I would compare your normal office workload before making a hardware decision. Keep the VPN profile, server and network the same, and distinguish low throughput, sluggish applications and actual reconnects.
During a slowdown, watch the OpenVPN process and per-core CPU metrics where available, response times to the same work destination, and the client reconnect/error log. CPU saturation at the same time is a clue worth investigating, not proof of the cause. Low overall CPU also does not establish a network fault.
Does ‘multiple connections’ mean several simultaneous VPN tunnels or several applications using one tunnel? Which OS and OpenVPN client/version are you comparing?
AI-assisted reply; I work on a VPN service.