I don’t have specific projects but maybe some ideas of where to start looking. This is a list of some advantages I see SAFE having over traditional server approaches.
Data as a cost - where the data is the primary reason for the service so forms a large portion of the cost of the service - content delivery networks, backup services, mirrors - SAFE should be cheaper.
Data as an obligation - where the data is not the primary purpose of the service but the service cannot function without supplying data - software updates, package managers - SAFE should be simpler to manage
Data as a requirement - some services require data in specific forms or within certain boundaries and the predefined structure of the network may assist meeting those needs - legal requirements such as tax records - SAFE should be easier to demonstrate meeting as these requirements
Data as a liability - where the data and the source must be isolated to prevent risks to the source - eg whistleblowers, piracy metadata (most piracy data is p2p already) - SAFE should increase anonymity and reduce operational risks
Data as a historical record - most of the internet is browsed in the most up-to-date form and records of changes or past versions are discarded - archives, libraries, media and journalism, knowledge graphs, encyclopedias - SAFE should reduce the complexity of managing history
Data as a public good - when private companies own public data it may put that data at risks which can be mitigated by using SAFE network - government services, maps, satelite data, public data sets, community sensor networks like weather stations - SAFE is owned by the public and is less prone to the whims and faults of private services being used for public goods.
Is the plan to replace an existing service with a new one on SAFE, or to augment the existing service to use SAFE ‘behind the scenes’?
What problem does the company have and how does SAFE specifically help with that problem? Has that problem statement come from the company, or has it been assumed by an outsider? Will the solution actually end up being adopted?
How does adding SAFE / moving to SAFE potentially cause problems or additional costs? Why would they not move to SAFE and how can these issues be addressed?
IMO data as a historical record is probably the strongest place to start a hackathon project due to the permanence of SAFE being such a unique offering.