Well... after you pushed the nerfs through and made the only remaining true caster tree in wizard (Pale Master) basically useless until epics (not here to debate that, see other threads by me if you want to debate that) we are at the point that wizard has lost.. any semblance of class identity.
There is not a single thing a wizard can do, that another class cannot do far better in heroic (and I am talking only heroic here) content. Divines are the better necromancers (and healers, and far more defensive all in one package), bards are the better CCs, paladins the better melees and sorcs do more raw damage.
This is a problem. One that could be fixed by finally overhauling the archmage tree.

Now, before we start with that, let's look at what I see as ... not so good options for the new tree:
- another single-element caster with a bypass-that-element that is a carbon copy of basically a sorc (and either makes them useless if stronger, or is pointless itself, if weaker) even if said element is "force"
- another DC-tree because that, even though now mangled and battered, would've been the role of the PM, if that tree ever gets overhauled to be useful again in heroics...
- a physical combat tree, because that is already filled by EK for melee and IQ for ranged
- the same as now, with slightly tweaked numbers, because this tree is fundamentally flawed and cannot be fixed by adjusting some numbers

So, what could be built on the basis of a wizard, that is both not already represented and at least somewhat innovative, given the limitations of a single enhancement tree (so no new spells, feats, etc...)?

A true master of metamagics.

While most casters have some support for metamagics included in their trees (cost reductions) none are really currently defined by it. And that is where the revised Archmage tree could go towards.

Thus I here present a "prototype" of a tree. The tree does not contain anything that has to do with numbers - no generic INT, no generic Crit enhancements, no DCs no nothing of the sorts - just the "interesting" parts of the tree, that don't boil down to numbers. That does not means it should not have that, only that I don't wanna suggest values for those - the devs should know that part better. I want to only present what would make the tree "unique", what would give it it's identity.

Core 01: Select a defensive spell whose effects will be incorporated into your very aura, making it permanent, from this list: Shield, Night Shield, Jump, Blur, Protection From Evil, Expeditious Retreat, Heroism, ...
Core 02: Select another one from the Core 01 list
Core 03: Select another one from the Core 01 list
Core 04: Select a defensive spell whose effect will be incorporated into your very aura, making it permanent, from this list: Lesser Globe of Invulnerability (immune to spells 3th lvl and lower), Fire Shield (incl. cold version), Mind Blank (immune to charms and commands), True Seeing, Greater Heroism, ...
Core 05: Select an option from Core 04 not yet chosen

Tier 1:
- Subtle Spellcasting or Brute Spellcasting (increase or decrease in threat from spells because why not support i.e. EK "tanks", they might be rare, but no hurt in supporting them)
- Keep Wand and Scroll Mastery (because that fits thematically very well here)
- Mind Mastery: pick any SLA from the following list: Grease, Hypnotism, Invisibility, Magic Missile, Resist Energy, Web, Otto's Resistable Dance, Knock, Protection From Energy, Teleport, ....

Tier 2:
- Mind Mastery: pick any other SLA from the list you haven't picked already
- Keep Efficient Metamagic

Tier 3:
- Mind Mastery: pick a third SLA from the list you haven't picked already
- Metamagic: Eschew Materials (like the Feat)
- Keep Arcane Bolt
- More Efficient Metamagics

Tier 4:
- Greater Mind Mastery: pick any SLA from the list: Flesh to Stone, Hold Monster, Dismissal, Fire Shield, Stoneskin, Haste, Dimension Door, ....
- Metamagic: Extend Spell (like the Feat)

Tier 5:
- Greater Mind Mastery: pick a second SLA from the list you haven't picked already
- Metamagic: Enlarge (like the Feat)
- Mastery of Metamagics: Your metamagic feats do not increase the cost of spells cast with them. Prerequisite: the efficient metamagics in the lower ranks
- Arcane Blast
- Arcane Supremacy: Activate to hone your skills and power into a spell frenzy, gaining +25% Spell Critical Chance and +100% Spell Critical Damage for 20s, but your metamagic feats have no effect on spells for the next 60s afterwards. (Implemented i.e. by a debuff that is applied to the character for one minute duration). CD: 60s, unlimited charges or 5 charges / rest. Exclusive to Arcane Mastery.
- Arcane Mastery: Activate to hone your skills and power into a spell frenzy, gaining +6 to all DCs and overcoming any spell resistance for 20s, but resisting your magic will be very easy for enemies for the next 60s afterwards (effectively a -99 DC to all schools, forcing you to use no-save spells). Exclusive to Arcane Supremacy.

