The short answer
ChatGPT Ads reached a $1 billion annualized ad revenue run rate within 200 days of launch, is now live in more than 60 countries, and travel has been an allowed advertising category since day one — Expedia was among the very first buyers in the February 2026 US pilot (OpenAI, August 31, 2026; Adweek, February 19, 2026). For hotels, tour operators and DMCs this creates a new dual-contest reality inside one chat window: the organic answer — which property ChatGPT recommends unprompted — is decided by generative engine optimization, while the labeled ad unit placed beneath that answer is sold to whoever bids, and OpenAI states ads do not influence the answer above them. Your hotel can be the one ChatGPT names while an OTA pays to sit directly underneath that recommendation, capturing the click on your own inventory.
But the demand-side number every budget owner should hold onto is this: Booking Holdings’ CFO told analysts on the Q2 2026 earnings call that traffic referred from large language models — paid and unpaid combined — remains below 1% of the company’s total room nights, and has not moved meaningfully in recent quarters even as its brands appear frequently inside AI answers (PhocusWire, August 2026). Research behavior is shifting to AI platforms fast; attributable bookings are not moving at the same pace yet. The rational Q4 2026 move for most properties is therefore sequenced: secure organic AI visibility first, test paid narrowly with hard guardrails, and measure both. A structured travel AI audit tells you where you stand in the organic contest before you spend a euro on the paid one.
What actually shipped: from pilot to $1B in 200 days
The speed of this rollout is the story. ChatGPT Ads launched as a US pilot on February 9, 2026, restricted to logged-in adult users on the free tier and the low-cost Go plan — subscribers on Plus, Pro, Business and Enterprise plans still see no ads at all (OpenAI, January 16, 2026, updated March 26, 2026). Within weeks the pilot extended to Canada, Australia and New Zealand; by mid-year to the UK, Mexico, Brazil, Japan and South Korea. Then came two inflection points in rapid succession:
- August 24, 2026 — 31 European countries added in a single expansion, the largest to date, taking the platform past 40 markets (OpenAI, August 18, 2026).
- August 31, 2026 — OpenAI reported a $1 billion annualized ad revenue run rate within 200 days of launch and opened self-serve Ads Manager access across India, Europe, the Middle East and North Africa (OpenAI, “A milestone in expanding access to AI,” August 31, 2026).
September then delivered two more milestones relevant to travel sellers. On September 16, 2026, OpenAI’s “Reimagining advertising with AI” announcement introduced Sponsored Agents — an ad format where clicking the ad opens a separate, labeled conversation with a brand-backed agent — plus a natural-language campaign builder in Ads Manager, AI creative tools, and first CRM/e-commerce integrations with HubSpot and Shopify (OpenAI, September 16, 2026; PPC Land, September 16, 2026). And on September 23, 2026, seven more Asian markets came online — Indonesia, Malaysia, the Philippines, Singapore, Thailand, Vietnam and Taiwan — pushing availability past 60 countries (Unite.AI, September 23, 2026).
For a hotel group in Milan, a DMC in Bangkok or a tour operator in Lisbon, this crossed a threshold: ChatGPT Ads is no longer an American curiosity. It is addressable inventory with global reach, self-serve buying in 50+ countries, and conversion-based bidding and geo-targeting added since the pilot.
The two contests: why an ad under the answer is not the answer
The mechanism matters more than the milestone. A ChatGPT ad renders as a small, clearly labeled unit placed underneath ChatGPT’s own answer, selected using the context of the live conversation — and, depending on market and user settings, broader context from that person’s use of ChatGPT. OpenAI has stated repeatedly that advertising does not influence the answer ChatGPT gives, and that advertisers never get access to user conversations (OpenAI ad policies, updated August 31, 2026).
The practical consequence: the brand ChatGPT recommends organically and the brand in the paid unit below it are two separate contests decided by two separate mechanisms. A property can win the recommendation and lose the click, because the OTA listing that same property’s rooms can buy the placement directly beneath the endorsement — on inventory that rates your brand name as prime ad context.
