We surveyed around 50 Data Scientists and IT Decision Makers to understand how their teams obtain the necessary compute for workloads and the consideration set used to make these types of decisions. 

Here are the key highlights: 

• Use of a public cloud is the most popular way to obtain the compute needed to manage AI workloads, though roughly a third of Data Scientists and ITDMs use a combination of sources. 
• Often time, teams consider the type of work they’re supporting when making decisions about which sources to rely on. Depending on the size of the workload or sensitivity of the data, they may choose one option may over another. Data Scientists and ITDMs admit that cost, flexibility, scalability, security, and performance are important factors.

What do you think of the results and how does your team secure the compute resources needed to handle AI workloads?