How many negative keywords can a Google Ads campaign have?

A single campaign can hold up to 10,000 negative keywords. For Performance Max this was a major change in March 2025, when Google raised the PMax per-campaign limit from 100 to 10,000 to align it with Search campaigns. Search campaigns already supported large negative counts. In practice most advertisers use only a few hundred campaign-level negatives, so the 10,000 ceiling is rarely the binding constraint on a single campaign. The constraint that actually bites at account scale is the shared negative list limit and the account-level negative limit, which are much lower and are what you have to design around when you manage many campaigns.

What are the limits on shared negative keyword lists?

The documented limits are 20 shared negative keyword lists per account, each holding up to 5,000 keywords, and a separate account-level negative keyword limit of 1,000. Shared lists became available for Performance Max campaigns by August 2025, so they now apply across Search, Shopping and PMax rather than Search alone. Some advertisers have reported being able to add more than 5,000 to a list in practice, but the documented ceiling is 5,000 per list, so design to it rather than relying on an undocumented allowance that could be enforced at any time. Twenty lists of 5,000 is 100,000 shared negatives, which is ample for almost any account if you organise them well.

Should I use campaign-level or shared negative lists for scale?

Use shared lists for anything that should apply across multiple campaigns, and reserve campaign-level negatives for exclusions that are genuinely specific to one campaign. The whole point of shared lists is single-point maintenance: change the list once and every campaign it is attached to updates, instead of editing the same negative in twenty campaigns and inevitably missing some. The failure mode at scale is fragmentation, where the same waste term gets added ad hoc to individual campaigns over months until no one knows what is blocked where. Consolidating recurring exclusions into a small number of well-named shared lists fixes that and is the main structural upgrade the expanded limits enable.

Do shared negative lists work in Performance Max now?

Yes. Google completed the rollout of shared negative keyword list support for Performance Max by August 2025, so a shared list can now be applied to PMax campaigns the same way it applies to Search and Shopping. This matters because before that, PMax negative management was campaign-by-campaign and capped at 100 per campaign, which made account-wide hygiene across PMax almost impossible. With shared lists and the 10,000 per-campaign ceiling, PMax negatives can finally be maintained centrally. Remember that PMax negatives only filter Search and Shopping inventory, not YouTube, Display, Discover or Gmail, so the list controls part of PMax spend, not all of it.

How many negative keywords do I actually need?

Far fewer than the limits suggest, if the list is structured by theme rather than by individual term. Most advertisers run a few hundred to under a thousand meaningful negatives per campaign, and a well-built shared list covers most recurring waste with a compact set of terms grouped into patterns. The goal is coverage, not count: one phrase-match negative for a recurring junk theme blocks more waste than fifty exact negatives for individual variants of it. Chasing the 10,000 ceiling is a sign the list is being built term-by-term instead of pattern-by-pattern, which is slower to maintain and easier to break with an accidental over-block.

Scaling Negative Keyword Lists: The Real 2026 Limits

The ceilings on negative keywords moved in 2025, and most accounts have not been re-architected around the new room. Google raised the Performance Max per-campaign negative limit from 100 to 10,000 in March 2025, aligning PMax with Search campaigns, and completed the rollout of shared negative keyword lists for PMax by August 2025. For years the binding constraint on account-wide negative hygiene was that PMax capped you at 100 per campaign and would not take a shared list at all. That constraint is gone. What is left is a different design problem: not “how do I fit my negatives in” but “how do I organise them so a human can still maintain them”.

This post is about that structure. The limits are generous enough that count is no longer the problem for almost any account — fragmentation is. The same waste term added ad hoc to fifteen campaigns over six months is the real failure mode, and the expanded limits plus shared-list support are the tools to fix it. We will walk the actual documented ceilings, why chasing the big numbers is the wrong instinct, and how to consolidate a scattered pile of per-campaign negatives into a small set of high-coverage shared lists that one person can own.

The real limits, stated plainly

Here are the documented ceilings you design around. A single campaign can hold up to 10,000 negative keywords — the March 2025 change that brought PMax up to parity with Search. Shared negative keyword lists are capped at 20 lists per account, each holding up to 5,000 keywords, which is 100,000 shared negatives in total if you ever needed them. Account-level negatives — the ones that apply across the whole account automatically — are limited to 1,000. Those three numbers, 10,000 per campaign, 20 lists of 5,000, and 1,000 account-level, are the frame for every structural decision below.

You will see field reports of advertisers adding more than 5,000 to a single list in practice, and they may be accurate, but the documented ceiling is 5,000 per list and you should design to it. Building an account that depends on an undocumented allowance is building on sand: Google can enforce the stated limit at any time, and a list that silently stops accepting additions is a hygiene gap you will not notice until waste appears. Treat 5,000 per list as the real number, use the 20-list allowance to get headroom, and you never have to care whether the soft limit is higher than the hard one.

