Microsoft and Amazon's earnings reports: Investors are more concerned about order backlogs being depleted by AI than revenue growth rates

As Microsoft and Amazon released their earnings reports after the U.S. stock market closed on Wednesday, Wall Street began to focus on a metric that typically does not appear in press releases—"remaining performance obligations," which refers to the sales backlog that cloud computing companies will recognize from multi-year contracts in the future. This figure is becoming a new benchmark for assessing whether the AI boom is overheating.

Since the release of ChatGPT, AI companies like OpenAI and Anthropic have made long-term commitments worth hundreds of billions of dollars to cloud service providers, and these massive contracts are reshaping the competitive landscape of the cloud computing market. Data shows that Microsoft's new order backlog has already surpassed that of Amazon's cloud computing business AWS, while Google is also nearly catching up to this market share leader.

However, investor concerns have also arisen: Can the sales from these commitments truly be realized? The case of Oracle serves as a warning—despite its backlog soaring to $460 billion due to OpenAI's $300 billion commitment, its stock price fell by a quarter, as the market questioned the profitability of this business and whether it could all materialize.

Bank of America analyst Justin Post stated, "Order backlog is a good leading indicator," especially when it diverges from revenue growth direction, investors tend to pay more attention." The market's reaction to these commitments when cloud computing companies report earnings will be an important signal for assessing the risk of an AI bubble.

Order Backlog Indicator Emerges

During cloud computing earnings season, investors typically focus on the revenue growth rates of AWS, Microsoft Azure, and Google Cloud Platform. However, the financial metric of "remaining performance obligations" is gaining more attention.

This figure shows the sales that cloud service providers will recognize from multi-year contracts in the future. As AI companies like OpenAI and Anthropic make substantial long-term commitments to cloud providers—these servers provide the computational power for their AI models—the scale and changes in order backlog are revealing the competitive dynamics of the AI era.

Bank of America analyst Justin Post pointed out that when order backlog diverges from revenue growth direction, investors pay special attention to this metric. Data as of September 30 of last year shows that this usually inconspicuous metric has become key to understanding the shift in the cloud computing market landscape.

Competitive Landscape Being Redefined

Although Amazon's cloud computing division pioneered the on-demand sales model for computing infrastructure and still holds the largest market share, since the release of ChatGPT at the end of 2022, in terms of the amount of new order backlog, Amazon has fallen behind the second-ranked cloud service provider Microsoft, and is even slightly behind Google.

Google previously had a lackluster performance in the cloud computing market, but with its cutting-edge AI technology, it has risen to become a strong competitor capable of challenging larger rivals. Justin Post stated, "Google is a huge success story," as the company has successfully attracted AI companies to pay for renting its custom AI chips.

This dynamic, along with Amazon's slower growth rate, explains investors' relatively pessimistic attitude towards AWS last year. Microsoft, as the primary cloud service provider for OpenAI, holds an advantage in order backlog growth, while Google's rise has rewritten the three-way competitive landscape of the cloud computing market

Concerns Behind the Numbers

The backlog data itself has limitations. Each company's statistical criteria vary slightly: Microsoft's figures include commitments from enterprises to its vast Windows and Office businesses, while Google's metrics also encompass its productivity suite and other businesses.

The greater risk lies in whether this cash can be realized as promised. During the pandemic, companies renegotiated contracts with cloud service providers, deferring expenditures to future years. Now, massive deals from a single cloud customer—often OpenAI—could unilaterally drive significant fluctuations in this figure.

Francine McKenna, a professor at Montclair State University and a correspondent on accounting issues, stated:

"The question is, how much of these remaining performance obligations they report are business far in the future, which may or may not be delivered, and they may or may not recognize revenue."

In the fourth quarter of last year, more large commitments were pouring in. OpenAI announced it would spend an additional $250 billion on Microsoft Azure and $38 billion on AWS. Anthropic committed to purchasing $30 billion in computing capacity from Azure as part of an investment agreement.

Oracle's Warning

Oracle's case has sounded the alarm for the market. Since the end of 2022, the company's backlog has surged by $460 billion, almost entirely reliant on OpenAI's commitment to lease $300 billion worth of space in Oracle's data centers in the future.

However, the company's stock price has dropped by a quarter, as investors question whether this business can be profitable and whether the complete deal can truly materialize. This market reaction shows that even if the backlog figures are astonishing, a lack of visibility on profitability or certainty of execution will still lead investors to vote with their feet.

As the latest commitments from Microsoft, Amazon, and Google are reflected in their respective backlogs, the market's response will become an important indicator of investors' concerns about a potential AI bubble. These figures not only reflect the scale of AI demand but also test the market's confidence in the sustainability of this technological revolution