Google, Microsoft, and Amazon pursuing investments and deals with generative AI startups, incentivizing use of their cloud services, raises regulatory concerns
Deals between Google, Microsoft and Amazon and ‘generative AI’ start-ups raise competition concerns
Context & Ripple Effects
Days after sources reported [[a:835040|Google taking a $300M, 10% stake in Anthropic with a requirement to spend the money on Google Cloud]], the Financial Times frames the pattern across all three hyperscalers: investment in generative AI startups doubles as a channel for locking those startups' workloads into the investor's cloud.
The concern is that equity plus committed cloud spend blurs the line between venture backing and customer capture. The coverage since has validated the worry: by late 2023 the three companies supplied 66% of the $27B raised by generative AI startups that year, and by early 2024 both the FTC had opened a probe into the Microsoft–OpenAI and Google/Amazon–Anthropic deals and the UK CMA had flagged an "interconnected web" of 90+ such arrangements.
First-order effects
- Generative AI startups like Anthropic get capital they could not raise from VCs at this scale, but accept terms that route their training and inference spend to the investor's own cloud platform.
Second-order effects
- Rival cloud providers without anchor model stakes must compete differently — AWS's response was to position itself as a neutral platform hosting LLMs from Anthropic, Stability AI, and AI21 Labs rather than betting on one captive partner.
Third-order effects
- If the pattern holds, antitrust enforcement shifts from app-store-style gatekeeping to scrutinizing how equity investments steer procurement — the FTC probe and the CMA's 90-deal map suggest regulators now treat startup financing itself as a competition issue.
The trend: Cloud infrastructure is being financialized into the dominant funding channel for frontier AI, converting venture capital into a lock-in mechanism regulators are only beginning to map.