According to a recent FTC report, Big Tech's growing investments in AI startups, such as Microsoft's $13 billion funding of OpenAI and Amazon's and Google's partnerships with Anthropic, are raising alarms about potential monopolization and competitive risks in the AI and cloud computing sectors. One of the major FTC concerns is the so-called 'circular spending' practice, under which startups spend their received money on services their investors offer, notes Bloomberg.
Training artificial intelligence models requires a lot of computing power that usually costs billions of dollars, an amount of money that startups like Open AI and Anthropic do not have and can barely rise. As a result, they turn to Big Tech companies with billions of dollars and massive amounts of computing power. For obvious reasons, companies like Amazon, Google, and Microsoft try to capitalize on their deals with AI companies, which usually means exclusivity. Also, tech giants want their partners to spend the money they receive on products and services these companies offer.
The report highlights concerns that these investments could allow major cloud service providers (CSPs) like Amazon, Google, and Microsoft to dominate the AI market by locking startups into exclusive agreements. These partnerships might restrict startups from collaborating with multiple firms, increasing dependency on their benefactors and reducing market competition.
"The FTC's report sheds light on how partnerships by big tech firms can create lock-in, deprive start-ups of key AI inputs, and reveal sensitive information that can undermine fair competition," said FTC Chair Lina M. Khan.
A major issue is the requirement that startups spend a portion of their funding on tech giants' products and services, a practice the FTC calls 'circular spending.' For example, much of Microsoft’s investment in OpenAI involved credits for its Azure cloud platform, creating a feedback loop that shields Microsoft from financial risks. In return, CSPs provide AI developers with discounted access to vast computing resources that cannot usually be obtained elsewhere.
Additionally, CSPs gain access to intellectual property, advanced AI models, and critical data related to training and development. The FTC also noted that some deals grant large tech firms access to sensitive information. In at least one case, a company received weekly reports about an AI startup's financial performance, including revenue trends and customer data, a type of access that could provide unfair competitive advantages.
Moreover, the report reveals that some agreements include provisions allowing tech companies to use the output of AI models—referred to as 'synthetic data'—for training purposes. This practice raises ethical and competitive concerns, as it could enable the giants to strengthen their own AI systems at the expense of their partners.
Another concern is that Big Tech companies leverage these partnerships to integrate AI models into their products or deploy them on their platforms, expanding their influence in the AI and cloud markets.
Finally, there is a fear of consolidating highly specialized AI talent. The FTC warns that these partnerships could concentrate skilled engineers within a few dominant firms, making it difficult for smaller companies to acquire the expertise needed to develop advanced AI models. While this concern might have merits, the number of AI startups that emerge every month does not prove that.
The report, based on data from September 2024 and public information until January 2025, aims to inform policymakers, regulators, and the public about the implications of these partnerships.
