NVIDIA’s AI platform
With Titan V, NVIDIA (NVDA) has launched a complete suite of AI (artificial intelligence) products that cater to all types of consumers. The suite includes the following:
- Tesla V100 for data centers
- DGX for HPC (high-performance computing)
- GPU (graphics processing unit) cloud services that rent GPU computing capability in the cloud
- NGC (NVIDIA GPU Cloud) services, which optimize deep-learning frameworks across its GPU cloud service, the TensorRT platform, for inference
- The Titan V, which bring AI (artificial intelligence) computing capabilities to PCs (personal computer)
NVIDIA launched its Tesla V100 at the end of 2Q17, and in just one quarter, it has seen widespread adoption across the major internet, cloud services, and server companies like Google (GOOG), Facebook (FB), Amazon.com (AMZN), Microsoft (MSFT), and IBM (IBM). Chinese giants like Alibaba, Baidu, and Tencent are also adopting Volta GPUs.
The growing adoption of the V100 saw NVIDIA’s data center revenues rise 20% sequentially to $500 million in fiscal 3Q18. NVIDIA believes that AI and data center to be a $30-billion market opportunity, and it’s only drawing $2 billion from this market so far, leaving plenty of untapped opportunities.
Notably, in October 2017, Jefferies analyst Mark Lipacis predicted that Volta would help NVIDIA command 80% of the AI market in the long term.
NVIDIA should be wary of competition
NVIDIA is currently dominating the AI space and is charging a premium price for its GPUs. It prices in the huge cost savings that the GPUs bring to customers, in addition to its efforts in designing the GPU and its efforts in writing the CUDA program.
But the rising prices of NVIDIA’s GPUs are fuelling innovation and encouraging companies to build cheaper alternatives. For instance, Google has developed tensor processing units to handle its AI workloads.
Many analysts have warned that several companies are investing heavily in AI technology, but NVIDIA will likely be cautious of these competitors—particularly of Intel (INTC).
Next, we’ll discuss growing competition in the AI space.