--- title: "Major Shift in the Role of Cloud Giants: From Computing Power Providers to AI Distribution Hubs" type: "News" locale: "en" url: "https://longbridge.com/en/news/288689966.md" description: "A UBS report points out that the role of hyperscale cloud providers has evolved from being mere GPU computing power suppliers to becoming core distribution hubs for enterprises to access cutting-edge AI models. This transformation has spawned three layers of high-margin revenue: first, model API services that allow enterprises to directly call AI models encapsulated in the cloud; second, distribution channels for third-party AI products, where platform service fees are charged by integrating tools such as those from OpenAI; and third, sales of proprietary AI products, such as enterprise subscriptions for Google Gemini and Microsoft GitHub Copilot" datetime: "2026-06-04T08:42:17.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/288689966.md) - [en](https://longbridge.com/en/news/288689966.md) - [zh-HK](https://longbridge.com/zh-HK/news/288689966.md) --- # Major Shift in the Role of Cloud Giants: From Computing Power Providers to AI Distribution Hubs The demand for AI computing power is fundamentally reshaping the growth logic of hyperscale cloud computing providers. According to Zhuifeng Trading Desk, a UBS report indicates that while the market still views the three major cloud giants as mere AI computing power providers, **their roles have quietly transformed—from selling GPU computing power to becoming core distribution platforms for frontier models like those from OpenAI and Anthropic to reach enterprise customers.** The incremental revenue brought about by this structural change has not yet been fully priced in by the market. In the first quarter of 2026, the combined revenue of the three cloud giants reached $84.8 billion, a year-on-year increase of 39%, with backlog orders hitting $2.1 trillion, a surge of 184% year-on-year. This growth rate and order magnitude exceed what can be explained by simple "computing power resale," suggesting that enterprise customers are paying higher premium "platform access fees" to connect to frontier AI models. The market had previously worried that soaring capital expenditures (expected to reach $673 billion in 2026) to support computing power expansion would erode profits. However, over the past six quarters, their combined operating profit margin has remained stable at 36%-37%. This indicates that **high-margin platform service revenue is becoming the main support for profits, rather than low-margin computing power resale.** Nevertheless, the market continues to price these companies under the framework of "computing power providers," with current forward P/E ratios for 2027 ranging only from 19 to 27 times. For a sector with an overall revenue growth rate of approximately 40% and an income structure shifting towards high-margin platform services, this valuation clearly underestimates the potential for value re-rating driven by the role transformation. ## **Three-Layer Revenue Model: How Cloud Giants Are Shifting from "Selling Computing Power" to "Selling Platforms"** Wall Street is accustomed to viewing hyperscale cloud service providers as "resale channels" for NVIDIA GPUs, meaning they purchase chips and output computing power to customers to earn the spread. However, UBS points out that this perspective overlooks the huge incremental revenue they generate as core distribution platforms for OpenAI and Anthropic. Google Cloud's performance is a typical example. It is clearly insufficient to explain its 35 percentage point leap in growth rate over the past year solely through "computing capacity expansion." **Cloud giants are replicating the successful path of 10 to 15 years ago—extending from underlying infrastructure to upper layers, evolving into "model distribution hubs" in the AI era.** The revenue sources from this distribution platform role are mainly reflected in three layers. **First layer: Model API services.** Enterprises do not need to build their own AI models but can directly call them via API interfaces provided by cloud vendors. Microsoft's Azure OpenAI API service has already formed a business volume with annual revenue in the billions of dollars. Amazon AWS Bedrock acts similarly to a "model marketplace," where customers can select and call various models; currently, its customer base has reached 125,000, covering 80% of Fortune 100 companies, with customer spending increasing by 170% quarter-on-quarter in the first quarter. **Second layer: Distribution channels for third-party AI products.** OpenAI and Anthropic are vigorously promoting their respective proprietary AI programming tools, but large enterprises tend to purchase them through existing procurement agreements with Microsoft Azure or AWS to obtain better security compliance guarantees and scale bargaining power. Cloud vendors charge platform service fees in this process, or drive additional consumption of underlying computing power. **Third layer: Sales of proprietary AI products.** Google's Gemini API and enterprise edition subscriptions are estimated to generate combined annual revenue of approximately $2 billion, covering more than 11 million seats. Microsoft GitHub Copilot's enterprise subscriptions have doubled year-on-year. The aforementioned revenues are directly accounted for in the cloud business segment. In addition, Google has adopted a differentiated strategy—directly selling TPU chips (an alternative to Nvidia GPUs) to specific external customers. Related revenue is recognized in full upon shipment, with the majority expected to be recorded in 2027. This is likely one of the important driving factors behind the sudden quarter-on-quarter increase of $220 billion in Google Cloud's orders in the first quarter. ## Valuation Expansion Is Far From Peaking UBS believes that this round of valuation repair for hyperscale cloud service providers is not yet over. There are currently **two obvious cognitive biases** in the market: first, **underestimating the growth slope of AI computing power demand**; and second, more critically, **severely underestimating the upside revenue potential brought by AWS and Azure serving as distribution platforms for the full suite of AI products from OpenAI and Anthropic.** From a valuation perspective, based on expected earnings for 2027, Amazon, Google, Microsoft, and Oracle correspond to P/E ratios of 24x, 23x, 19x, and 27x, respectively. For a sector with an overall revenue growth rate of approximately 40%, strong growth momentum, and resilient profit margins, a valuation range of 19 to 27 times is at a relatively reasonable level. ### Related Stocks - [MSFT.US](https://longbridge.com/en/quote/MSFT.US.md) - [GOOG.US](https://longbridge.com/en/quote/GOOG.US.md) - [GOOGL.US](https://longbridge.com/en/quote/GOOGL.US.md) - [MSFL.US](https://longbridge.com/en/quote/MSFL.US.md) - [MSFU.US](https://longbridge.com/en/quote/MSFU.US.md) - [GGLS.US](https://longbridge.com/en/quote/GGLS.US.md) - [GGLL.US](https://longbridge.com/en/quote/GGLL.US.md) - [UBS.US](https://longbridge.com/en/quote/UBS.US.md) - [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md) - [AMZN.US](https://longbridge.com/en/quote/AMZN.US.md) - [ORCL.US](https://longbridge.com/en/quote/ORCL.US.md) - [NVD.DE](https://longbridge.com/en/quote/NVD.DE.md) - [ORCL-D.US](https://longbridge.com/en/quote/ORCL-D.US.md) ## Related News & Research - [Meta Delays New AI Model Again — What It Could Mean for Investors](https://longbridge.com/en/news/288678053.md) - ['Google Is the Clear Beneficiary,' Says Top Analyst as AI Search Takes Off](https://longbridge.com/en/news/288300466.md) - [IBM shares hit record highs after $350 analyst target and Google Cloud AI pact](https://longbridge.com/en/news/288751014.md) - [If You Want to Invest in AI Without Losing Sleep, Buy Google Stock Now](https://longbridge.com/en/news/288316541.md) - [UK regulator enforces new competition requirements on Google search](https://longbridge.com/en/news/288522733.md)