When speaking of loot, it is good to look at what is expected of it.
The first thing that came to mind is what place the loot can take in Fate to increase the narrative. For this aspect of loot, I came up with a couple of goals.
Loot should add to the narrative. At it's core Fate is a storytelling game. If the hero picks up a random sword, it doesn't add in an overall way to the narrative. However, the Bloody Blade of the Unstoppable Warlord Kharm tells a story, and over time can add to the narrative.
Loot should adapt to the narrative. Over time, as the character changes and adapts because of the story, if the loot that he obtained in his first adventure never changes with the character, it makes one part of him static. Boring. As the character changes and grows more powerful, so should his loot.
The second part is the general player satisfaction. One part about seeing a chest after a long fight in a dungeon is what secrets does it hold, and how will it make me more powerful? If the player doesn't receive something crunchy from the loot obtained, it might fall short in satisfaction for a long road slogged to get it.
The third part goes hand in hand with player satisfaction- balance. If only one of the players is going to get the Bloody Blade of the Unstoppable Warlord Kharm, will he suddenly become out of balance with the other players? Will you start having to balance towards his character rather than the group?
So, taking these in mind, this is what I did for a fate-core fantasy game that I recently ran.
In the Extras chapter in the Fate Core book, the rules refer to Weapon and Armor ratingsFC277. In a game with loot, these are a necessity IMO. They differentiate the types of weapons, and give more mechanical heft to the items found, and their lack.
Once this is in place, I dug further into the extras (and the yet to be release Fate Toolkit), and looked at how I viewed the parts of the game as they related to things that were not a part of the character- namely Aspects, Skills/Stunts, and Fate points- and how they related to advancement.
The one concession that I first made was that the point was to balance in general, and not at any given point in time. After the person picks up the Bloody Blade of the Unstoppable Warlord Kharm, for a short time after he is able to use it, he will be more powerful. But by the time of the advancement, I want them to be on equal or near-equal footing. This also makes the placement of items within the story important.
The other thing I did was treat any weapon worthy of being loot as a Fractal. This was overkill for many cases, i.e. I only want to add an aspect to a weapon... is this necessarily a fractal? But what I found in the long run, looking at my goals, this actually increased player satisfaction with those special items, and over time gave them personalities of their own as they advanced alongside the player.
With that out of the way, the first thing to look at is Aspects. Aspects make a character more flexible within the narrative, and there is no explicit way in advancements to get another aspect- only to rename them or change them around. This either makes new aspects very powerful (in that you don't get more), or more of a narrative issue (in that you can switch them with narrative changes at any milestones excepting the high concept/trouble).
I chose to view aspects as not an issue other than with flexibility, and so associated no cost (other than having the item) with aspects on items. So if the Bloody Blade of the Warlord Kharm has the aspect Hungers for the Blood of Warriors, I wouldn't say that was anything other than a characteristic of the blade. You can invoke this (and it can be compelled against you) only while the blade is in on your person. But other than this limitation, you can use it just as you can any of your other aspects.
The next thing is Skills and Stunts. These definitely mechanically make the character not only more flexible, but also more powerful. The limiting factors to this are similar to stunts - the cost of Fate Points to activate. But just having the ability is an advantage also, even if you don't have the use of it- you have the option to do so.
With this in mind, I treated these similarly to stunts, one and all. I rated them for refresh cost, and if the character chose to keep the weapon, they had to pay for the cost in refresh- either out of their current refresh, or incur a refresh debt to be paid over time (see below for more on refresh debt)
The last thing is Fate points. This is a very powerful thing to put into an item. Adding Fate points to an item is to give it refresh. In the Dresden Files, it made it very clear that refresh was akin to how much free will a character had- so I looked at items that were in this category as items that had their own will, i.e. artifacts. The could use them for the character's benefit... but this is a two edged sword. Again, this is something that comes into play with the concept of a refresh debt)
With that out the way, the last thing I added was something of my own- what I call a Refresh Debt (countered by Refresh Investment).
The more refresh an item costs (either from Fate points or the cost of an ability), the more powerful the item is. It takes a very strong willed person to tame such items- they have to invest a part of themselves into the item to get it to obey their will. Until they have done so, that will is in the hands of the object, i.e. the user can find themselves in an escalating contest of wills with the object. Any Fate that is in the hands of the item is also always in question.
Take the example from above- the Bloody Blade of the Unstoppable Warlord Kharm. It has the aspect Hungers for the Blood of Warriors, it gives a +2 bonus to any Fight attack made in when fighting more than one opponent (1 refresh). It is a 1 refresh item, and costs 1 refresh to the character that will wield it. The character can pay 1 refresh immediately, and dominate the weapon. But what if he doesn't have the refresh?
In that case, the amount of Refresh Debt is utilized in contests of will, and when determining the frequency of them. It is basically used a Fate points that the item can use towards asserting its dominance. In the case that items have their own Fate points to use, this is always counted as refresh debt. The item has its own will at that point and is not just a character in mechanics, but a character in truth, with its own personality, and can be bargained with. In those cases, I also gave the item a mental stress track and applicable skills that could be used in these conflicts.
Refresh Investment is the counterpart- being how much refresh, i.e. how much if the character's free will is invested in the artifact. This takes care of the fact that the artifact is an item that can be lost/stolen. This refresh investment can be used as Fate points in actions to locate and recover the artifact.
