OpenAI’s lawsuit against Microsoft – filed on 10 January 2024 in a U.S. federal court – sent shockwaves through the tech world, not least because it pits the world’s leading foundation‑model developer against the cloud giant that powers its most profitable products. The complaint accuses Microsoft of breaching an exclusive‑licensing pact for OpenAI’s models on Azure and of abusing a dominant position in the market for cloud‑based AI services – allegations that echo directly into Europe’s antitrust arena.
Within days, regulators on both sides of the Atlantic began to wonder whether the dispute could become the first high‑profile test of the EU’s new AI‑centric competition toolbox. The filing’s timing coincides with a wave of European legislative activity: the Digital Markets Act (DMA) entered its public consultation in July 2025, a dedicated AI questionnaire followed in August 2025, and the AI Act’s first obligations for general‑purpose models took effect in August 2025, with full implementation slated for early 2026. The confluence of these measures means that any exclusive cloud‑AI arrangement now sits under the simultaneous gaze of Articles 101 and 102 of the TFEU, the DMA’s gatekeeper duties, and the AI Act’s transparency regime.
European competition authorities have a well‑defined playbook for tackling the conduct OpenAI alleges. Under Article 102, the European Commission applies the Bronner test – confirming a dominant position, an exploitative or exclusionary practice, and a detrimental effect on competition. If Microsoft’s alleged exclusivity clause forces Azure users to accept OpenAI’s models on terms that foreclose rival providers, the Commission could launch a formal probe, as it did in the Google advertising self‑preferencing case that resulted in a €2.95 billion fine. Likewise, a cartel‑type arrangement that coordinates pricing or ties services could trigger Article 101 enforcement, reminiscent of the €329 million penalty imposed on Delivery Hero and Glovo for an online‑food‑delivery cartel.
The DMA amplifies the stakes. By designating large platforms that provide essential digital infrastructure as “gatekeepers”, the Act obliges them to grant non‑discriminatory access, ensure algorithmic transparency and uphold data‑portability. Should Azure be classified as a gatekeeper for AI infrastructure, Microsoft would be required to open its cloud‑AI services on fair, objectively justified terms – a breach that can attract periodic penalties of up to 10 % of worldwide turnover. The July 2025 consultation, which drew more than 450 contributions, already called for extending the DMA’s scope to cover AI and cloud services, signalling a regulatory appetite for exactly the kind of exclusivity dispute OpenAI raises.
The AI Act adds a third, technology‑specific layer. Providers of foundation models now must publish detailed summaries of their training data and conduct rigorous risk assessments, while downstream users must avoid prohibited uses such as untargeted facial‑scraping. This regime dovetails with competition scrutiny: the European Commission has warned that “pseudo‑mergers” – where a platform licences a startup’s IP while sidelining merger‑review thresholds – could be examined under both antitrust and AI‑Act provisions. In practice, a court‑ordered remedy could force Microsoft to disclose the algorithmic criteria that govern Azure’s AI offerings and to offer comparable licensing conditions to all AI developers.
Legal scholars see the OpenAI‑Microsoft clash as a litmus test for how the EU’s overlapping frameworks will operate in concert. The “Bronner criteria” provide a clear analytical route for abuse‑of‑dominance claims, while the DMA’s gatekeeper duties supply a procedural hook for ensuring market access. The AI Act’s transparency mandates, meanwhile, create a factual baseline that competition enforcers can rely on when assessing whether a dominant platform is leveraging its position to lock out rivals. In short, the case forces regulators to coordinate across antitrust, digital‑market and AI‑specific law – a coordination that the European Competition Network is already prepared to facilitate.
Beyond the legal mechanics, the dispute feeds a broader strategic narrative about Europe’s digital sovereignty. A January 14 2026 Atlantic Council report framed AI governance, data control and supply‑chain resilience as pillars of an independent European tech ecosystem. By signalling that exclusive AI‑cloud contracts will be subject to rigorous scrutiny, the EU is sending a clear message: lock‑in by a non‑European provider will not be tolerated, and firms must build interoperable, transparent partnerships if they wish to operate on the continent.
In the months ahead, the trajectory of OpenAI’s suit will likely dictate whether European authorities move from watchful observation to formal investigation. If the Commission or national competition agencies deem Microsoft’s conduct to breach Article 102, the DMA, or the AI Act, the penalties could be severe enough to reshape the architecture of AI‑cloud collaborations across the EU. For European start‑ups and mid‑size firms eyeing the lucrative AI market, the outcome will determine whether they can rely on a level playing field or must navigate a landscape dominated by a single, gatekeeping cloud titan.
Image Source: www.dreamstime.com
