On 31 August 2026 the European Commission designated ChatGPT a Very Large Online Search Engine under the Digital Services Act, seven days after OpenAI switched ChatGPT Ads on across 31 European markets. ChatGPT's search engine designation makes official what media buyers already priced in: the AI answer is a search medium. What Brussels classified is the surface where the ad runs, not the supply chain that produces the answer.

  • ChatGPT joins Google Search and Bing as one of only three designated Very Large Online Search Engines in the EU, with compliance due by January 2027.
  • Designation pulls AI answers into the DSA advertising regime: labelled ads, explained targeting, and a public, API-accessible ad repository under Article 39.
  • Brands gain regulator-grade disclosure on the AI answer surface, which makes it defensible as a media line rather than an SEO experiment.
  • Publishers gain nothing from the designation, because the content grounding those answers sits outside the regulated transaction entirely.

What the ChatGPT search engine designation actually obliges

A Very Large Online Search Engine is a search service that reaches at least 45 million average monthly users in the EU and therefore carries the DSA's heaviest obligations. The Commission's 31 August announcement gives ChatGPT, Reddit and Roblox four months, until January 2027, to comply with systemic risk assessment, independent audit and data access duties. Non-compliance exposes up to 6% of global annual turnover.

For anyone buying media, the operative clause is Article 39. Designated services must publish a searchable, API-accessible repository of every paid ad carried on the service: the creative, the advertiser, who paid for it, the main targeting parameters and the reach broken down by member state, retained for the exposure period plus at least one year.

The regulated SERP and the regulated answer

In the open-web economy, regulation followed the money into the SERP. Ads had to be labelled, targeting had to be explained, and the repository made the auction inspectable. Publishers sat inside that same economy through traffic and network revenue, so oversight of the ad layer indirectly protected the people supplying the content.

In the agentic web that loop is cut. The regulated ad layer sits inside ChatGPT's own answer. The content grounding that answer sits on publisher domains that receive neither the impression nor the click. Supervising the ad slot no longer supervises anything a publisher gets paid for.

What Gizmodo reported, and what the record shows

Gizmodo's 31 August piece is accurate on substance and needs one correction of emphasis. It presents designation as automatic above 45 million monthly users; the Commission's own text says the services declared reaching that threshold, which makes self-reporting the trigger and Commission supervision the consequence. Politico reported the Commission hesitated between platform and search engine, and OpenAI has since stated publicly that ChatGPT search operates as a search service under the DSA, which is the same reading.

Why the antitrust route will not produce a publisher price

The strongest case for regulation as the compensation mechanism

The serious counter-argument is that European enforcement is already converging on payment. The Commission opened a formal antitrust investigation in December 2025 into Google's use of publisher content for AI. The European Publishers Council filed an Article 102 complaint on 10 February 2026 demanding control, transparency and a fair remuneration framework. France's APIG followed in August 2026, citing the national regulator's estimate that AI summaries account for 33% to 38% of publisher traffic losses in markets where they run.

Why classification and disclosure are not pricing

Each of those instruments allocates rights, not rates. Antitrust remedies arrive as settlements years after the audience is gone, and they attach to one dominant defendant rather than to a market. The DSA is blunter still: it makes advertising inside the AI answer inspectable and says nothing about who supplied the evidence the answer was built from. A regime can force full disclosure of what a brand paid ChatGPT without ever creating an invoice a publisher can send.

What this means for brands and for publishers

For CMOs, media buyers and agencies: the AI answer is now a nameable media line

Designation removes the last procurement excuse. The surface carrying your brand into AI answers is legally a search engine, subject to ad labelling and to a public repository you will be able to query from January 2027. Treat it as search inventory: name the budget line, set a cost per mention, and use the repository as competitive intelligence on what rivals pay to sit beside the answer.

For publishers: compliance visibility is not revenue visibility

Your content will ground answers on a service Brussels now supervises, and nothing in that supervision routes a cent to you. Licensing and litigation are worth pursuing, but budget them as one-off recoveries rather than as run rate. Recurring revenue comes from being paid as inventory in the answer economy, which means selling placement on the pages AI engines cite instead of waiting for a remedy.

Three signals to watch before January 2027

First, the shape of ChatGPT's Article 39 repository, which becomes the first public record of what advertisers pay to appear inside AI answers in Europe. Second, whether the Commission's Google case produces a remuneration principle or another behavioural remedy. Third, whether any AI engine voluntarily discloses payments to cited sources, because the first to publish that number sets the benchmark every other engine gets measured against.

Conclusion

Hold on to this: ChatGPT's search engine designation gives the AI answer a legal identity and an advertising rulebook, while leaving the publishers who supply the answer entirely outside the transaction. Regulation classifies value; only a market prices it, which is why the missing layer is Generative Engine Advertising, native ads placed on the publisher pages AI engines cite, buyable by brands and paid through to the media sources feeding the answer. That is the category Smalk AI is building on both sides of the market. Watch the first Article 39 repository entries in January 2027, because the moment Europe publishes what brands pay to appear in AI answers, the absence of a publisher price becomes impossible to defend.