Just keep in mind that if they don't schedule hardware replacement, the hardware will eventually schedule it for them as a fun surprise.
From a realistic standpoint, if the processors, RAM, and network cards haven't caused problems by now, they probably won't. The fans, spinning drives, and power supplies are the things which become faulty over time. Of those, the drives can be replaced with off-the-shelf units should the need arise. Just pull the old drive out of the rail, slap a new one in, insert it, and probe the SATA endpoints to force the new drive to be recognized.
I would feel fine running old hardware (e.g, to avoid spending money and time on a datacenter being decommissioned) as long as I had some spare fans and power supplies. If you don't have some spares on hand, get some now so you don't have to panic-buy them later after a failure. Save them from other decoms. Find some on eBay. Whatever. Once you have spare parts, test them. Plan a testing window, pull a part out of the standby member of the cluster, and swap in a spare. Make sure it works.
If you're okay taking on the responsibility of dealing with failed hardware, that just leaves the software. The 12k series refuses to install R81 or newer. This is ultimately another reason I can't stand Check Point's branded boxes. I can upgrade a much older IBM x3650 to R81.10 with no problem. Still, R80.40 can get software support for another year (until January 2024).