Are Big Tech Firms Quietly Monopolising AI?

Big Tech companies like Meta, Microsoft, Amazon, Google, and Nvidia are making strategic investments and forming exclusive partnerships in the AI sector, stopping just short of formal acquisitions to avoid antitrust scrutiny. This approach, termed 'non-acquisition acquisitions,' allows them to secure access to AI models, data pipelines, and research talent.

The tactics used by tech giants echo those of John D. Rockefeller's Standard Oil, which involved silent partnerships and intermediaries to absorb competitors. This led to the breakup of Standard Oil under the Sherman Antitrust Act.

Meta's $14.8 billion investment in Scale AI, framed as a 49% non-voting stake, grants Meta exclusive access to Scale’s data pipeline and talent. This deal structure raises concerns about de facto control, input foreclosure, and talent consolidation, potentially drawing FTC scrutiny.

Microsoft's $13 billion investment in OpenAI secures exclusive commercial rights to its models, integrated into Azure, without formal equity or voting control. The FTC is investigating whether this constitutes a de facto acquisition. Conflicts have arisen due to OpenAI's unique structure, affecting Microsoft's influence.

Amazon and Google have invested billions in Anthropic, hosting Claude on their cloud platforms. Nvidia invests in AI startups, linking investments to hardware access and support, creating a self-reinforcing loop that raises conflict of interest concerns.

Meta is reportedly in talks to hire AI investors Nat Friedman and Daniel Gross, potentially buying out their venture capital fund. This 'acquihire' strategy, gaining talent and stakes in AI startups, is under scrutiny.

Traditional antitrust enforcement focuses on ownership and control, but influence can be exerted through contracts and infrastructure. Regulators are adopting a 'substance over form' approach, scrutinizing the economic reality of deals. Guidelines address minority stakes, resource blocking, talent hiring, shared board members, and dominance entrenchment.

Despite current competition in AI, firms are trying to build moats. Concerns exist that winner-take-all dynamics, like in search, could emerge. However, the ease of replicating models may prevent dominance. Regulatory scrutiny is increasing, potentially hindering anti-competitive deals.

Big Tech firms are betting on AI as a structural shift, consolidating influence through strategic entanglements rather than mergers. These deals, designed to circumvent FTC rules, may be anticompetitive by blocking access to key services. The Economist argues against breaking up Meta, citing the rise of TikTok as proof of competition.