Nvidia’s transformation from a GPU provider to a platform provider
In the previous part of this series, we saw that Nvidia (NVDA) excels in its core product of gaming GPUs (graphics processing unit). For years, the company has been working to transform its graphics processors to general purpose processors by shifting to a platform model. This has helped it shift to new paradigms for its Data Center and Automotive sectors.
The transformation took place in three stages:
- virtualization of graphics and applications
- using computational science to make GPUs suitable for general scientific computing
- using data science to use GPUs for deep learning and AI (artificial intelligence)
These efforts started to materialize in fiscal 2017 as more cloud companies and supercomputers started using GPUs for their AI functions.
Nvidia’s data center revenue increased
As seen in the above chart, Nvidia’s (NVDA) data center revenue has increased threefold from $82 million in fiscal 3Q16 to $240 million in fiscal 3Q17, accounting for 12% of the company’s revenue. The segment witnessed strong demand in every area: AI, supercomputing, GRID, and virtualization.
Nvidia reported strong demand from cloud service providers such as Amazon (AMZN) Web Services, Microsoft (MSFT) Azure, and Alibaba (BABA) Cloud for their AI needs. Moreover, fiscal 3Q17 revenue also included sales of the DGX-1 AI supercomputer. Nvidia has shipped the DGX-1 to major universities, research groups, and multinational companies.
What’s new in Nvidia’s AI portfolio?
The growth in Nvidia’s (NVDA) Data Center segment is likely to continue in fiscal 4Q17 as more companies adopt AI. This growth would be further fueled by Nvidia’s newly launched Tesla P40 and P4 GPUs.
Nvidia has also launched TensorRT deep learning framework, which would accelerate the training process, as well as AI production workloads in hyperscale data centers.
Nvidia’s president, Jen-Hsun Huang, said that the benefit of GPU acceleration in the data center space is not just 20% or 50%—it is 20x–50x. The growth that Nvidia has started to report in its Data Center segment is just the start.
The company notes that it is likely to grow exponentially as the connected world leads to a data explosion. This dramatic increase would create the need for real-time analytics and AI in almost all fields, including finance, healthcare, retail, and transportation.
Nvidia has already started moving in this direction, with a variety of partnerships and next-generation technology. We’ll explore this initiative in the next part of the series.