When you start out with things like this, you immediately disqualify anything that follows. PC's, the VAST majority of the time, are operating at "room" temps.
You don't get it, do you?
When the thermal paste is applied on a cold CPU, yes, both are at room temperature. Within seconds of turning on the computer, the CPU is already above room temperature. That heat transfers to the IHS as well as the exterior thermal solution (pad, grease, heatsink, waterblock, maybe a combination of many).
Arctic MX-5 does not remain at room temperature on a CPU in operation.
Go ahead and fire up a typical contemporary desktop PC with a bare CPU. I bet you the computer tries to shut itself down in a minute or two. And please feel free to touch the bare CPU just before it does. Leave your finger on there for a few seconds. I'm sure the sensation will be one that you richly deserve.
Then take your finger off the CPU and use it to operate a laser thermometer. Point it at the ready-to-shutdown CPU and measure the temperature. That would be the temperature of a thermal grease applied to that surface.
And for the record, oil "seperation" doesn't happen until well above 150C for most substances that would be used in a PC/technology application. It's a chemistry thing.
Unfortunately, Arctic selected an oil for MX-5 that does separate below 150C. I'm sure they didn't mean to but the crux of the problem is they did. Arctic admitted it themselves and they deliberately put the product into EOL status.
I didn't tell Arctic they made a bad product.
They told us they made a bad product.
And it was probably mostly a financial decision. What is a tube of MX-5 at retail? About $8? What are COGS? Let's say $5. So Arctic has a gross margin of $3 per tube (let's remove wholesale distribution out of the equation). But every time there's an MX-5 RMA, the administrative cost (replacement product, staff salaries, shipping and logistics) to send out a replacement tube is probably around $18-20. So all of a sudden, Arctic needs to sell 6-7 more tubes of MX-5 to cover the expense of just one RMA.
Arctic probably has a good idea of the failure rate of MX-5 now based on more extended testing in their own labs and their RMA volume. Continuing to sell MX-5 likely pointed to continued expense and headaches for Arctic so they killed it off. Themselves.
The fact that you have not seen problems with your own tubes of MX-5 is irrelevant. Arctic acknowledged the problem and decided that killing of their own product was the best way to handle it.
Unfortunately Chomiq's test was useless. I can spread a paste with oil in it at room temperature and not see any oil separation. But if that paste changes temperatures, it can change and sometimes this leads to oil separation. If you ever spend any time in the kitchen, you'd know this.
Chomiq could have spread peanut butter, Nutella or mayonnaise on the CPU and not see fat separation. Great. That doesn't mean fat separation can't occur at a different temperature.
Could we expect to see oil separation from a CPU in operation using Nutella as thermal grease? I'm sure most people here would probably say yes. Could we expect to see it from a CPU using MX-5 as thermal grease? Most people at the start of this thread (March 2021) would have said no, but over time, more and more people did. In fact, so did Arctic eventually.
Maybe you're frustrated that Arctic killed off a thermal grease that you liked. Well Arctic liked it too (and the revenue it generated) until the RMAs started piling up.
Ultimately Arctic did the right thing: take a defective product off the market.
My guess is some day Arctic will launch MX-6 but I bet they will test for oil separation very, Very, VERY thoroughly before they do.
