---
title: "Paying to reach curious readers in 2026: ChatGPT ads, feed ads, paid newsletter recommendations, and the labelling rules"
kind: "Article"
topic: ""
date: "2026-10-11"
url: "https://canopticon.com/articles/paying-to-reach-curious-readers-in-2026"
author: "Chris Telford"
license: "CC BY 4.0"
---
Advertising inside AI assistants went from announcement to a working market within this year. It is a real place to buy attention, but not yet a cheap or well-measured one, and the older channels still undercut it. Reading the prices side by side, and the rules that now govern what an ad may pretend to be, shows where a curious reader can actually be bought, and on what terms.

## ChatGPT became an ad platform in 2026, with thin inventory

The timeline is short. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="TL;DR">OpenAI announced them Jan 16, 2026, and they went live Feb 9, 2026.</q> Buying started expensive and closed: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Ads launched with a $60 CPM and a reported six-figure minimum.</q> CPM is the price of impressions, the standard unit for ad space. Within months the door opened: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="TL;DR">Self-serve opened May 5 with no minimum spend.</q> By the end of summer the business was large: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="TL;DR">ChatGPT ads reached a $1B annualized run rate "in less than 200 days after launch," according to OpenAI's August 31, 2026 announcement.</q>

The format is unlike a search ad. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="TL;DR">The ads are sponsored cards below answers, targeted by plain-language "context hints" rather than keywords.</q> The rules around them are strict: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">No political ads. Health and finance outside the US generally prohibited. No ads in Temporary Chats, to logged-out users, to under-18s or to paid tiers.</q> And the early experience was rough: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Early advertisers complained they couldn't prove results or even spend their budgets, because inventory was thin.</q>

The audience is mostly not shopping. OpenAI's head of global ads solutions said that <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">roughly 20% of queries in ChatGPT have a direct commercial intent</q>, which the research reads plainly: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">So most conversations are curiosity, not shopping.</q> For a Canadian buyer there is an extra wrinkle: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="TL;DR">Canada has had ads since spring 2026, but whether self-serve buying works for a Canadian account is reported inconsistently</q>.

## Perplexity left advertising; Google and Microsoft are moving in

Not every AI answer engine wants the money. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Perplexity stopped taking advertisers in October 2025 and confirmed in February 2026 that it was out.</q> An executive explained why to the Financial Times: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">A user needs to believe this is the best possible answer... The challenge with ads is that a user would just start doubting everything.</q>

The larger companies made the opposite call. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Google (AI Mode formats announced May 20, 2026, testing in the US) and Microsoft (Copilot ads via Microsoft Advertising; AI Max for Search pilot April 21, 2026) are going the other way.</q> Their formats are closer to search advertising than to content, and the performance claims are their own: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Microsoft claims 73% higher CTR than traditional search (first-party, unverified)</q>

## What a curious reader costs on each platform

Against the feeds, the assistant slot is expensive. The research's paid inventory map sets the prices side by side:

| Platform | Typical cost | Fit for curious, idea-dense audiences |
|---|---|---|
| ChatGPT ads | $25–60 CPM (default max $60); $3–5 recommended CPC; oCPC available. Minimum daily budget reported at $25 USD for US accounts (vendor source). | High intent, thin inventory, premium price |
| Meta (FB/IG/Threads) | ~$6.59 CPM avg (Gupta, Oct 2025); most accounts $10–18 | Broad; curiosity pockets via interests and lookalikes |
| Reddit | ~$3–12 CPM; ~$1.25 median CPC (agency data) | Very high for technical and curious users |
| X | Lowest CPMs among majors (Emplifi) | Mixed; brand-safety concerns |
| TikTok | ~$4–7 CPM; campaign minimums apply | Young, video-native |
| LinkedIn | ~$31–65 CPM | Professional, expensive |
| Beehiiv Boosts | $1.63 avg (early access); $50 minimum; $2–3 typical (third-party) | High: people already read newsletters |
| SparkLoop Partner Network | ~$4–7 per subscriber | High; better early opens |

*Rows copied from the research’s paid inventory map, late September 2026. Most figures are agency or vendor benchmarks.[Details](doc:content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s)*

