In an AI gold rush, those selling the proverbial pickaxes are surest to win: cloud companies provide scalable managed computational resources as a subscription service now used by most businesses to store their data, and as a primary ingredient to build and use AI. Just three companies—Amazon, Microsoft, and Google—control two thirds of global cloud compute market share, collectively comprising “Big Cloud.” This highly concentrated market raises concerns regarding digital sovereignty, slowed innovation, and a concentration of corporate power.
In this report, we explore an underrecognized manner in which AI ecosystems increasingly depend on Big Cloud: Big Cloud’s investment in other companies. We show how Big Cloud companies are prolific investors widely deploying hundreds of billions of dollars over thousands of deals, often in smaller, lesser-known startups. We find that:
- While some regulators have begun to scrutinize the largest of these deals—such as Microsoft’s investment in OpenAI or Google and Amazon in Anthropic—the ecosystem-wide scale of this investment is hard to overstate: Big Cloud invests as frequently and at similar amounts to the largest venture capital firms and startup accelerators. Further, Big Cloud invests about ten times as often as other Big Tech companies, and ten to a hundred times more in total dollar amounts.
- Via accelerator programs, Big Cloud companies lock startups into their cloud infrastructure. Big Cloud ensnares young startups in their cloud ecosystem via cloud credits while requiring startups use the Cloud company’s other tech, and incentivizing strategies with particularly heavy cloud needs, such as generative AI.
- More so than when other Big Tech companies or VC firms invest, startups funded by Big Cloud are more likely to rely on Big Cloud as their lead or sole investor. These relationships allow Big Cloud to exercise significant influence over startups and bend them to their interests.
- Amid concerns that vertical integration may give one firm too much control over AI supply chains—such as chips, cloud, or data—our work shows that Big Cloud is investing in a way that brings many of the same risks as conventional forms of vertical integration: when Big Cloud invests in an AI supply chain company—such as a Data, X-as-a-Service, or Internet infrastructure company—that company is often more likely to be dependent on that Big Cloud company as their sole or lead investor, compared with other investors.
- Intensifying concerns about threats to global digital sovereignty, we find that American Big Cloud companies make global investments at a far greater pace than other investors we compare against. Just over half of all Big Cloud investments are made internationally, about twice the frequency of large VCs, top accelerators and other Big Tech companies. Big Cloud also invests through accelerators abroad much more often than at home, highlighting the need for global regulatory scrutiny of startup accelerator programs.
While these practices merit creative regulatory and policy responses, we emphasize that such interventions should proceed in light of the following overarching implications: — Dependence on Big Cloud is not just technical or contractual. It is also financial, as a source of investment. This compounds the need for structural separation: Amazon, Google, and Microsoft must be compelled to split their cloud business from their other businesses that run on the cloud, per past calls, so that they do not both provide infrastructure and compete with the customers and investees relying on that infrastructure.
— Big Cloud companies are huge investors, which sets them apart from all other large tech companies. Any one of these investments may be small and insignificant, but they cumulatively shape the startup and developer ecosystem in Big Cloud companies’ interest. Thus, in addition to “deal by deal” scrutiny, in which only the largest deals receive attention, regulators and researchers should monitor and scrutinize these investments and their effects in an ecosystem-wide, cumulative, and ongoing manner.