Recently, Nvidia (NVDA) has consistently remained a focal point in the market. The company’s stock performance significantly outpaced the broader market between 2023 and 2025. Looking ahead to 2026, the next stage of development in the artificial intelligence field may bring an even more profound impact. At the CES exhibition, Nvidia officially launched its new-generation Rubin AI platform, marking the establishment of an annual update cadence in its AI chip segment. The platform integrates six new types of chips, achieving notable improvements in inference costs and training efficiency. The first batch is expected to be delivered to customers in the second half of 2026. Several major cloud service providers, including Microsoft (MSFT), will be among the first to deploy it. Microsoft’s next-generation Fairwater AI super factory plans to configure NVIDIA Vera Rubin NVL72 systems, scalable to hundreds of thousands of super chips. Meanwhile, CoreWeave (CRWV) will also serve as an early supplier of Rubin systems. This progress addresses some market concerns regarding the sustainability of AI spending and intensifying competition. Nvidia remains optimistic, expecting the relevant market size to reach trillions of dollars in the long term.
Nvidia’s third-quarter results showed that its data center GPUs were “sold out,” despite the quarter’s total revenue reaching $57 billion, with the data center segment contributing $51.2 billion. To meet persistently strong demand, the company is comprehensively expanding production capacity, including coordinating with upstream supply chains and adjusting internal production lines. According to reports, Nvidia has begun cutting production capacity for certain product lines, such as gaming chips, to prioritize the production of higher-margin and more urgently demanded data center GPUs. While the gaming business brought in $4.3 billion in revenue in the third quarter, reallocating capacity helps alleviate supply pressure for AI chips. Nevertheless, due to substantial customer demand, fully meeting all orders remains challenging. This also means that when other AI operators turn to proprietary chips or competitors (such as Advanced Micro Devices), it is not necessarily due to a lack of competitiveness in Nvidia’s products but rather stems more from its own supply constraints. If this situation persists, it may help Nvidia maintain its product pricing advantages and profit margins.
In April 2025, the export license for Nvidia’s H2O chips specifically designed for the Chinese market was suspended, impacting the company’s business. Although related sales were once projected to reach $8 billion, Nvidia recently obtained approval to resume sales, albeit with an additional 25% fee. Regardless of which party bears this cost, Nvidia will re-enter this important market, which is expected to drive revenue growth in 2026.
Nvidia anticipates that global data center capital expenditure could grow from approximately $600 billion in 2025 to $3-4 trillion by 2030. If this prediction materializes, Nvidia stands to benefit continuously from this process. Through multiple measures, including capacity optimization, new product launches, and market expansion, the company is fully committed to seizing the opportunities presented by the new stage of AI development. These strategies collectively outline a clear path for Nvidia’s future growth and lay a crucial foundation for its performance in 2026 and beyond.
