Google Ads is removing language targeting from Search, and in a region that runs three languages that is not a minor detail
On August 14, 2026, Google confirmed something it had hinted at months earlier: starting in late September 2026, the campaign-level language targeting control disappears from Search, and from the Search Network portion of Performance Max too. Google wrote it up as a technical tidying. For any account that splits budget across Spanish, Portuguese and English (close to any relevant account in this region) it is a structural change in how audiences get split, how performance gets read, and how an account gets built.
What Google actually says
The Google Ads Help page ("About language targeting") puts it plainly: "Starting in September 2026, the way language targeting works is changing for Search campaigns and the Search Network portion of Performance Max campaigns." Further down: "The campaign-level language targeting setting will be removed. Search ads will automatically match based on the language of your ads." The line most likely to get quoted without reading the rest is this one: "No action is required by advertisers for this change." That is true in the literal sense (there is no button to press, no migration to run), but it describes what happens technically on the day of the change, not what is worth doing in the weeks before it.
What replaces the setting
The mechanism that remains is not new, it was already running in parallel. Google automatically matches based on the language of the ad and the landing page, combined with signals about which languages a given user understands, with AI-based ad group prioritization to surface the most relevant language for that person at that moment. Inside Performance Max the distinction matters: the setting keeps applying to YouTube, Display, Discover and Gmail. It only stops applying to the portion that runs on Search.
Single-language campaign structures (one campaign in Spanish, another in English, each with its own budget) keep working. What changes is what holds them apart behind the scenes. It is the same move already visible when Google started writing the description on your Shopping ads: a control that sat on the advertiser side moves to the platform side.
Language as a wall, not a label
Google built its own examples around Miami and Quebec: markets where language splits the audience in two, and it pays to serve the right ad to each side. Nobody has written the version with Spanish, Portuguese and English coexisting in the same region yet: Brazil running separately from the rest of LATAM, campaigns aimed at US-Hispanic audiences, border markets where the same product sells in two languages at once. The region was never a homogeneous block, something retail media fragmentation in LATAM already made plain: tools designed for specific bilingual markets do not always anticipate three languages living inside one account.
In many of those accounts, language targeting was never just an eligibility filter, it was a full segmentation tool: same product, two campaigns, two budgets, two performance readings. The language split was, in practice, the market split. Once that split stops being guaranteed by a setting and starts depending on how well the system matches the ad's language to the user's, the question worth asking is not whether Google removed a button, but which accounts used it as a wall (hard control over who sees what) and which as a label (a way to name campaigns already split by audience, geo or product).
What is left as the only signal
With the setting gone, what each advertiser has left to influence matching is the actual language of the copy and the landing page. That was already true before, but there used to be a safety net behind it: if the copy had any ambiguity, the campaign-level setting still narrowed who the ad reached. That net is gone now. A Spanish-language ad pointing at an English-language landing page, more common than it sounds in accounts that grew fast, now depends entirely on the system resolving that mismatch correctly.
Before late September, it is worth checking, case by case, that every ad and its landing page match on language, especially where language was standing in for market segmentation. And there is a detail worth not underestimating here: a campaign named "ES" in the panel does not stay Spanish-only just because the name says so. It is a label someone set a year ago. The actual behavior, after September, is defined by the language of the copy and the landing page, not the campaign name.
Reporting across a before and after
There is a reporting consequence worth anticipating. If a Spanish campaign and an English campaign used to get compared on the assumption that each one reliably captured traffic in its own language, that comparability breaks the moment the split stops being a setting and becomes a probabilistic outcome of matching. Comparing September against August on those accounts is going to mix a real performance change with a change in how the sample gets built. The pattern, an operational decision that depended on a manual control and now depends on a model, is the same one reshaping the work inside paid media agencies.
Before the change lands, it makes sense to export the current segmentation (which campaigns had which language configured, what volume each drove under that setup), not because that data will be useful to revert anything, but because it will be the only clean comparison point once the setting no longer exists. It is also worth reviewing whether language there was acting as a real filter or whether the market split was already being held up by other variables (geo, audiences, extensions) and language was only reinforcing it.
What is still unknown
Google has not given an exact date within September, or whether the change lands on the same day across all markets or in a staggered rollout. It also has not published anything about Google Ads Editor, bulk upload templates, or the API, beyond the general guidance to stop setting language criteria on new Search campaigns. And it has not explained why the change is landing only now: the signal had surfaced months earlier, and it ended up shipping nearly nine months later, with no public reason given for the gap. For anyone managing accounts through scripts or integrations, that documentation gap runs into what is already shifting across the ads platforms that speak MCP.