How Cloud Giants Cement Their Power w/ Cecilia Rikap

Episode Summary

Executive Summary: Cecilia Rikap argues that Amazon, Microsoft, and Google gain power not just from AI models but from controlling cloud infrastructure, venture capital, and research networks that make governments, firms, and universities dependent on them. She says generative AI accelerates this dominance, and the alternative is public, democratic, and internationally coordinated digital infrastructure.

Main Topics: Cloud infrastructure as corporate power (Priority: 5/5): Cloud computing is presented as rented, centralized infrastructure that makes organizations dependent on Amazon, Microsoft, and Google for storage, compute, software, and data services. Dependency of startups, firms, and states (Priority: 5/5): Small startups and even large corporations choose cloud services for flexibility and cost, but over time become locked in and pay more as usage and AI reliance grow. Generative AI as an accelerant (Priority: 5/5): The rise of ChatGPT and similar tools massively increased AI adoption, pulling more companies onto cloud platforms and deepening reliance on big tech ecosystems. Corporate venture capital and ecosystem control (Priority: 4/5): Big tech invests in AI startups to steer development, gain early access, and keep startups in a dependent peripheral role rather than acquiring them outright. Academic capture and double affiliation (Priority: 4/5): Big tech shapes AI research through scholars with dual appointments, funding, and collaborations, influencing university agendas while keeping key pieces of work secret. Open source as strategy, not disruption (Priority: 4/5): Open source AI is framed as a way for big tech to widen adoption and standardize its models, while keeping critical data, compute, or components under control. Public alternatives and democratic planning (Priority: 5/5): Rikap calls for public cloud infrastructure, data solidarity, international solutions, and democratic planning oriented toward social and ecological needs rather than profit.

Key Arguments: Cloud services are not neutral efficiency tools; they create structural dependency because AI and digital work require compute, software, and data hosted on platforms owned by a few firms. The cloud is attractive because it reduces upfront costs, but long-term usage typically becomes more expensive and harder to exit, especially once systems are built around proprietary services. Large companies use cloud infrastructure to escape uncertainty and update technologies faster, even when it increases reliance on big tech and raises costs over time. Generative AI intensified cloud lock-in by making AI easier to adopt and by encouraging firms to use AI across marketing, product development, drug discovery, and other functions. Big tech’s venture capital investments are strategic: they help steer startup development, secure early access to technology, and preserve a core-periphery structure in the AI market. Academic collaborations with big tech are a form of control beyond ownership: companies can influence research agendas, even when they do not fully own the institutions or outputs. Open source can be used by dominant firms to set industry standards and recruit free labor, while the most valuable inputs—data, compute, and some code—remain proprietary. A better model would treat cloud and data infrastructure as public utilities and focus AI development on concrete social problems such as healthcare and ecological transition. Data should be governed through solidarity, not pure privatization or unrestricted extraction; some data uses should be forbidden, while socially useful data should be shared. Regulation must address control without ownership, because existing antitrust frameworks often miss how big tech governs ecosystems through dependence rather than direct acquisition.

Data Points: Global cloud market share of Amazon, Microsoft, and Google: 66% - Rikap says the three firms together dominate the global cloud business. OpenAI investment by Microsoft: $1 billion - Microsoft’s 2019 corporate venture investment in OpenAI. Top AI startup funding share in 2023: Two-thirds - Two-thirds of all money invested in generative AI startups came from Microsoft, Amazon, and Google in 2023. Google top-investor count: 2,445 companies - From ChatGPT’s release until February 2023, Google was a top investor in 2,445 companies on Crunchbase data. Double-affiliation institutions identified: At least 100 institutions - Rikap found universities and public research organizations with at least one scholar holding a double affiliation with big tech. Advance access to OpenAI tech: 6 months - A Microsoft source said the company got access to OpenAI’s technology about six months early. Energy use of generative AI queries: Around 15 times more energy - Rikap says models like Gemini or ChatGPT can consume around 15x more energy per query than a typical Google search. Years cloud has existed: Over a decade / more than 15 years - Rikap notes the cloud has enabled the startup boom for more than 15 years.

Pivotal Quotes: "we need to have this infrastructure operating as the water pipes, this needs to be provided as a public service, needs to be a public utility." — Cecilia Rikap: Her proposal for publicly owned cloud infrastructure and digital sovereignty. "they are controlling without owning controlling others without owning them" — Cecilia Rikap: Her core thesis about how big tech exerts power across AI, cloud, academia, and startups. "if we really want to have this infrastructure operating as the water pipes, this needs to be provided as a public service" — Cecilia Rikap: A concrete analogy used to argue that cloud infrastructure should be treated like essential public infrastructure.

Implications: The episode suggests AI policy should target infrastructure, finance, and research capture—not just models. Without public cloud alternatives and stronger regulation, big tech will keep expanding control over the digital economy.