Ribera's Big Tech meetings put AI competition on the agenda—not a ruling

The European Union's antitrust chief held high-level meetings with CEOs from major tech firms, including Google, Meta, OpenAI, and Amazon, to discuss AI-related antitrust, data use, and market dominance issues. These talks reflect Europe’s proactive stance in addressing potential risks associated with the rapid expansion of generative AI technologies.

• A European Commission agenda showed Teresa Ribera scheduled to meet Sundar Pichai, Mark Zuckerberg and Sam Altman on 24 March 2026, followed by Amazon chief Andy Jassy. The report announced meetings; it did not announce a case, allegation, commitment or remedy.
• AI competition can involve traditional antitrust, merger control or service-specific Digital Markets Act duties. The AI Act addresses safety and fundamental rights. A concern discussed in a meeting is not a finding under any of those regimes.

The reported event was a set of first meetings

The Reuters report carried by Investing.com said EU competition chief Teresa Ribera would hold her first meetings with Alphabet's Sundar Pichai, Meta's Mark Zuckerberg and OpenAI's Sam Altman in San Francisco on Tuesday, then meet Amazon's Andy Jassy on Wednesday. The evidence cited was the Commission agenda. It did not report minutes, requests for information, a case number or a decision arising from the talks.

Reuters added that Ribera had separately launched investigations into practices by Google and Meta and had said she was examining the AI stack, including chatbots, training data and cloud infrastructure. Those existing matters explain the regulatory context. They do not turn every company at the meetings into a respondent in one common AI case.

The competition concern spans inputs and distribution

The Commission's 2024 generative-AI competition brief maps a value chain covering chips, cloud infrastructure, data licensing, models and downstream applications. It identifies possible bottlenecks, vertical integration, privileged access to compute or distribution, tying, self-preferencing and restrictions on rivals as issues that may warrant scrutiny. These are theories and market tendencies; liability requires facts, a legal basis and procedure in a concrete case.

That distinction matters for OpenAI. Alphabet, Amazon and Meta are DMA gatekeepers for specified core platform services, according to the Commission's gatekeeper register. OpenAI is not listed as a gatekeeper. Even for a designated company, a DMA duty attaches to the relevant designated service and obligation—not automatically to every AI model, cloud contract or meeting topic.

Four legal lanes should not be collapsed

Article 101 competition law addresses anticompetitive agreements; Article 102 addresses abuse by a dominant undertaking. Under the Commission's Article 102 procedure, an authority first defines the relevant market and assesses dominance before it can establish abusive conduct. Opening an investigation is therefore not a finding of dominance or infringement.

Merger control examines transactions and structural effects. The DMA imposes ex ante rules on designated gatekeepers and core platform services. The AI Act is a risk-based regime for health, safety and fundamental rights, including rules for high-risk systems and general-purpose models. It is not the legal test for market dominance, foreclosure or an anticompetitive agreement.

A useful example is the Commission's investigation of Google's use of online content for AI. Associated Press reported that it was opened under longstanding competition law rather than the DMA. The investigation examined possible unfair terms and privileged access; it did not establish an infringement when opened. That procedural label is more informative than a generic claim that “AI regulation” is tightening.

A meeting can inform enforcement without being enforcement

Regulatory engagement can help officials understand technical architecture, commercial dependencies and proposed compliance. It can also precede requests for information, market studies or formal proceedings. But without a published legal act, it creates no company-specific finding. The joint statement by EU, UK and US competition authorities sets out general principles on fair dealing, interoperability, choice and scrutiny of control over key inputs; it is a policy signal, not a decision against the four companies.

The old article predicted tougher oversight, changed investment flows, lower entry barriers and global precedent. Those outcomes are possible scenarios, not facts created by the meetings. Enforcement may end with closure, commitments, interim measures, a prohibition, a fine or an appeal. Market effects depend on the precise conduct, service, remedy and judicial review.

What would turn scrutiny into an enforceable event

Watch for published minutes or agenda descriptions, information requests, a named case and legal basis, market definition, an opening decision, preliminary findings or a statement of objections, commitments, interim measures, a final decision and appeals. For the DMA, also identify the designated gatekeeper, core platform service and article at issue.

The defensible conclusion is narrow: Ribera used a US trip to engage leaders across models, platforms and cloud while the Commission was studying competition across the AI stack. The meetings are evidence of attention and information gathering. They are not proof that any entity violated EU law, that a sanction was imminent or that investment and market structure would move in a particular direction.