If he somehow wins every single game he plays in, then yes, obviously.
Practically? I dont think there's much of a chance of it. The weighting of ratings makes it impossible for him to raise his rating with anything but straight wins. I believe at his peak Fischer was the only player in the 2700 range, with Spassky trailing behind at about 2690, and the only reason Fischer was able to be +100 points ahead of the competition was that he just kept on winning.
In the current pool he's the best, no doubt, but versus previous generations, there is no way to tell if lets say, Capablanca was worse than Carlsen. You could go over their games, but Capablanca had no computers, his theory was different then ours, so compared to the talents of other generations, he is overrated.
Below are the results of the 2009 Pearl tournament with ratings:
| Participant | Rating | 1 | 2 | 3 | 4 | 5 | 6 | Points | Tiebreak | Performance | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 |  Magnus Carlsen (NOR) |
2772 | * * | 1 ½ | 1 ½ | 1 ½ | 1 ½ | 1 1 | 8.0 | 3002 | |
| 2 |  Veselin Topalov (BUL) |
2813 | 0 ½ | * * | ½ ½ | ½ ½ | ½ 1 | ½ 1 | 5.5 | 2789 | |
| 3 |  Wang Yue (CHN) |
2736 | 0 ½ | ½ ½ | * * | ½ ½ | ½ ½ | ½ ½ | 4.5 | 2735 | |
| 4 |  Teimour Radjabov (AZE) |
2757 | 0 ½ | ½ ½ | ½ ½ | * * | ½ ½ | ½ 0 | 4.0 | 20.00 | 2695 |
| 5 |  Peter Leko (HUN) |
2762 | 0 ½ | ½ 0 | ½ ½ | ½ ½ | * * | ½ ½ | 4.0 | 19.25 | 2694 |
| 6 |  Dmitry Jakovenko (RUS) |
2742 | 0 0 | ½ 0 | ½ ½ | ½ 1 | ½ ½ | * * | 4.0 | 17.25 | 2698 |
He performed at 3002 while his rating was 2772. Next year his rating was 2900+ for the same tournament. So he is definitely capable of performing at that level. I am not sure if even kasparov has had 2900+ rating for any individual tournament. But like Carlsen himself said he doesnt think performing at 3000 consistently is easy but feels 2900 is acheievable. There was a time when kasparov was on 2800+ while only player in 2700s was Karpov. So maybe after 20 years, we might see about 10 GMs in upper 2800s, and Carlsen maybe be close to 3000. As one GM said, "Carlsen may not have peaked yet, which is frightening!"
Cheers
I Think that it is possible for Magnus Carlsen to reach 3000 elo !! Maybe More. !! Just consider that chess engines like KOMODO are reaching elo 3500 pretty soon , now GM's like Carlsen surely play and study with these chess engines and improve there Human chess knoledge . We made chess engines better , with time they will make Human chess better. !!! Just a matter of Time and Study. !!
Well, I dunno, four years ago I was unconvinced, a year ago I was unconvinced, but I think that tide of exclamation marks may have finall persuaded me.
Why people talk about 3000 when no one has ever reached 2900?
NEVER!!!
Unless there is massive rating inflation I don't think it is possible for any human to reach 3000. You would have to score incredibly heavily and consistently against 2800 players and basically not ever draw against anyone lower than that. I agree with Drakodan.
No, I don't think so.
Anything is possible so possibly
Magnus should have been at 3010 against Ding Liren. But FIDE don't count score if the outcome is determined ...
I Think that it is possible for Magnus Carlsen to reach 3000 elo !! Maybe More. !! Just consider that chess engines like KOMODO are reaching elo 3500 pretty soon , now GM's like Carlsen surely play and study with these chess engines and improve there Human chess knoledge . We made chess engines better , with time they will make Human chess better. !!! Just a matter of Time and Study. !!
Not a matter of time and study, a matter of math, you need a bigger pool or do you seriously believe Calsen will be the only one to improve? Other study too! Carlsen seems to have reach (or even passed) a peak now. Mean that he's not improving faster than others any more! And please, it has been said again and again, ELO is not objective strength, only relative to a pool. All the pool benefits improvement of chess level all over the world so forget the idea that studying improves ELO rating.
And now forget your argue about chess engines as well. Chess engines aren't FIDE rated at all! Who put this in your head? It's not even en evaluation, it's only an interpolation. It's more like "if there were an infinite pool of Carlsens and other GM for the engines to fight against and get ELO points from, then the engine level would be..." If these engines were to get a FIDE rating and human players being forced to play against them, then the scenario would be totally different, far from the interpolations we have now, and Carlsen's ELO rating would drop dramatically loosing again and again and again versus stockfish.
He beat Kasparov's peak 5 years ago.
This is an old thread. It's interesting that many people who seem to know quite a bit about chess don't really understand how FIDE chess ratings work. It would be almost impossible for a human to break 3000 FIDE any time soon. Carlsen is the most dominant player on the planet, and has been for about a decade now. But in order to break 3000, he would have to be even more dominant than Morphy was over his peers.* That just isn't going to happen.
_________
*According to chess metrics, Morphy was ~120 points better than the next best player (Adolf Anderssen) in May, 1859. Source: www.chessmetrics.com/cm/CM2/MonthlyLists.asp?Params=185705SSSSS3S000000000000111000000000000010100
If Carlsen were 120 points better than Caruana, today, that would still put him at only 2942. That may seem kinda close to 3000, but it's miles away in terms of dominance. Carlsen is clearly the best player in the world, but the next ~15 best players are all extraordinarily strong and no human player will ever again dominate the chess scene the way Morphy did.
According to chess metrics, Morphy was ~120 points better than the next best player (Adolf Anderssen) in May, 1859.
According to FIDE, Kasparov and Karpov were ~120 points ahead of everyone else around the late 80s early 90s
http://fidelists.blogspot.com/2008/03/january-1989-fide-rating-list.html
And Fischer was 125 ahead of Spassky in 1972, 120 ahead of Karpov in 1973.
Yes. 120-130 rating points appears to be the maximum. I don't see any human becoming that dominant again unless chess becomes so unpopular that few people see the benefit of getting good at chess. And even if they did become that dominant, there's still a long way from 130 rating points to ~180 rating points.
The very impressive thing with Carlsen is how he dominates all speeds the same way. After him, the 2d 3d etc are very different whether you look at blitz, rapid or classic, but he has the same domination over all. That for me shows he is already AS dominant as can be. Again, it makes really NO sense to assume Carlsen will improve but not others. Most are younger than him now. Probability is he will now run into people improving faster than him. And this mean ELO decreasing, even if his level keeps improving.
Now if someone here is so sure that Carlson will hit 3000 this year or next, instead of arguing, please contact me to make a bet. I bet as many as you want he won't. Good chance. (And easy money for me).
I know all that, I'm simply saying he's not overrated because the rating system isn't a comparison of him versus previous generations, but against the current pool.