On 31 August 2026, the European Commission designated ChatGPT a Very Large Online Search Engine under the Digital Services Act. The number that triggered it — 159.1 million average monthly active recipients in the EU — makes ChatGPT the second-largest designated search engine in Europe, ahead of Bing's 119 million and behind only Google Search. The designation does not require ChatGPT to cite or link to anyone; what it does is put an AI answer engine, for the first time, under the same audit and data-access regime as Google.
Most of the coverage framed this as "ChatGPT joins Reddit and Roblox." That comparison hides the more useful one. Reddit and Roblox were designated as platforms. ChatGPT was designated as a search engine — the same legal category as Google Search and Bing — and on the Commission's own register it now sits above Bing.
Key takeaways
- ChatGPT was designated a VLOSE on 31 August 2026 — a Very Large Online Search Engine, not a Very Large Online Platform. Reddit and Roblox were designated VLOPs the same day.
- 159.1 million average monthly active recipients in the EU is the figure on the Commission's register — more than three times the 45 million threshold in Article 33 of the DSA.
- ChatGPT now outranks Bing in the EU by users: Google Search 364 million, ChatGPT 159.1 million, Bing 119 million.
- The deadline is January 2027, not December 2026. The Commission's own notice gives four months from designation. Several outlets reported December.
- Maximum penalty is 6% of total worldwide annual turnover (Article 74). The Commission has issued €870 million in DSA fines to date across three decisions.
What exactly did the European Commission designate?
The Commission designated ChatGPT as a Very Large Online Search Engine — VLOSE — under the Digital Services Act on 31 August 2026. On the same date it designated Reddit and Roblox as Very Large Online Platforms.
The distinction matters and is routinely blurred. A VLOP is a hosting service that disseminates content users upload. A VLOSE is a service that lets users search and returns results. ChatGPT landed in the second category because of its live web-search capability, not because it is an AI product. There is no "AI" category in the DSA; the designation followed the function.
ChatGPT was designated as a VLOSE only. It is not currently on the register as a VLOP.
Is ChatGPT really bigger than Bing in Europe?
By the numbers the Commission publishes, yes. The register of designated services lists three search engines:
| Designated search engine | Provider | Avg. monthly active recipients (EU) |
|---|---|---|
| Google Search | Google Ireland Ltd. | 364 million |
| ChatGPT | OpenAI Ireland Limited | 159.1 million |
| Bing | Microsoft Ireland Operations Limited | 119 million |
Two caveats worth stating plainly. First, these figures are self-declared by the providers and measured over different periods — OpenAI's covers the six months to 31 March 2026. They are not a like-for-like market-share study. Second, "active recipient of a search engine" is not the same as "person running a query with commercial intent." A large share of ChatGPT's 159.1 million are not searching the web at all.
Even with both caveats, the direction is unambiguous. On the EU's own official register, the second-biggest search engine in Europe is a chatbot.
What does a VLOSE designation actually require?
Designation triggers the DSA's heaviest tier of obligations. These go well beyond the baseline rules every online service already follows.
| Obligation | What it means in practice |
|---|---|
| Systemic risk assessment | Assess risks from the service and its algorithmic systems — illegal content, effects on minors, physical and mental well-being, fundamental rights, elections, public security |
| Risk mitigation | Adopt measures that actually reduce the risks identified |
| Independent audit | Be audited by an independent auditor at least once a year, and act on the recommendations |
| Compliance function | Maintain an internal function that ensures identified risks are mitigated |
| Vetted researcher data access | Give vetted researchers access to platform data to detect and understand systemic risks in the EU |
| Ad repository | Maintain a publicly available repository of advertisements |
| Non-profiling recommender option | Under Article 38, offer at least one recommender option not based on profiling |
| Regulatory data sharing | Share data with the Commission and national authorities for compliance monitoring |
Enforcement of this tier sits with the Commission directly, not with national Digital Services Coordinators.
When does OpenAI have to comply?
By January 2027. The Commission's notice gives designated services "four months, i.e. by January 2027, to comply with the additional DSA obligations."
This is worth flagging because a number of outlets reported the deadline as December 2026. Four months from a 31 August designation lands in January, and January is the date the Commission published. If you are tracking this for a compliance or content calendar, use the Commission's date.
What happens if OpenAI doesn't comply?
Article 74(1) of the DSA lets the Commission impose fines "not exceeding 6 % of its total worldwide annual turnover in the preceding financial year."
That ceiling has not been theoretical. The Commission has issued three DSA non-compliance decisions:
| Decision | Date | Fine | Grounds |
|---|---|---|---|
| X | 5 December 2025 | €120 million | Deceptive "blue checkmark" design, ad repository transparency, researcher data access |
| Temu | 28 May 2026 | €200 million | Failure to assess and mitigate risk of illegal products |
| AliExpress | 20 July 2026 | €550 million | Failure to assess and mitigate risk of illegal, unsafe or counterfeit products |
That is €870 million in total, and the trajectory is steeply upward. Note what X was actually fined for: two of the three grounds were transparency failures — the ad repository and researcher data access — not content moderation. Those are precisely the obligations ChatGPT now inherits.
