TBH there are few projects with the same concept - decentralized computing power. But what I think is different here in Kumo is the pay per execution model. I could be wrong tho, haven't checked all the other whitepapers.
This is a question we've had asked a lot and it's a good one - what makes Kumo different from the other projects out there? We were recently asked this in our discord chat; the TLDR comes down to Kumo's focus on latency, full cloud platform functionality, and a purpose built blockchain for running applications. The longer answer from our chat is here:
Q: What is the difference between kumo and the other alternatives out right now? Golem, iexec and the elastic project.
A: With networks such as Golem and similar the focus is creating a "global supercomputer" which is great for tasks such as image rendering or specific tasks within applications that are not time sensitive/interactive such as batch processing.
With Kumo we are focused on tasks that have time sensitivity - ie the tasks that applications have to complete quickly as they involve a user feedback loop or are part of a sever/container/task workflow where tolerances or internal/external SLA's are measured in milliseconds vs minutes (or hours/days as in image rendering).
Essentially Kumo is focused on building the network, and distributed cloud platform, to run entire applications in real time in direct performance competition with major IaaS providers and allow developers to easily migrate existing jobs/tasks/portions of their applications (or entire applications) onto the Kumo network as they would between existing cloud providers such as AWS or Azure.
iexec would be most similar however the focus there (at least from what I understand currently) is to run on Ethereum to route tasks specific to dapps to available compute resources vs serving existing commercial/industrial tasks or applications in general that would run on a traditional cloud deployment.