It's a bit difficult to answer this question without first coming to an understanding of how moderation is currently getting done... So, let's do that.
Here are some statistics on various moderator actions performed during the past 365 days, separated into two columns: Moderators means people with diamonds next to their names; Community means both people without diamonds and the system itself (usually triggered indirectly by normal user actions):
Action Moderators Community
--------------------------------------- ---------- ---------
Users suspended¹ 13 15
Users destroyed 447 0
Users deleted 3 0
Users contacted 40 0
Tasks reviewed²: Suggested Edit queue 287 2079
Tasks reviewed: Reopen Vote queue 100 209
Tasks reviewed: Low Quality Posts queue 712 1263
Tasks reviewed: Late Answer queue 130 1205
Tasks reviewed: First Post queue 300 5803
Tasks reviewed: Close Votes queue 725 3251
Tags merged 10 0
Tag synonyms proposed 10 0
Tag synonyms created 10 0
Tag highlight language set 12 0
Revisions redacted 10 0
Questions unprotected 1 1
Questions reopened 53 2
Questions protected 92 43
Questions migrated 54 0
Questions merged 7 0
Questions flagged³ 84 13924
Questions closed 1341 454
Question flags handled 1100 12908
Posts unlocked 5 16
Posts undeleted 24 107
Posts locked 11 2230
Posts deleted⁴ 1152 5585
Posts bumped 0 5110
Escalations to the CM team 3 0
Comments undeleted 26 0
Comments flagged 546 632
Comments deleted⁵ 1728 1840
Comment flags handled 790 388
Bounties canceled 5 0
Answers flagged 50 3114
Answer flags handled 2601 563
All comments on a post moved to chat 22 0
Footnotes
¹ The system will suspend users under three circumstances: when a user is recreated after being previously suspended, when a user is recreated after being destroyed for spam or abuse, and when a network-wide suspension is in effect on an account.
² This counts every review that was submitted (not skipped) - so the 3 suggested edits reviews needed to approve an edit would count as 3, the goal being to indicate the frequency of moderation actions. This also applies to flags, etc.
³ Includes close flags (but not close or reopen votes).
⁴ This ignores a good chunk of deletions that happen automatically in response to some other action.
⁵ This includes comments deleted by their own authors (which also account for some number of handled comment flags).
Observations
A couple of things I'd like to draw attention to in the list above:
The community does the majority of close and reopen reviewing, but moderators are responsible for the majority of questions closed and reopened - this suggests that there simply aren't enough reviewers to close or open questions without moderator assistance (outside of duplicates, it always takes 5 normal-user votes to close or reopen).
The vast majority of flags are handled by moderators. This suggests that there aren't enough people in Low Quality review to completely handle posts that end up there either.
Finally, answers to your questions
Lets face it, we do not have a lot of people doing close vote handling. If we now take 2 active ones out of this pool we have even less.
You're not "taking anyone out of the pool"; you're giving two of the folks in the pool better tools: unlimited, binding votes. Obviously, you should be sure you're giving them to folks with good judgement who'll use them to assist the rest of you and won't use them in ways you wouldn't agree with... But assuming you do that, this shouldn't hurt your collective ability to get things done.
So how big do you guys see this risk is, and what can we do to mitigate this?
Well, as I just said... The obvious mitigation is to make sure you vote for competent candidates.
Beyond that... Encourage more people to use their moderation abilities. The more normal folks you have closing and reopening, the more normal folks you have reviewing, the fewer mods you need and the less they need to act. Moderators' primary function is to pick up the slack when there just aren't enough active user-moderators to do what needs to be done; the less of that there is, the less you need mods.