In its fully-powered form, this deck doesn’t generally have many glaring weaknesses. Having played more than a thousand games with it, I’ve found only a few weaknesses, and attempted to avert them wherever possible.
First, we are a spell-based, storm-esque combo deck, so we tend to have the same weaknesses as that archetype generally does. Stax restrictions such as Arcane Laboratory, Spirit of the Labyrinth, et cetera tend to hit us very hard, if they interfere with our ability to cast many spells or our ability to draw cards. I’ve found the best way to work around this sort of thing is to either use our tutors to dig for removal before we go off (Firestorm can pull double duty here, and most of our answers can be found with Spellseeker/Mystical Tutor/etc), or to just generate an army of Locusts and win with a backup aggro approach. This isn’t a crippling weakness in that matchup, however, as we can often establish a large amount of board presence before they can lay down their stax pieces.
Second, hard control decks mess with us, as should be expected for a deck that wins primarily through combo. I’ve played Teferi, Mage of Zhalfir and/or Defense Grid in the past to combat this weakness, but at the moment the primary answers for control here are in the form of our own counterspells. If you are concerned that your opponents have a counterspell, it’s usually worth holding up that mana. And if they didn’t have a counter, hey! It’s extra mana to combo off with. Pact of Negation and Force of Will are important here, because they give us ways to interact without spending limited mana.
Third, faster combo decks can cause us a problem, if they are true cEDH builds (i.e. winning T2-4 on average). I’m generally not too worried by something like Food Chain Prossh, because they usually can’t match our counterspells, but decks like Kess Doomsday-Storm or Tymna/Thrasios can be an issue, because they win about half a turn faster than us on average (and have access to tutors that we do not). Unless you have an exceptionally fast hand, the best plan here is usually NOT to race them to a combo. Instead, hoard your counterspells and aggressively target their wincons, manabase, and card advantage spells. Our deck is usually more resilient than them if you can survive the initial onslaught, because we run a broader suite of wincons and thus are more able to recover after a blowout. Mystic Remora is a priority in this matchup, and I used to run Mindbreak Trap to handle this sort of thing.
Fourth, consistent removal on our commander can hurt. He doesn’t pay commander tax, usually, but if he is removed in response to a wheel, it can deny us Locusts that we were planning to rely upon. Moreover, a 6-CMC commander already isn’t cheap, so if he gets countered or exiled a few times, things can get tricky (and push us towards infinite-mana-centric combos). I’m running Cavern of Souls (“naming God”) to help here, but I’ve still had issues occasionally. As this deck has tilted more towards combo rather than aggro, this weakness has lessened: many games I don’t even cast The Locust God until I already am in the process of comboing off (particularly under Paradox Engine or Dramatic Scepter).
Fifth, mass artifact removal HURTS. Our deck is built around artifact-centric combos, and relies VERY heavily on mana rocks to give our deck speed. This means that effects like Null Rod, Stony Silence, Bane of Progress, Vandalblast, etc hit us very hard. Hopefully you’ll save a counter for these, but…can’t always rely on that, unfortunately.
Sixth, the deck can be vulnerable to early aggro damage. Unless The Locust God is in play, we have almost zero blockers, meaning that our life sometimes gets uncomfortably low if the table decides to play Archenemy. Our answers to aggro are not very good in UR; other than Blasphemous Act, most board wipes cost too much mana to be viable. Propaganda can help patch this issue, and has gone in and out of the deck at different points.
Seventh, if playing against proper cEDH decks, don’t wheel until you absolutely HAVE to. Our wheels are playable here because our mana curve is low enough to break parity: we can empty our hand very very quickly, then wheel when our opponents still have many cards in hand. However, other cEDH decks have curves as low (or even lower) than us, so a wheel can be sharply in their favor. If this is the case, things get tricky, and I try to only wheel once I’ve either 1) completely run out of gas and have no other choice, 2) I can nullify multiple tutoring effects, or 3) I’m planning to win that turn.
And finally, sometimes the non-infinite Paradox Engine combos simply fizzle, if you wheel and draw a hand full of lands. The deck uses as many wheel effects as possible to avert this, and our card selection usually solves this problem, but it's still something to be aware of. I’d say this happens maybe 3% of the time that you resolve a Paradox Engine?