Several principles to keep an eye on here:
Bulk powder, like the military and most manufacturers buy, can last a very long time. The military puts stockpile age limits for munitions loaded with double-base powders at 20 years and single-base at 45 years. In good conditions, those are conservative life expectancies for new powder in bulk grade.
Alliant inherited a sample of the first ever lot of Unique made, IIRC, in 1913. They keep it stored under water and periodically dry out and test a sample, and thus far, it is still good. I have communicated with someone who fired some M1 Ball ammunition made in the late 1920's and found it still worked.
On the other hand, you can find several examples online, of M1 Garands damaged by old ammunition. In one instance, 1947 manufactured M2 Ball ammunition burst the receiver. I, personally, bought some .308 Winchester ammunition in 1993 that was made by Sellier & Bellot in 1982 and that was already seriously deteriorated, causing squibs that wouldn't cycle my M1A and, due to the acid radicals in the deteriorating powder, caused severe bore after-rust, even worse than corrosive primers cause.
In bad conditions (high heat of 140°F) the Naval Ordnance folks have been able to seriously deteriorate double-base spherical propellant in about 18 months, resulting in a large increase in pressure because the deterrent compounds in the powder had deteriorated faster than the nitrocellulose and nitroglycerin did, leaving the remaining powder with a much faster burn rate than it had originally. That is undoubtedly what blew up the Garands damaged by old ammunition. It is not uncommon for folks using surplus ammunition to find it seems awfully hot. When that occurs, powder at end of life is to be suspected.
The reaction responsible for breaking down stabilizer and making the powder go bad will double in rate roughly every 10°C (18°F) increase in temperature. So keeping the powder in the deep freeze can make it last centuries.
But all the above is more problematic for the handloader. Handloaders are not sold bulk grade powder. They are sold canister grade powder. This is because the burn rate in bulk powder varies too much from lot-to-lot to be reliably safe with published load data. To make published load data work, burn rate must be more tightly controlled. Canister grade powder is the more tightly controlled type.
To make canister grade powder, the manufacturer checks the burn rates of every bulk lot of each type of powder he makes. When a lot comes out exceptionally fast or exceptionally slow, he sets it aside. Later, when he gets a new bulk lot of that same powder type that is a little too slow burning, he blends some of the setaside extra fast lot with it to adjust the net burn rate close to where it should be ideally. If he gets a lot that is a little too fast, he blends some of the setaside extra slow lot with it to adjust the burn rate up closer to its nominal value. This has to be done with fast or slow lots of the same type of powder or the combustion curve characteristics will be thrown off, altering the progressive performance of the powder.
That's all great for making your load manual load data close enough to be useful, but it has a drawback. Instead of the powder being 100% fresh, it is now a mix of older powder and new powder. Since acid radicals from the breakdown of old powder can attack new powder, this means the life expectancy of your canister grade powder is only as good as the old powder blended in with it. Since powders gradually consume the stabilizer, the old powder's life will be shorter than the new.
So, now we come to the $65K question, how long will your powder be good for? Who knows? Norma's manual says to expect they will last at least 10 years in normal conditions (not in your freezer and not sitting in the trunk of your car all summer long). The powder makers really should test the diphenylamine content of their lots held back for blending before they mix it with a new lot. If it is too near the end, they shouldn't use it. I think failure to take that step is why you get occassional callback and warnings. IMR had a
warning about several lots of IMR4007SC because of early deterioration. Around 2000 there was also a lot of IMR4350 that a number of folks had going bad in just two or three years. A few years ago, Internet reports of one lot of Vihtavuori N140 from the 1990's going bad caused me to look at a sample I got from my dad that was sitting on the back of a shelf in the basement. Acid fumes had left the steel capped plastic bottle I had it in looking like this after about 20 years:
In fairness to VV, that was not an approved container. Even though the basement is pretty dark, light getting through the plastic won't have helped matters. The "powder" inside was fused into one oily looking lump. But I've had powder in proper containers go bad. Last year I mixed a pound of Accurate 3100 I probably bought around 1990 into my lawn fertilizer and spread it around. I had been going through old containers and sniffing them, and that one smelled like nitric acid, which is a little hard to describe to those who haven't ever smelled it, but it is a sharper and less pleasant smell than the ether fumes good powder normally gives off a little. Red vapor in the air is another sure indicator and this stuff was starting to clump, too. So, out it went. Whether it actually helps the yard or not is debatable, but it can't hurt it. Not around here, where the soil is powdered limestone that can neutralize even the heaviest application of sulfur in no time, so a little nitric acid won't be a challenge to it.
The bottom line is I recommend you load fresh and don't delay too long in using the ammo up once the container of powder is empty. You can keep it longer if you still have the original container of powder to check, and if it is still good, so is your stock (assuming it has been kept in the same conditions). If you are trying to keep an inventory of loaded ammo, keep it sorted by age and shoot it based on the FIFO principle, using up the oldest first and replacing it with the newest.