Nvidia on Thursday announced the launch of its latest graphics card for artificial intelligence.
CEO Jensen Huang showed off the Titan V graphics processing unit for desktop PCs, which employs Nvidia's seventh-generation Volta architecture, for the first time at a major AI conference in Long Beach, Calif.
The new GPU delivers 110 teraflops of performance, or nine times that of the comparative model from Nvidia's previous architecture, the Titan Xp, which was announced in April. But like Nvidia's Tesla V100 GPU for data centers, which was unveiled in May, the new Titan V is more clearly oriented toward AI than previous generations.
The news comes amid a rush to deliver ever more powerful computing resaurces for AI. In the past few years Nvidia's GPUs have become widely used for a trendy type of AI called deep learning. Intel recently announced plans to develop its own GPUs that will compete with Nvidia, but, at the same time, Intel has recently begun shipping Nervana chips tailor-made for AI.
The new Nvidia GPUs are meant to go inside people's PCs, rather than in data centers. But researchers, scientists and engineers can take advantage of GPUs from cloud infrastructure providers like Amazon, Google, IBM and Microsoft. Google has revealed tensor processing units that can be used in the place of GPUs — but they will only be accessible from Google's cloud, unlike the new Titan V.
Nvidia stock has risen 101 percent in the past year, according to FactSet.