Higher education institutions will need to have a certain AI footprint in the future, and for most, that footprint is going to require some computing upgrades. Universities doing research likely already need some type of high-performance, GPU-enabled computing, but that computing power may soon be necessary outside of that research. In addition, current higher education networks are built for moving data in a pre-AI world, and that’s not going to cut it going forward. Institutions will need more bandwidth and lower latency on their networks to make AI tools generate the outcomes they want, in a hurry. Network demands will be a barrier to AI implementation for institutions that don’t act soon.