The last part of the equation is Advancement. When the character reaches a milestone in which they have been utilizing the artifact to a great extent, the moderator can talk with the player and choose to apply a milestone to the artifact. This milestone can be something that the character and artifact strive towards (i.e. unlocking an additional effect that the artifact has), or something that the artifact itself strives towards (i.e. increasing its dominance and will). If the latter, it should have been something obvious in the narrative, and talked over with the player, and especially evident in the actions that the artifact has taken. At that point, the player can choose to spend the character's advancement re-asserting domination, or let it carry the character further into refresh debt.
Best Answer
I'm also a big fan of Fate and Magic the Gathering, so here's to hoping I do your question justice!
Balance and Feasibility
To answer some of your concerns first, I will say that such a system is totally feasible and can be "balanced" just fine, as long as you and your players all agree upfront what it should look like, what the limits are, and how common it should be. Fate handles tweaks, modifications, and homebrew spectacularly well (it's basically designed for that), so the limiting factor is generally what everyone at the table actually thinks is fun.
For instance, if only one player is at all interested in summoning, and the party never faces anyone who does it, that player might A) seem "better" than his comrades if he and his summons get to act each turn in Conflicts, Contests, and Challenges, and B) get frustrated that the game is so much more complex for him than everyone else!
On the other end of the spectrum, if everyone and their mother in your setting is waltzing around with a small army of Goblin Tinkers (all exchanging Ornithopters for Darksteel Golems every turn, of course), the complexity problem might rear its head in a different form: turns take so long for everyone and their forces to act that you feel like you're playing a plodding wargame rather than an exciting RPG based around action, inventiveness, and collaboration.
If your players are interested in doing this, and you all agree upfront how complex and powerful summoning should be to fit your desires and goals, then there is no problem implementing it. You don't need to be concerned about it "only" costing a Stunt, Skill, or Aspect, if it's designed around the amount of cost you assign it. In fact, depending on how powerful and common it is, playing with the cost involved (maybe it sucks up Refresh, an Aspect Slot, and a relevant Skill!) is a great way to keep it from overwhelming other skills and ways of playing the game.
Unless, of course, it's your intention to have Summoners stomp all over paltry swordsmen and politicians and songwriters and whatever other kinds of "classes" are common in your gameworld!
The "Storm Summoning" System
First and foremost, I'd suggest giving the Storm Summoners magic system proposed in the Fate System Toolkit a read. It plays with the notion of summoning in Fate by tying it to elemental "storms" of power inhabited by creatures who can be summoned or bargained with through the use of a "Conjuration" skill.
As a brief summary: for basic summons, you roll against the power of the elemental you want to summon (from Average [+1] Wisps up to Great [+4] Attendants) as a Create an Advantage Action requiring material components. Success grants you control over the creature for a set amount of time, while failure summons an uncontrollable elemental.
The skill is also used to attack (banish) other conjurations or to defend against their attacks with magical wards, as well as to renew existing summons. The total number of summons allowed is also determined by your skill.
The summons generally have only one skill (___ Elemental) used for any action they could reasonably accomplish specifically by being one (e.g., fire attacks, walls of ice, control of water, etc.). Smaller ones are weak, but have bonuses to speed-related skills, while larger ones have more Stress and Consequence slots. Each element also grants an additional ability (e.g., Fire grants a Weapon:x rating, while Earth elementals have more Stress boxes).
If a character makes a lasting (and expensive) bargain with a specific element, he gains bonuses when summoning it and gets access to even more powerful (+5 and higher) summons. This bargain is usually represented, mechanically, by an Aspect denoting it, which of course can be invoked (bonuses on conjuration rolls) and compelled (being forced to do work for your chosen elemental storm).
In combat, the Elementals use teamwork rules like Mooks (strongest one sets the base attack power, others act as +1s), and can aid the summoner on Skill rolls that make sense or be sacrificed to "absorb" blows on behalf of the summoner.
Specific M:tG Tweaks
The first obvious tweak is to replace the elements with the five colors of Magic, or whatever other obvious power-separations make sense in your own setting (I'd start there; save multi-color or similarly complicated summoning for later revision or see if it's necessary at all).
You can then tweak the summoning specialty chart, as well: Maybe Red-aligned creatures gain a bonus to Athletics or a Weapon rating, while Blue-aligned creatures can be sacrificed to banish an equally powerful-or-weaker creature controlled by someone else, for instance.
You might also consider replacing the single Conjuration skill with separate ones for each flavor of magic in your setting. That's more complex, but my guess is, if you want to have summoners running around arenas trying to bash each other to bits with giant gorillas and the like, players in the game will want to be as cool and unique as possible while doing so!
In that case, you can modify the skill's usefulness from just Summon/Banish/Defend against Elements to more flavorful uses. Perhaps White Magic replaces Lore/Medicine/etc. as the default Skill used to recover Physical Consequences with an Overcome action, or maybe the Black Magic Skill gets a Stunt to take Stress in order to increase your rolls with it.
For easiness' sake, I would strongly recommend you keep up the idea of having all summons act as a mook mob (single action per exchange), particularly if the party will have or face multiple summoners (in fact, if everyone is going to be summoning, I'd even consider allowing a Summoner or his summons to act each turn, although still granting each other Teamwork bonuses when not "in use.").
Artifacts could probably be represented by Extras (built using the Golden Rule--a Cloudpost artifactor even Stunts