Meta is the reference point: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Gupta Media's tracker put Meta at a $6.59 CPM and $0.68 per link click (October 2025).</q> On the same measure <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">ChatGPT's $25–60 CPM puts it in LinkedIn territory.</q> Reddit sits at the cheap end with an audience that suits ideas: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Reddit reportedly runs a $3–12 CPM with a median CPC around $1.25 (agency data), with an older, more technical audience.</q> Its users are also the least forgiving of advertising that reads like advertising: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Promoted posts must read like native posts. Community tone is unforgiving; give the full answer in the post.</q>

Prices move through the year, and the figures are not audited. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">CPM seasonality: Q4 spikes of up to about 66%.</q> <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Caveats">Most cost benchmarks come from agencies or vendors (Gupta, Stackmatix, Newsletrix, SparkLoop, Billo). Treat them as ranges, not promises. ChatGPT ad pricing changes monthly.</q>

## For a newsletter, the measure is a reader who keeps opening it

Impressions and clicks matter less to a newsletter than the price of a subscriber who reads. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">One operator acquired engaged subscribers on Meta at $0.65–$1.60.</q> <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Another reports large B2C newsletters happily paying $3–$5 at scale, while some clients need under $1.25 to break even on day one.</q>

The other route is to pay other newsletters to recommend yours. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Beehiiv lists a $1.63 average cost per subscriber for Boosts (early-access data) and a $50 minimum.</q> <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">A vendor analysis puts SparkLoop at $4–7, but with 35–45% first-30-day open rates versus 22–30% for Meta lead ads.</q> Recommendations can also pay the other way, which is how some publishers offset their ad spend: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">One customer says "we're actually making more money from paid recs than we're spending on ads."</q> <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">That's a vendor testimonial, but the mechanism is straightforward.</q>

The measurement trap is the cheap signup. The research's rule: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Judge by cost per engaged subscriber, not raw signups.</q> The reason: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">One operator found the cheapest signups were people with no interest in newsletters, and that existing subscribers clicking ads made CPA look artificially low.</q> Hard numbers are scarce in any case: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Public, audited payback numbers from media companies are rare. Most figures come from operators and vendors.</q>

## An ad that gives the whole answer selects its readers

The common practice is to pay only for what has already worked. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">This is the "test free, pay for proven" loop newsletter operators use.</q> The research's argument for making the ad itself useful is about who it attracts. Giving the complete answer in the feed, rather than teasing it, <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">reverses the curiosity-gap clickbait pattern</q>, and <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">People who subscribe after being fully satisfied are self-selecting for the product, which is why their open rates hold.</q> That is reasoning, not a measured result. The honest version of a reason to come back is a promise with a date: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Clickbait withholds; open loops promise a dated resolution and deliver it.</q>

The economics are tight either way. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="TL;DR">The loop breaks when revenue per subscriber (often only $1–4 a year from sponsorship alone) can't repay CAC before churn.</q> CAC is what it cost to acquire the reader. Sponsorship needs a list of some size before it pays at all: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Sponsorship also requires scale: sponsors rarely buy lists under a few thousand subscribers.</q> Readers bought with ads tend to leave sooner: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Paid cohorts churn faster than organic ones.</q> though the research flags that rule as a heuristic rather than a study. Paid tiers convert slowly and lose readers steadily: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">0.62% median free→paid conversion (beehiiv)</q> and <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">beehiiv reports monthly churn of 5.06% to 16.67% depending on vertical.</q>

## The 31 percent figure measures minutes of scrolling, not people

A common shorthand says a large share of scrollers have attention problems. The study behind it measured something narrower: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Allcott, Gentzkow and Song ("Digital Addiction," AER 2022; about 2,000 US adults, 2020)</q> estimate that self-control problems <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">account for an average of 31 percent (48 minutes per day)</q> of social-media time. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">It is a share of minutes, not a share of people, and it isn't a clinical attention diagnosis.</q> The effect is also uneven: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">for 22% of participants it adds less than 10 minutes a day</q>. For anyone buying attention, the distinction is the difference between reaching curious people and harvesting compulsive minutes.

