At a recent industry exchange event, NVIDIA CEO Jensen Huang sent a clear signal to investors: despite the company's quarterly revenue approaching the $100 billion mark, its growth momentum has not slowed down but is instead accelerating. In response to external concerns about increased competition from ASICs (application-specific integrated circuits) and product roadmaps, management provided an active response, reiterating the company's core competitiveness in the AI computing market.

Regarding previous market rumors that the next-generation flagship architecture, Rubin Ultra, would be delayed until 2028, Huang explicitly denied this. He emphasized that Rubin Ultra will still follow the planned schedule and ship next year. Although some rack designs are undergoing architectural optimization to meet the demands of large-scale computing clusters—such as introducing new designs to support 800V high-voltage power supply and rack-to-rack optical interconnect technology—this is part of normal R&D iteration and does not change the overall release schedule.

The shift in customer structure became the focus of this exchange. According to disclosure, a cutting-edge AI model project that previously heavily relied on self-developed ASIC chips now uses nearly 50% of NVIDIA GPU computing power. This shift effectively counters claims that NVIDIA is losing its competitive edge and also confirms the significant advantages of NVIDIA's full-stack solution in "Token cost control" for both training and inference scenarios.

In terms of business growth engines, NVIDIA is focusing on diversification. In addition to traditional hyperscale cloud service providers, AI labs, sovereign AI, and enterprise customers are becoming new growth drivers. The company clearly stated that as governments around the world place greater emphasis on local computing infrastructure, these emerging markets tend to adopt NVIDIA's highly integrated platform solutions.

Additionally, NVIDIA is accelerating its penetration into the general-purpose server market. The revenue from the CPU business is expected to reach $20 billion this fiscal year, and the next-generation Vera CPU is positioned beyond traditional management nodes, directly entering the broader general-purpose server computing market.

As its market value steadily approaches $5 trillion, NVIDIA's capital strategy is beginning to align with value investment logic. With strong free cash flow, the company is expected to increase stock buybacks in the future to reward shareholders. Morgan Stanley analysis points out that the biggest challenge currently facing NVIDIA is no longer the sustainability of demand, but how to further improve delivery efficiency under physical resource constraints such as power, network, and cooling, transforming massive orders into high-quality financial growth.