current trends, the surge in AI applications and demand is expected to drive greater output in AI manufacturing and infrastructure, positioning electronics production and AI applications as the core engines of national economic growth.