AI Infrastructure for Cloud

In this next installment of Hot Topics in AI Infrastructure, NVIDIA® and Pure Storage® discuss the role of cloud in artificial intelligence (AI) deployments. Should you build on-premises infrastructures for your AI needs or is it better to look to cloud-based AI?

We designed this video series to help educate IT specialists about how to get involved in AI initiatives. Each two-minute episode covers a single topic as Tony Paikeday and I tour NVIDIA’s Silicon Valley headquarters.

One reason AI adoption is growing is that cloud providers are offering simple AI cloud services that give you quick access to what can be unlimited compute to build and train models. This makes it very easy to get started with AI.

On the other hand, AI requires access to large amounts of data for effective training and results which may be in your data center. With the cloud, you might see escalating costs as datasets grow and cost of I/O and storage for this data increase with each experiment. This is one inflection point at which your developers might benefit from a fixed cost solution, making an on-premises AI solution the better choice.

How do you choose? Or should you consider a hybrid model in which you can experiment with AI in the cloud before building out on-premises capabilities? Take a look and see what is right for you.

Learn more about how to build the right IT platform that can support your AI initiatives. Register for our “Designing IT Infrastructure to Accelerate AI Innovation” webinar. We’ll see you back in two weeks for another episode. 

Want more video content from Pure Storage? Subscribe to Pure’s YouTube Channel. Stay current with the latest AI infrastructure resources and assets by following us on Facebook, Twitter, and LinkedIn