TL;DR: If it were me, and I was spending $45k on a used robot with welder, I'd probably cough up another $10-15k for a brand new one. This is my opinion after being in the business for 20 years. If you want to read about why I state this, well, keep reading.
There's a lot to know about the world of used / refurbished robots that you should be aware of. A pretty paint job doesn't always mean a great robot. But even before that, there are things to consider about your project itself when deciding to go with new or used. The scale and budget of the project obviously has a lot to do with it, but so does the process and exposure or liability for downtime. If you're integrating a welding robot in a small fab shop, you have a few thousand parts lined up to weld and you want to get started welding with robots, by all means a used/refurbished robot is your jam. You'll get your feet wet with a smaller investment, and you'll most likely add more robots as you grow. We integrated one of our customers with custom cells using refurbished robots, and a few years later they bought almost 20x brand new dual arm full welding cells from us. That customer is probably the biggest reason we are now no longer refurbishing robots and focusing on integrated systems. So a few used robots can go places for many people if that tool is used correctly. On the other hand, if your project is many hundred thousand dollar robotic welding cell with auxiliary axis positioners, multi-arm, complex fixturing, and requires a large amount of integration, you really need to consider the risk / reward factor by putting a used/refurbished robot into that type of system. We have done many projects this way, and most were very successful. There were some that had service issues with the robots, and even more-so with the used welders. These were companies who were welding automotive parts and could not have downtime. It was very stressful for the customer obviously, but also for us as we had to resolve it immediately. Looking back, it would have been wise to have gone with brand new equipment on a project that critical. Also look at your ROI. If you're happier with a longer term, a new robot might not be a bad choice. Things have changed with used robots, some good, some bad.
Some of the good:
You can find them very inexpensive at auction if you can deal with the risk, buyers premium, rigging, and freight. I am talking sub $10k for late model robots. Also some of the older unit prices are starting to drop substantially, so if you can deal with a robot and control that's 2 generations old, there are great deals to be had.
Some suppliers have large inventories of late model robots at very reasonable prices. Some who would rather avoid the drama of an auction would have a wide selection to choose from and would have a guaranteed running robot when it arrived. It's up to you to do your research on the supplier of your choice. Some of the bad:
The later model welding robots are lighter, faster, have smaller gearboxes and larger payloads, and have longer grease service intervals. That doesn't equal a robot lasting as long as they used to. A lot of used welding robots are hammered, have a lot of wrist slop, and a lot of hours running at very fast speeds. Stay away from welding robots with more than 5k hours of move time in my opinion, the less the better honestly.
You may have to dish out many thousands of dollars for software and software option licensing depending on the manufacturer of that robot. Please do your homework here so you are not surprised later.
If you are looking to buy for example a Fanuc Arcmate 100iD with R30iB control and a robotic weld package. List on something like that might be only slightly higher for a brand new robot with factory warranty. If you look for the same arm on the used market, you may find one "refurbished" with a weld package for $45 to $50k... so, what are you really saving at that point? If you're short $10k and that's a line in the sand for you, that's fair enough, you'll have to go with used. Or if you can go back a few generations to an R30iA, you'll save money that way. But you're buying a 12 year old robot. Is that wise? Maybe. It depends on a lot of things. If you're going to a used robot supplier, have questions for them.
How long do you intend on using it?
How many hours are on it (actual run hours)? If they tell you "we don't know, it was init started" that usually means "it had a ton of hours but it's still pretty tight so we wiped it".
Is "refurbished" or "used" or "PM serviced"? Can your supplier define what that means? Company A might have a different definition of those vs Company B.
Is your supplier an authorized integrator or service outfit for the robot(s) they're selling?
Can your supplier provide you references?
Are they able to buy factory spare parts or are they buying used parts?
Can they supply a metrology level documented repeatability test?
Is there a service report?
Is the software licensed to your company?
What is the measured backlash on the gearboxes? Is it documented?
Does your supplier do run tests at load? Is it documented?
A lot of used welding robots have a ton of wrist slop - has that been checked? It shouldn't be ignored - it doesn't go away, it only gets worse. Find a tight manipulator, specifically check J5 on weld robots that are hollow wrist. Most robot mfg's do not sell replacement internals, a new wrist is typically $15k and up.
It really doesn't hurt to visit the supplier and see the robot for yourself. It's worth your time to go look at a few. This isn't much different than buying a car. Do you buy one without test driving multiple? If so, that's not too wise is it? Why would a robot be any different?
How was the robot greased? Does your supplier use a pneumatic pump? Those are prone to push seals, and hand pump grease guns are even worse. We have a utility patent on a de-coupled mechanical pneumatic greasing process that is used in our GreaseBots. The system is designed to never put out over 20psi to the gearbox. It's worth finding out how your robot is being serviced before you get it.
Does your supplier have a warranty? Can they provide a legal document showing their warranty? You have to read the fine print on warranties. You know this, you've seen them. They're all the same, this doesn't exclude robot warranties. Read it, understand what it says, and understand that they will not accept your terms down the road after you've accepted theirs. As they shouldn't. So don't skip this, talk to them about it, ask them if you have to ship the robot back - because some require it. Some don't. Some will do field service, some won't. This could be a major impact on your budget if you have to de-integrate a robot and control to have something fixed on it. So make sure you know what you're getting into there. There are some who found out later, probably best to know up front, eh?
Now when you look at used welding power supplies - completely different story. If you get a great deal on a robot with a welder, great. If you have an outfit who will let you visit and watch it weld in front of you before you buy it, even better. A one year warranty would be what I'd want, or no deal. You'll never get that with used though. In my opinion, refurbished welders are not "refurbished" as most say in the description. They may blow the fans out or replace broken ones, maybe reload firmware - but there's VERY little to refurbish on a welding power supply, or feeder for that matter. Please keep that in mind. In my opinion - if you can avoid a used welder, consider avoiding a used welder. I can't tell you the nightmare stories I've experienced in my time with used welders. I swore time and again I'd never sell one again, and I kept selling them and I kept getting dragged. I won't sell used welders anymore. Ever.
Used/Refurbished robots is an interesting market. Over the years it's been interesting watching how things change as the larger outfits all fight for the same saturated market. There are times when used/refurbished robots are in high demand - Covid being one of those times as supply chains were severely disrupted. But as a whole - the used robot market is mostly flooded with robots - mostly out of automotive companies. What I've noticed over the years is that some of the robots coming out today are actually late model, still in very good shape, and really don't even need to be refurbished. For a very long time, a lot of the robots that would hit the used market were beat, in very poor shape, and required quite a lot of actual refurbishment. That does not seem to be the case in what I've seen in the late model market. It's worth asking your supplier what their inspection / refurb process is at that point since most of the newer units don't really need much, and most of the time their factory paint jobs are fine.
In my opinion that business model is not at all different than a used car lot. How is it different? If you have a robot with 2,000 hours, it doesn't even need the grease changed yet. Maybe batteries and update software if it's required, but other than that?? So in my mind - if you see late model robots at auction and are comfortable with buying directly from the auction, you can save $20k most likely. The risk on a robot with that low of hours is very small. Make sure you inspect it in person, power it up if you can, move it around even better... Some of these later model R30iB and R30iB+ "refurbished" robots are selling for not a whole lot less than new. Keep that in mind as well. Overall - there are good deals to be had on used robots, stick with late model units, try to avoid high hour welding robots, and consider your options based on your budget and your project scope. There's going to be a point in your project where a used robot probably just doesn't make sense, and there are going to be some projects were a new robot doesn't make sense.