Does this mean ChatGPT has to cite or link to your site?
No. This is the part worth being honest about, because the designation is being sold in some quarters as a win for publishers, and it is not one.
Nothing in the VLOSE regime requires an answer engine to attribute sources, link out, or send referral traffic. The DSA governs systemic risk, transparency and accountability. It does not regulate how a search engine ranks, how it summarises, or whether it drives clicks to the sites it learned from. Google Search has been a designated VLOSE since 2023, and that designation did nothing to slow AI Overviews.
If you are waiting for regulation to restore your referral traffic, this is not it.
The obligation SEOs should actually watch: researcher data access
Here is the part with real second-order consequences, and the reason this designation matters more than a compliance headline.
Under the DSA, a designated service must let vetted researchers access platform data to study systemic risks. Applied to ChatGPT, that opens a door the search industry has never had: the possibility of independent, non-vendor data about how an AI answer engine selects, weights and surfaces sources.
Everything the GEO and AEO industry currently believes about getting cited by AI is inferred from the outside — from prompt testing, log files and correlation studies. No one has ever seen the inside of the selection process. Researcher access is scoped to systemic risk research, not to marketing analytics, so no one is going to publish a ranking-factors list. But research into how an answer engine amplifies or suppresses sources is squarely a systemic-risk question, and that research would be the closest thing to ground truth the field has had.
It is also the obligation X was fined for refusing. Watch whether OpenAI complies quickly or quietly.
What should you actually do about it?
Nothing in this designation changes the mechanics of AI visibility. It does confirm the scale you are optimising for. A useful order of operations:
- Confirm OpenAI's crawlers can reach you. ChatGPT's live search can only surface what it can fetch, and OpenAI runs separate agents for search retrieval and for training — allowing one is not the same as allowing the other. Check what your robots.txt actually permits with the crawler checker.
- Stop treating ChatGPT as a rounding error in reporting. A surface with 159.1 million EU users is not an experimental channel. If your dashboard has no row for it, the dashboard is wrong — see the AI search KPIs worth tracking.
- Write for extraction, not for sessions. Answer engines quote sections, not pages. Clear question-shaped headings and self-contained answers get lifted; narrative build-ups do not.
- Measure brand mentions, not just clicks. The traffic may not arrive even when the citation does. How to get cited by AI covers what actually moves that number.
Frequently asked questions
Is ChatGPT legally a search engine now?
In the European Union, yes — for the purposes of the Digital Services Act. On 31 August 2026 the European Commission designated ChatGPT a Very Large Online Search Engine, the same legal category as Google Search and Bing. The designation applies under the DSA in the EU and does not change ChatGPT's legal status elsewhere.
How many EU users does ChatGPT have?
159.1 million average monthly active recipients in the European Union, according to the Commission's register of designated services. OpenAI's figure covers the six months ending 31 March 2026. The DSA designation threshold in Article 33 is 45 million.
Is ChatGPT bigger than Bing in Europe?
By the figures on the Commission's official register, yes. ChatGPT is listed at 159.1 million average monthly active EU recipients against Bing's 119 million. Google Search remains far ahead at 364 million. These are self-declared figures measured over different periods, not a market-share study.
When must OpenAI comply with the DSA obligations?
By January 2027. The Commission gave designated services four months from the 31 August 2026 designation. Some coverage reported December 2026; the Commission's own notice says January 2027.
What fine can the EU impose on OpenAI?
Up to 6% of total worldwide annual turnover in the preceding financial year, under Article 74(1) of the DSA. The Commission has so far issued €870 million in DSA fines across three decisions: €120 million against X, €200 million against Temu and €550 million against AliExpress.
Does the DSA force ChatGPT to link to publishers?
No. The DSA imposes obligations on systemic risk assessment, auditing, transparency and data access. It does not require an answer engine to cite sources, attribute content or send referral traffic. Google Search has been a designated VLOSE since 2023 and that did not prevent AI Overviews.
Will Gemini, Claude and Perplexity be designated too?
The designation followed a function — live web search at scale — and a self-declared user count crossing 45 million EU recipients. Any AI assistant with a search capability that crosses that threshold and declares it becomes designatable on the same reasoning. None have been designated as of this writing.
Where this leaves you
The EU has now formally recognised what your analytics have been implying for two years: an AI assistant is a search engine, and in Europe it is the second-biggest one there is. The regulation will not send you traffic. It does settle the argument about whether this channel is big enough to plan around.
Start with what you control. Check whether AI crawlers can actually reach your content with the free crawler checker, confirm your pages are describing themselves correctly with the meta tags checker, and if you want the fuller picture of how AI search is reshaping discovery, read AI search visibility: why clicks are falling and brand demand isn't.