Nvidia will be the first company to achieve an annual revenue of $1 trillion just by selling AI equipment. This bold statement was made by Jensen Huang, Nvidia’s CEO, during the opening speech at the GTC 2026 conference held last week in San Jose, in the southern part of Silicon Valley.
To reach this goal, Nvidia is changing its strategy. Instead of being a chip supplier, it is transforming into an AI server manufacturer, operating under a white-label model. Previously, most of its clients were hyperscalers that used Nvidia’s GPUs to offer AI services online. Now, Nvidia plans to focus more on supporting private installations, which are in high demand.
By doing this, Nvidia aims to boost the business of data center equipment suppliers, who have been losing customers to the public cloud for years. However, Nvidia will no longer sell them chips to manufacture their own servers, storage racks, or network switches. Instead, Nvidia wants them to brand the pre-designed motherboards it produces.
But can the market truly abandon in-house engineering, which each brand uses to differentiate itself from competitors? One technical point of concern is that while Nvidia has always claimed to have designed the best semiconductors for AI operations, the most unexpected solution announced at this conference is based on Groq’s LPU chip.
Jensen Huang provided clarification on his strategy during a closed-door session. He emphasized that Nvidia’s GPUs have become dominant in the field of inference, especially with the introduction of NVLink-72 for direct networking of 72 GPUs and NVFP4 encoding in 4 bits. He claimed that Nvidia’s GPUs outperform x86 clusters significantly in terms of performance per watt, reducing energy costs by 35 times.
He also highlighted the collaboration between Nvidia’s GPUs and Groq’s LPUs for various AI models, stating that these partnerships will enhance the overall performance of AI operations.
Huang also pointed out that Nvidia’s revenue diversification strategy includes targeting 40% of its revenue from enterprises, private clouds, industrial sectors, and automotive manufacturers that prefer on-premise AI infrastructures over the public cloud. This shift reflects the increasing demand for private AI solutions globally.
In conclusion, Nvidia’s investments in new ventures like CoreWeave, Nscale, and Nubius are aimed at expanding its ecosystem and supporting agile partners for rapid growth. Despite some skepticism from observers, Nvidia’s approach seems to be centered on securing a stable revenue stream through hardware sales and dividends from recommercialization, potentially boosting its stock performance in the future.
