This i[nterview on Bloomberg Podcasts](https://www.youtube.com/watch?v=zbKDmkJPVvI) highlights the Circularity of the Hypersclaler's AI revenues on Anthropic and Open AI.
* The AI revenue streams of cloud giants (Google, Amazon, MSFT) are heavily dependent on two unprofitable, capital-hungry companies—OpenAI and Anthropic—creating systemic concentration risk.
|Metric|Google Cloud|Microsoft Azure|Notes|
|:-|:-|:-|:-|
|2024 AI Revenue from OpenAI & Anthropic|27%|13%|UBS (Google), Barclays (Microsoft)|
|2025 AI Revenue Projection|48%|18%|Significant increase in revenue reliance|
|OpenAI 2025 Losses|*Not specified*|$20.9 billion|Massive losses reported by OpenAI|
|Infrastructure Cost (estimate)|\~$100 billion|*Not specified*|Mainly covered by cloud providers|
|Data Center Capacity Planned (GW)|190 GW|*Not specified*|From Sightline Climate report|
|Estimated Infrastructure Revenue Need|$1.6 trillion annual|*Not specified*|To support planned data centers|
* OpenAI is deeply intertwined with Microsoft, which holds a near-controlling economic interest and provides its core compute infrastructure.
* Anthropic has taken on major investments from Amazon, Google, Microsoft, and NVIDIA
* These AI companies burn immense amounts of cash and rely on investor capital inflows, making their sustainability uncertain.
* these AI companies "do not pay their bills out of existing cash flow," those receivables represent a concentrated credit risk for the hyperscalers, contingent entirely on the AI companies' ability to continuously raise external capital from investors (including the hyperscalers themselves).
* The required compute infrastructure is vast and costly, with a growing gap between capacity and paying customers.
* Delays in AI companies going public and profitability questions raise red flags for investors expecting robust returns from AI-driven growth.
* The current market resembles a form of circular financing where cloud providers own the infrastructure and also fund the primary customers, which could magnify risks if growth slows.
* Without substantial productivity gains from AI, the justification for such heavy investments and high valuations remains precarious.
* Cloud providers must manage the challenge of balancing infrastructure expansion with a narrow customer base and potentially diminishing returns.
This evolving situation requires close monitoring from investors, regulators, and market participants, as the AI ecosystem faces critical tests of scalability, profitability, and technological impact in coming years.