## The label stays; the rules target fake people and hidden sponsors

Every paid placement carries its label, and the research treats that as no obstacle to useful content. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Deception starts when you hide who's speaking or why: fake users, fake reviews, undisclosed paid relationships.</q> In the United States the FTC's rule <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">bans reviews and testimonials by people who don't exist, "such as AI-generated fake reviews."</q> <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">It also covers fake social-media indicators such as bought followers or views.</q> <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Civil penalties run up to $53,088 per violation.</q> Enforcement has begun: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">The FTC sent its first warning letters to 10 companies on Dec 22, 2025.</q>

Canada's rule is framed around disclosure: under the Competition Act <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">anyone promoting a business interest must disclose material connections "as visible as possible," and reviews must reflect real experience.</q> The advertising standards body extends that to synthetic presenters: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">An AI influencer still has to disclose brand connections and can't give testimonials about experiences it can't have.</q> Europe's rules reach the same point by another route: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Deepfake audio and video must be disclosed as AI-generated when shown to EU users.</q>

The platforms enforce their own versions. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Meta auto-labels AI-edited ads with "AI info."</q> <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">OpenAI prohibits ads that imitate ChatGPT's interface or voice, or use false endorsements.</q> Readers want the disclosure even when they cannot spot the difference. In an Ipsos test, <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">only 25% could confidently tell which ads were AI-made, but 79% said companies should disclose AI use</q>. Canadian results run the same way: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">In Canada, the Ad Standards/Caddle survey found 78% say AI influencers must disclose their identity, and 57% say AI reduces their likelihood to buy.</q> Not everyone minds: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Klaviyo's March 2026 survey of 7,998 consumers in 8 countries found that 32% say brand use of AI-generated content makes them trust brands less, while 61% are neutral.</q>

> **Objection: Paying to send readers to your own pages is attention arbitrage**
>
> The arbitrage the industry worries about has a name and a size. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Made-for-advertising (MFA) sites buy cheap clickbait traffic and resell it as ad impressions.</q> In one audit <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">MFA sites made up 21% of impressions and 15% of ad spend</q>, and <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Industry groups define MFA partly by "little-to-no organic audience" and dependence on paid clickbait traffic.</q>
>
> The test that separates them is whether readers come back when nobody pays them to.

## Synthetic crowds work as test rigs, not as audiences

Cheap AI-made content invites a shortcut: fill the room with synthetic people. The experiments so far argue against it. The agent-only network Moltbook was a sharp lesson: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">It went viral partly on posts that turned out to be humans posing as agents through an insecure API.</q> A research version built to study such networks found they reproduce the worst of human ones. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">With only posting, reposting and following, it reproduced echo chambers, elite concentration and amplification of polarized voices, and the fixes they tested helped only modestly.</q>

Feeding machines their own output degrades them as well. A study in Nature found that training on recursively generated data <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">causes irreversible defects… tails of the original content distribution disappear.</q> The research draws the consequence: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">An AI-only feed homogenizes, and human-origin data becomes more valuable.</q> Synthetic panels can still help rank ideas before money is spent. <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">Argyle et al. report correlations around 0.90–0.94 on ANES items</q>, but <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Details">they show less variance than real people, and regression relationships can differ or flip sign.</q> The research's verdict: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Key findings">Synthetic societies are useful as mirrors and test rigs, not as fake crowds.</q>

## AI-assistant ads sell a reader mid-question, at a premium

It explains why the new ad market has not displaced the old one. What an assistant sells is a reader caught in the middle of a question, and it charges accordingly, with thin inventory and weak measurement while it learns. The cheap curious readers are still in the feeds and in other newsletters' recommendations, and the cost that matters is the one for a reader who keeps opening. The research's own advice for a small publisher follows from the prices: <q doc="content-that-is-the-ad-buying-curious-attention-in-late-2026-and-the-s" section="Bottom line">Use them as a small, high-intent probe, not the main channel.</q>

The rest of the finding is about honesty rather than price. An ad can be the whole of something worth reading and still say it is an ad. The rules, the platforms and the readers all draw the line in the same place: at fake people, invented testimony and hidden sponsors, not at ads that are useful.