Google Searches For "ChatGPT Ads" Rose At Least 13-Fold In A Year. Searches For "ChatGPT" Rose About 2-Fold Or Less.

The numbers

Three markets. For each one, each table shows the multiplier for the median method (M1), plus the range across all three methods I ran (M1, M2, M3 - see Methodology).

US

"chatgpt ads" grew at least 18.9x on every method I ran. "chatgpt" grew at most 2.0x.

UK

"chatgpt ads" grew at least 24.0x on every method I ran. "chatgpt" grew at most 1.8x.

DE

"chatgpt ads" grew at least 13.4x on every method I ran (the lowest figure of all three markets). "chatgpt" grew at most 1.5x.

On the lowest of the three methods in each market, "chatgpt ads" still rose 18.9x in the US, 24.0x in the UK and 13.4x in Germany. On the highest, "chatgpt" rose 2.0x, 1.8x and 1.5x.

"google ads" sits in the table as a control: a mature, high-volume query with no obvious reason to move. Its median stayed inside the same volume bucket in all three markets on M1 and M2. On the sum method (M3) it fell 31% in the US and 10% in the UK, and stayed flat in Germany. That argues against a tool-wide rise in high-volume queries, though small queries could still have moved for tool reasons.

What this means

If you own a business: people are typing "chatgpt ads" into Google far more than a year ago. That's a sign of curiosity about the channel, across three markets. Whether it fits your product and margin is something only a test in your own market would show.

If you run ads for a living: the volume is still small next to Google Ads itself. 15,100 median monthly searches for "chatgpt ads" in the US against 550,000 for "google ads" is under 3% of it. What moved is the slope, not the size. A pattern like this, holding across three markets and three counting methods, is worth tracking before it shows up in a client brief.

Neither reading is a service recommendation, and this report doesn't price a pilot or promise a return.

What this does not show

Someone searching "chatgpt ads" might be planning a campaign. They might also be asking whether ChatGPT itself carries ads, looking up a competitor's spend, or researching a blog post like this one. Keyword Planner counts the query, not the intent behind it.

Before February 2026 there was no ads product to search for, so part of the earlier volume in this data is likely people looking for ads made with ChatGPT, or for ChatGPT's own marketing. The comparison mixes both meanings, and this report can't separate them.

Methodology

Three ways of cutting the same 24 months, so the result doesn't depend on one choice of statistic or window:

• M1 (median, Sep-Aug): median monthly volume for September 2024 through August 2025, against September 2025 through August 2026.
• M2 (median, Dec-Aug): the same comparison starting in December, so the three months around the September-October 2025 jump don't pull the medians (see Limitations for what this jump is and isn't).
• M3 (sum, Sep-Aug): the twelve-month total for each window, same boundaries as M1, to check the median isn't hiding a skewed distribution.

The three cuts show the direction holds whichever statistic you pick. Bucket noise stays in all three. Read M2 as a sensitivity check on M1 and M3 (see Limitations).

Keyword Planner reports volume as bucketed ranges. Above about 100 searches a month, one bucket step is roughly 20 to 25%. Below that, steps are as small as 10 searches, so a move from 20 to 30 shows as +50% on paper. The rows with the lowest starting volume - UK "advertise on chatgpt", UK "openai ads", DE "openai ads", DE "chatgpt werbung" - carry the most of that noise.

"chatgpt advertising" returned numbers identical to "chatgpt ads" in the US and the UK (not pulled for Germany). Keyword Planner treats them as the same demand bucket, so this report states the finding once, with this note about the merge, instead of double-counting it.

All queries pulled from one Google Ads account (CID 1649397392), Keyword Planner network, Google Search only. Geo and language: US (geo 2840, English), UK (geo 2826, English), Germany (geo 2276, German). Pull date: 17 September 2026.

Download the raw data (JSON): the full 48-month series behind every number on this page, one query per key, per market.

Limitations

Why this report says nothing about agency searches. I first tried to measure searches for "chatgpt ads agency" and similar phrases; the pattern failed to hold: it broke on exact phrasing, shifted when a close variant like "ppc company" was added, and depended on which twelve-month window was chosen. Underneath that sits a step change in September and October 2025 that hit the agency-related queries I tested, in all three markets, at the same time. That pattern points to a change inside the tool. I'm not publishing conclusions about agency-related searches until that pattern is confirmed against a source outside Keyword Planner.

That same jump sits between the two windows compared here, and "chatgpt ads" moves through it: from August to October 2025 it went from 1,000 to 4,400 in the US, 260 to 1,300 in the UK and 140 to 390 in Germany. A tool change can inflate the multiples in this report, but the rise had started well before the jump: between September 2024 and August 2025, "chatgpt ads" already went from 210 to 1,000 in the US, 30 to 260 in the UK and 20 to 140 in Germany. M2 keeps September to November 2025 out of both medians, but its two windows still sit on either side of the jump, so any tool change shows up in M2 too.

Queries under 100 searches a month carry more bucket noise than the headline numbers, on all three cuts.

Disclosure: I manage Google Ads accounts for clients and offer ChatGPT Ads management, so I have a commercial interest in this channel. This report counts searches and recommends neither service.

Full data

Ratio of query growth to "chatgpt" growth (% change of the query divided by % change of "chatgpt", same method):

Note on M2 ratios: on M2, "chatgpt" itself grows by close to one bucket step in every market (US +22.8%, UK +22.1%, DE +22.1%, all driven by the same single Keyword Planner bucket move). Dividing by a number that small inflates the ratio. The M2 ratios above are a sensitivity artifact of that denominator. None of them appear in the title, key finding or meta tags.

Full monthly series, 48 months per query: download JSON.

FAQ