The implications of AI are crucial for merger control enforcement, emphasized Margrethe Vestager, the European Union’s antitrust chief and digital EVP, during a recent seminar. She cautioned that “wide-reaching” digital markets could result in unforeseen economic consequences. Her comments were directed at preventing tech giants like Microsoft, Google, and Meta from establishing monopolies in the AI space, signaling an intent for more rigorous scrutiny over their operations.
“We must carefully examine vertical integration and ecosystems, considering AI’s impact in our merger assessments. We also need to be aware of how AI could foster new forms of algorithmic collusion,” Vestager stated.
These remarks indicate that the EU is poised to intensify its evaluations of technology mergers and cozy AI partnerships. Recently, the EU launched an investigation into whether Microsoft’s investment in generative AI powerhouse OpenAI warrants a merger review.
Additionally, Vestager highlighted that the challenges of competition are intrinsic to the development of cutting-edge AI, pointing out “barriers to entry everywhere.” She explained, “Large Language Models (LLMs) require vast data, cloud resources, and specialized chips. Tech giants possess the means to attract top talent, which sets a high barrier for newcomers. Disruption in AI will likely stem from established tech ecosystems rather than from upstarts.”
The rapid ascent of generative AI has spotlighted how innovations are often controlled by a few companies intimately connected to major tech platforms. Examples include OpenAI's strategic partnership with Microsoft, investments from Google and Amazon into OpenAI competitor Anthropic, and Meta’s use of its extensive social media data to develop its foundational AI models (known as LLaMA).
During the seminar, participants raised concerns about how European AI startups can compete when they lack access to similar AI infrastructure.
Challenges and Uncertainties
Tobias Haar, general counsel for German AI startup Aleph Alpha, questioned whether LLaMA 3 would also be open-sourced, expressing concerns over the future availability of open-source LLMs. Haar emphasized the necessity for his startup to build its own foundational models within an independent data center to ensure sustainability.
Aleph Alpha operates Europe’s largest commercial AI cluster with 512 A100 Nvidia GPUs. However, Haar pointed out that this is still insignificant compared to the computing capacity of Big Tech, citing Microsoft's announcement of a plan to deploy approximately 10,000 GPUs in the UK as part of a £25 billion investment.
“In context, our 512 A100 GPUs afford us some independence, but they are mere fractions of what major organizations possess to train and refine their LLMs. OpenAI may leverage Microsoft’s resources to further enhance its models, creating a disparity in the competitive environment,” he remarked.
While Vestager's presentation did not outline specific strategies to level the playing field for European generative AI startups, it suggests the EU is preparing for a tougher stance on scrutinizing tech mergers due to emerging AI challenges.
In the past, Vestager approved Google’s acquisition of Fitbit under specific commitments, allowing Google to utilize data for purposes other than advertising. With recent advancements in AI, the era of Big Tech cherry-picking acquisition targets may soon be coming to an end.
Vestager implied that the EU plans to utilize existing competition tools, such as the Digital Markets Act (DMA), to shape AI market dynamics. Although there have been questions regarding the DMA's applicability to generative AI, as no cloud services have been classified under “core platform services,” she asserted that the time for decisive action is still at hand.
In her speech, Vestager conveyed a sense of urgency, stating that the “window of opportunity” to achieve beneficial outcomes for the economy, citizens, and democracies is narrowing. However, her address raised more questions than answers regarding how to address the complexities of AI, including concerns over democratic integrity, intellectual property, and ethical considerations.
“There are significant issues regarding the respect for intellectual property rights and the ethical deployment of AI. In all these considerations, competition policies play a role. Moreover, the enforcement of AI regulations may impact market openness,” she stated, indicating potential conflicts between regulating AI risks and fostering a vibrant ecosystem.
Despite her call for swift action and collaboration, the overall tone of Vestager's comments suggested managing expectations rather than offering immediate solutions for curtailing Big Tech's ambitions in AI.
Solutions
Discussion became heated when Barry Lynn from the Open Markets Institute proposed fundamental changes, suggesting the separation of cloud services from core tech operations. “Transforming cloud into a utility is one of the simplest solutions to reduce their leverage,” he asserted.
Lynn also advocated for a non-discrimination policy for platforms to prevent unfair pricing and data manipulation, emphasizing that the aggregated data collected from users should belong to the public rather than tech conglomerates like Google.
Microsoft's director of competition, Carel Maske, who was present during the panel discussion, appeared unsettled by Lynn's bold suggestions. He argued that structural separation could undermine necessary investments in cloud infrastructure, steering the conversation back to discussions of effective competition tools rather than structural reforms.
Another panelist, Andreas Mundt of Germany’s Federal Cartel Office, reflected on the challenges faced by regulators, noting the difficulties they've encountered while assessing the Microsoft-OpenAI partnership. Although they evaluated the collaboration for merger control, they ultimately decided it did not meet the necessary criteria, despite recognizing its competitive implications.
Mundt expressed his desire for a more decisive outcome regarding this significant partnership and voiced his satisfaction that the European Commission is now investigating the situation further, alongside similar inquiries by the UK's competition authority.
Mundt assured attendees that the authorities would closely scrutinize future collaborations between tech giants to ensure compliance with merger regulations. Vestager emphasized the importance of teamwork among competition and regulatory bodies worldwide to tackle the multifaceted challenges posed by AI's evolution.
“Collaboration is our strongest asset. By working together and acting early, we can maximize our impact," she concluded, urging stakeholders to participate in ongoing consultations regarding generative AI and virtual environments.
Antitrust regulators worldwide are racing to develop strategies for effectively addressing the challenges posed by AI technologies.
Ranking
