When Google’s parent company Alphabet reported its latest quarterly results on July 22, investors balked at its negative free cash flow and increased capital expenditures on data centers.

But Google is not like OpenAI or Anthropic, both of which are still trying to find a profitable business model as they burn through cash; it remains a cloud-computing giant, and the results of its cloud division for the second quarter were nothing short of extraordinary.

Google Cloud’s revenues grew by 82 percent year on year and are now equivalent to 40 percent of the revenues generated by Google Search. Its operating margin was equally impressive, rising by 15 points to hit 36 percent. The era in which Google generated profits exclusively from advertising is long gone.

Artificial intelligence is the Trojan Horse for these cloud giants; more AI hype means more cloud usage, which expands their businesses.

Cecilia Rikap

Two other tech giants, Microsoft and Amazon, are in a similar boat. Microsoft’s cloud business is its fastest-growing division, generating record profits in 2026. Net sales by Amazon Web Services have jumped by 37 percent year on year, in what the company described as the fastest increase in 18 quarters. Operating income at AWS grew even more, by over 60 percent compared with the same quarter in 2025. Just like Google, Amazon’s free cash flow turned negative during this period. Given its other results, however, this is not a signal of weakness but of a quest for future growth.

Artificial intelligence is the Trojan Horse for these cloud giants; more AI hype means more cloud usage, which expands their businesses. Because they own the cloud, they control the whole digital stack.

Everything from data and software to semiconductors and frontier AI models is offered through the cloud, which is the indispensable supermarket for those developing or deploying (consuming) digital technologies. Because third parties offer digital technologies as services, too, they expand the appeal of the cloud as a one-stop shop. The result is a vast network of thousands of companies — including other giants such as SAP, Oracle and IBM — that offer their technologies through platforms owned by Amazon, Microsoft or Google.

Thus, the cloud is where digital technologies are co-produced by the many for the disproportionate benefit of the few. Controlling the cloud has distinguished Amazon, Microsoft and Google even among other massive tech companies such as Oracle and Nvidia, whose own advantages depend narrowly on AI — although Nvidia, a chokepoint for chip design, has more leeway within the AI value chain, whereas Oracle is trying to reinvent itself as an AI infrastructure provider after deciding in 2023 that it would operate its cloud infrastructure services on Microsoft’s data centers. Similarly, Meta’s recent infrastructure investments mostly reflect its AI ambitions.

Unlike Oracle or Meta, the cloud giants are investing in data centers not only to train and run AI models, but also to bring more clients into their existing ecosystems. Companies from every industry, along with governments and even international organizations, rent vast, bespoke suites of digital services from Amazon, Microsoft and Google. These services include storage on virtual servers, content-delivery networks, cybersecurity software, technology for routing user traffic, and hundreds of others.

This much more diverse business explains why the cloud giants are less exposed to AI-industry risks than Oracle, Nvidia or Meta, not to mention companies such as OpenAI and Anthropic. For the cloud hegemons, AI is just one more feature to integrate into an existing suite of thousands of other cloud services.

If the promise of AI fails, it is companies such as Oracle and Nvidia that will feel the squeeze.

Cecilia Rikap

The greater the competition between AI models, the stronger the cloud giants will become. More AI models means more development and consumption of digital technology, and thus more cloud businesses. Because AI labs such as OpenAI and Anthropic suffer every time a Chinese or US startup releases a competitive open-weight model, they cannot pursue a strategy independent of the cloud giants. They must focus all their time and attention on releasing even more advanced models, which pushes them to consume more cloud services.

If the promise of AI fails, it is companies such as Oracle and Nvidia that will feel the squeeze. Given the growing discontent over data-center construction and the increasingly suspect narrative about AI’s return on investment, these companies have good reason to worry. But Amazon, Microsoft and Google would emerge as strong as ever, perhaps even stronger, owing to their greater financial leeway and more diversified business.

The companies and public bodies migrating to the cloud out of a sense of AI “FOMO” (fear of missing out) will remain cloud-bound even if the AI investment boom turns out to be a bubble. Digital architecture is increasingly the backbone of most organizations and activities, and building similar systems outside of the cloud hegemons’ ecosystems would be incredibly costly and challenging, if not impossible.

Regardless of what happens with AI, the cloud giants’ control over our economies and societies will only grow, becoming unaccountable accretions of private power that no democracy can accept. With all eyes on AI, we risk failure to see the forest for the trees.

  • Cecilia Rikap, associate professor of economics and head of research at University College London’s Institute for Innovation and Public Purpose, is the author, most recently, of “The Rulers: Corporate Power in the Age of AI and the Cloud” (Verso Books, 2026). ©Project Syndicate