WLCG
Offered resources not matched to requirements (e.g. HPC, GPS, etc)
Requirement of some FA's to provide HPC or GPU as part of the pledge may lead to resources matched to requirements, or could mean that some resources can only be used for some workflows.
Some architectures are not well adapted to the production workflows of the experiments, or may be very inefficient for some, or may be only suitable for specific types of work. In some countries we may be required to accept HPC systems as part of the pledge, whcih would be potentially inefficienct for general workloads, including those requiring signicficant data flows.
As with other technology related risks, the mitigation of to ensure that the experiments' core software is as portable and adaptable as possible, so that they can run reasonably efficiently on all potential srchitectures. Failure to do this and to be able to well use new architectures - some of which will be mainstream - will lead to the need to reduce the overall scope of physics, or to prioritse it.
Similar conclusions to other technology and funding related risks. The real mitigation is adequate (& significant!) investment in software skills, in code re-engineering to modern architectures, and to build a new skills base in the HEP community.