And travel was never gated. OpenAI’s category policy allowed travel and experiences from the start of the pilot (alongside household goods, local services and digital products), while financial services, healthcare and legal required case-by-case approval and gambling and political ads are excluded outright — casino resorts may advertise lodging and entertainment, not the casino floor (OpenAI ad policies, 2026). Expedia’s presence in the original February pilot, confirmed by an OpenAI spokesperson alongside Best Buy and Qualcomm, tells you at least one major intermediary priced this inventory as worth buying from week one (Adweek, February 19, 2026).
The attribution gap: 33% research share, under 1% of room nights
Here is the tension that should shape your budget line. On the research side, Phocuswright found that general search engines’ share as the most-used trip-research tool among US travelers fell from 51% in late 2024 to 36% in the second half of 2025, while generative AI platforms rose to 33% usage for trip research over the same period (Phocuswright, “The AI Surge,” 2026). Adobe’s panel tells the same story in traffic: AI-source visits to US travel sites grew 194% year over year in May 2026 — a record — and are up more than 2,200% since Adobe began tracking in October 2024 (Adobe, June 2026).
On the transaction side, the only large disclosed number runs the other way: Booking Holdings — whose brands appear constantly inside AI answers — says LLM-referred traffic, paid and unpaid combined, is below 1% of room nights and essentially flat (PhocusWire, August 2026). OpenAI is building the measurement stack to close that gap: a tracking pixel, a conversions API, and third-party measurement integrations, with early advertiser results cited as evidence the tools are starting to work (OpenAI, August–September 2026).
Both things are true at once. The channel is real, fast-growing, and cheap to test relative to what it may cost in a year. The attributable-booking evidence for material hotel spend does not exist yet. That is not a reason to ignore ChatGPT Ads; it is a reason to sequence.
Paid placement vs. organic recommendation: what each actually buys you
| Dimension | ChatGPT ad (paid) | Organic AI recommendation (GEO) |
|---|---|---|
| How you win | Auction: bid, creative, context match | Machine-legible reputation: ratings, reviews, structured data, citations |
| Where it appears | Labeled unit below the answer | Inside the answer itself, as the recommendation |
| Per-transaction cost | Pay per click / conversion, every time | Zero marginal cost per recommendation |
| Who can compete | Anyone with budget — including OTAs listing your rooms | Only the properties the model trusts and can verify |
| Reach | Free + Go tiers only (~part of users) | All users, all tiers, including subscribers |
| Durability | Stops the moment spend stops | Compounds; also feeds Gemini, Perplexity, Google AI Mode |
| Evidence base | $1B run rate, early attribution tools | Booking: brands visible in answers while referrals stay <1% of room nights |
The strategic read for hotels: the organic answer is the asset; the ad is a toll booth under it. If you lose the organic contest, an ad defends nothing — you are renting visibility next to a recommendation for a competitor. If you win the organic contest, a small paid test can capture demand your own recommendation generates but that would otherwise leak to an OTA’s ad. This is why AI visibility for hotels is the prerequisite investment, not the optional one — and why defending your brand and destination queries comes before any prospecting spend.
The Q4 2026 decision framework
1. Audit the organic layer first (weeks 1–2). Run a structured travel AI audit across ChatGPT, Gemini, Perplexity and Google AI Mode for your brand queries, your destination + property-type queries, and your competitor set. The first causal audit of LLM hotel selection (12 models, 60,000+ calls, June 2026) showed guest rating (+31.6 percentage points), price perception (−30.0) and review volume (+8.3) move recommendation probability far more than tactics vendors oversell — so fix the signals the machines actually read before buying media next to them.
2. Defend your brand queries with paid, narrowly (weeks 2–4). The one clearly rational early use of ChatGPT Ads for hotels is brand defense: ensuring that when a traveler asks ChatGPT about your property by name, the clickable unit under the answer leads to your direct booking engine, not an OTA listing at 15–22% commission. Cap it, geo-target your feeder markets, and measure against your commission line — not against clicks.
3. Hold prospecting until attribution proves out. Open-to-all destination campaigns are where budgets go to die in new ad platforms. Keep prospecting spend at zero until your pixel/conversion data — or Booking’s disclosed sub-1% figure moving — gives you a defensible cost per attributable booking. Re-check both before the 2027 budget cycle.