The goal of such a tree is to make a wizard that has a reason to use more then one school of magic and more then one element out of those, because he has no bonuses to anything in particular, but is rather defined by being the only caster in the game, that can utilize a lot of metamagics on each spell. If you can activate eschew, heighten, maximize, empower, quicken and enlarge you might, finally, not feel like a weak sorc tossing that fireball, but like a viable alternative. Not strictly better (you get NO other spell power, no crit, no nothing...), the raw numbers will probably still be lower. But *your* fireball has more range, your buffs last longer, you don't need components. And that is something no other class or tree currently brings to the table.

It also could really free up gearing. If you have a good reason to use any element, but none in particular, you don't gear with the 35 or so different effects for one spellpower (equipment, insightful, exceptional, artefact, profane, even-more-fancy, a-lot-more-special ....) but maybe stick to one or two - i.e. on switchable weapons, but use all spells - just with more metamagics attached to them. The danger of this is abuse by other classes, which is somewhat remedied by the interesting (free metas) enhancement being in T5, thus locking out other T5s. If you are to worried about that, change the text to read to "... do not increase the cost of wizard spells you cast with them."

If you implement this correctly, a wizard that has taken Arcane Supremacy can still use (non-quickened, non-heightened) DC spells while it is on CD / fatigue and a Arcane Mastery can use spells that either have no save or at least deal half damage with other spells. This will encourage actually using all the spells slots a wizard gets, because the reality right now is that there are fast stretches of spell levels where you just don't have any useful spell to slot. If you however gotta "weave" past the downsides of the T5 enhancements, you can can actually start slotting spells that fit those circumstances. Just to give you an example, spell level 8 is considered by most to be... *cough* just slightly subpar *cough*, but if you have Arcane Mastery on CD (and thus very low DCs) a spell like polar ray that has no save attached to it.. could actually, with all metamagics active, still do good damage. And even if you are not a necromancy specialist, why not fish for low rolls with FoD/FtS/PK/... while your Arcane Supremacy CD/fatigue ticks down. And you could even carry a secondary gear set, i.e. swap a ring or two, encouraging actually using the full breadth of all (or at least most...) of the spells a wizard has to offer, instead of blasting through all heroic content on the same spells.

This also serves to further differentiate a wizard from a sorc. Because right now, a wizard feels like a poor man's sorc. They cast the same maybe 10 to 15 spells all throughout heroics, because these are simply numerically the best, only that the wizard is worse at it. Focusing on just a few spells, in my humble opinion, should be the domain of the sorc, the wizard's domain should be versatility, but without any support from the trees, that just ain't happening. Why use a dismissal or hold monster if a finger of death will do? Well, maybe because dismissal can be quickened and enlarged for free or because the hold monster can be enlarged, extended, heightened and quickened for free. That means they can be cast from further ahead, without any cost and are more likely to land then your non-metamagic FoD. Which gives them a purpose, other then "filler content nobody ever uses because there are better alternatives".

The rework of Arcane Supremacy also makes it a lot less frustrating when it procs just as the last orc at 10% hp dies when the boss and the 20 adds are already dead but gives a noticeable downside to it - think the equivalent of a barbarian rage for casters, which is also something no caster currently *really* brings to the table (there is the reaper boost and a weak effect on warlocks, but nothing mayor). And since not everyone plays dmg casters, there is a DC caster option, whose downside should be that you cannot really hit anything for a minute afterwards, again lending to the idea of a burst of power with fatigue attached to it.

If you want to reinforce on that theme, you could also do a T4 enhancement that grants say 200 MRR (that is not affected by the cap ofc) but lowers DCs by 4 and universal spell power by 100 that can be activated and lasts.. 20s or so? Again reinforcing that the caster is not more powerful then other casters, but far more experienced in bending magic in whichever way the caster needs at the moment - which to me is the definition of an archmage (compared to say a high level sorc that is just "raw, unadulterated power"). This would also make an EK / AM hybrid interesting.