Count is not the goal — coverage is

The instinct when a limit jumps from 100 to 10,000 is to fill it, and that instinct is wrong. The number of negatives a campaign needs is a function of how you build them, not how much waste exists. A single phrase-match negative for a recurring junk theme blocks every query containing that theme; fifty exact negatives for individual variants of the same theme block fifty specific strings and miss the fifty-first. Building by pattern rather than by term is the difference between a list of a few hundred that covers your waste and a list of thousands that is slower to maintain and still has gaps. If you find yourself approaching the 10,000 ceiling, that is a signal you are building term-by-term.

This is the same discipline as grouping search-term patterns instead of adding single negatives, applied at account scale. It also protects you from the opposite failure: a giant term-by-term list is exactly where over-negating throttles good traffic, because the more individual strings you block the higher the odds one of them clips a query you wanted. A compact, pattern-based list is easier to audit for exactly that risk. The expanded limits are headroom for organisation, not an invitation to inflate the count.

Consolidate fragmented negatives into shared lists

The highest-value move the new limits enable is consolidation. Most accounts that have run for a while carry the same negatives duplicated across many campaigns, added one at a time whenever someone spotted waste, with no central record of what is blocked where. That is unmaintainable: when a new waste theme appears you have to add it to every campaign individually, and when you want to unblock something you cannot find all the copies. Shared lists fix this by giving the exclusion a single home — edit the list once and every attached campaign updates. Now that PMax accepts shared lists too, this finally works across the whole account rather than Search alone.

Audit what you have before you build. Export the negatives from every campaign, find the terms that recur across many of them, and those recurring terms are your shared-list candidates — they are, by definition, exclusions that should apply broadly. Move them into shared lists and detach the per-campaign copies, so the shared list becomes the single source of truth. The framework for deciding what belongs at account, campaign or shared scope is covered in structuring negative keyword lists across account, campaign and shared scope, and consolidation is where that framework pays off most.

Organise your 20 lists by theme, not by campaign

With 20 shared lists to work with, organise them by the kind of waste they block, not by which campaign they attach to. A common and durable structure is a small set of purpose-named lists: universal junk (jobs, free, cheap, DIY, careers), competitor brands you never want to compete for, off-intent modifiers for your category, and any industry-specific waste themes your account accumulates. Each list has a clear job, so anyone can look at its name and know what it blocks and whether a new negative belongs in it. Then you attach the relevant lists to the relevant campaigns — a brand campaign might take only the junk list, a broad prospecting campaign might take all four.

Organising by theme rather than by campaign is what keeps the system legible as it grows. If you name a list after a campaign, the moment you have two campaigns wanting the same exclusions you are back to duplication. If you name it after what it blocks, one list serves every campaign that needs that protection and the mapping stays obvious. Twenty lists is a lot of thematic buckets — more than most accounts need — so you are unlikely to run out, and if you do it usually means two of your themes should be merged. The limit is generous precisely so you can afford clean separation.

The Performance Max specifics to remember

Shared lists in PMax come with a boundary worth keeping in front of you: PMax negative keywords only filter Search and Shopping inventory, not YouTube, Display, Discover, Gmail or Maps. So a shared negative list attached to a PMax campaign controls part of that campaign's spend, not all of it, and the queries you block are the search-side queries. That is still a large and important slice, and before the 2025 changes it was almost unmanageable, but do not mistake a full shared list for full control of where PMax sends money. The channel-level picture is a separate audit.

For the broader question of controlling PMax spend, the negative list is one lever among several, and the others — brand exclusions, channel controls, the search terms report itself — are covered in Performance Max negative keywords and PMax brand exclusions versus negative keywords. Read those alongside this if PMax is a meaningful share of your spend. The scaling story is the same everywhere though: the limits are now high enough that the work is organisation, not accumulation, and the accounts that win are the ones whose negatives a single person can still explain from memory.

Keep the structure from drifting back

Consolidation is not a one-time project; without a rule it quietly reverts. The way accounts fragment is that someone spots a waste query mid-week, adds a campaign-level negative to stop the bleeding, and never promotes it to the shared list — and six months later you are back to scattered duplicates. The fix is a simple maintenance rule: campaign-level negatives are for genuinely campaign-specific exclusions only; anything that could recur elsewhere goes into the relevant shared list. Enforcing that one habit is what keeps the 20-list structure clean over time.

Tie the promotion step to your regular search-terms review. When you work the search terms report and find a new waste theme, decide in that moment whether it is a one-campaign quirk or a recurring pattern, and route it accordingly — shared list for patterns, campaign negative for quirks. Done consistently, the shared lists stay the source of truth, the per-campaign negatives stay small and specific, and the generous new limits do what they are actually good for: giving a maintainable system room to grow, instead of a term-by-term pile room to sprawl.