4. Prepare for Sponsored Agents, don’t chase them. The September 16 format — an ad click that opens a conversation with a brand-backed agent — is the first ad unit that converges with agentic booking. When your turn comes, the property that can answer rate, availability and booking questions inside that conversation wins; that readiness work (structured inventory, machine-readable policies) overlaps almost entirely with what a travel GEO agency setup already builds, and with the MCP readiness checklist for AI-bookable inventory.
5. Measure both contests on one dashboard. Track AI referral traffic and its conversion rate alongside paid metrics — Adobe’s data shows AI-referred visitors now convert materially better than average, and hotel panels have shown roughly double the conversion rate of organic search sessions (Adobe, August 2026; The Hotels Network panel, May–June 2026). If your AI-referred sessions convert at multiples of your blended rate, the organic contest deserves the marginal euro more than the auction does. Pricing for audit-and-visibility work that covers both layers is on the Palmtree pricing page.
What this means for tour operators and DMCs
Tour operators and DMCs face a version of this with one extra wrinkle: your category is as open as hotels’, but your inventory is less standardized, which makes organic AI visibility simultaneously harder and more defensible. Most multi-day tour products still can’t be bought inside any chat — which means the recommendation ChatGPT gives for “small-group food tours in Bologna” or “incentive program Morocco” routes the traveler to whichever operator’s content the model can actually read and verify. Paid placement can intercept that intent; it cannot manufacture trust that structured content, reviews and citations build. Agencies running multi-market programs should treat the audit layer as shared infrastructure across every client property — see how we approach this in the travel GEO agency model.
FAQ
Can hotels advertise on ChatGPT right now? Yes. Travel and experiences has been an allowed category since the February 9, 2026 pilot, and as of September 2026 ChatGPT Ads is live in 60+ countries with self-serve Ads Manager access in 50+ (OpenAI, August–September 2026). Ads reach logged-in adults on Free and Go tiers; paid subscribers see no ads.
Do ChatGPT ads influence which hotel ChatGPT recommends? No, according to OpenAI’s published policy: ads are visually separate, placed below the answer, and do not affect the answer’s content. Advertisers do not receive access to user conversations (OpenAI ad policies, updated August 31, 2026). The recommendation and the ad are separate contests — which is why organic AI visibility remains the primary asset.
How big is ChatGPT’s ad business? OpenAI reported a $1 billion annualized ad revenue run rate within 200 days of the February 2026 launch (OpenAI, August 31, 2026) — faster scaling than almost any digital ad format before it.
Are travelers actually booking hotels through AI? Research behavior has shifted sharply — AI platforms reached 33% usage for US trip research in H2 2025 while general search fell to 36% (Phocuswright, 2026), and AI-source traffic to US travel sites was up 194% year over year in May 2026 (Adobe, 2026). But attributable bookings lag: Booking Holdings says LLM-referred traffic remains below 1% of its room nights (PhocusWire, August 2026).
Should a hotel run ChatGPT ads in Q4 2026? For most properties: audit organic AI visibility first, then run a capped brand-defense campaign that keeps your own booking engine clickable under answers about your property. Hold broad prospecting spend until conversion tracking shows a defensible cost per attributable booking.
What are Sponsored Agents? Announced September 16, 2026: ad units that open a labeled conversation with a brand-sponsored agent when clicked, alongside AI campaign-building tools and HubSpot/Shopify integrations (OpenAI, September 16, 2026). For travel, it is the first visible bridge between advertising and agentic booking.
The bottom line
ChatGPT Ads is the fastest-scaling ad surface in the market, and OTAs moved on it within weeks of launch. But the paid unit sits under an answer your property can win for free — and the evidence says the organic recommendation, not the ad beneath it, is still what decides which hotels travelers shortlist. Win the answer first. Then decide whether the toll booth is worth renting. If you want to know where you currently stand in that organic contest across ChatGPT, Gemini, Perplexity and Google AI Mode, that is exactly what a Palmtree travel AI audit measures — and what our AI visibility program for hotels is built to improve. Related reading: what actually moves AI to recommend your hotel, the agentic booking stack for hotels, and how AI referral traffic converts into direct bookings.