Author: US Stock Research Institute" (ID: meigushe)

On July 26, 2023, Beijing time, Microsoft disclosed its financial report for the fourth quarter and full year of fiscal year 2023. Driven by businesses such as the productivity and business process department and the intelligent cloud department, Microsoft’s fourth-quarter revenue was US$56.189 billion, an increase of 8% year-on-year; net profit was US$20.081 billion, a year-on-year increase of 20%; diluted earnings per share were US$2.69, a year-on-year increase twenty one%.

Such an eye-catching performance has not attracted investors' pursuit. On July 25, Eastern Time, Microsoft's stock price rose 1.7% in regular trading, but with the after-hours disclosure of the financial report, Microsoft's stock price fell by 3.7%.

On the face of it, Microsoft's stock price drop appears to be due to its earnings outlook missing expectations. It is understood that Microsoft expects revenue in the first quarter of fiscal year 2024 to be 53.8-54.8 billion US dollars, with a median of 54.3 billion US dollars, an increase of only 8% year-on-year, which is lower than the average analyst forecast of 54.94 billion US dollars.

But in fact, the capital market pays more attention to the growth of enterprises. As the overlord of AI, Microsoft's AI story has not been effectively fulfilled, which may be the main reason why some investors choose to "escape".

01 PC business continues to slump, AI becomes Microsoft's top priority

Since it bet on the dark horse of OpenAI as early as 2019, with the explosion of ChatGPT, Microsoft finally got rid of the nightmare of declining PC shipments, and described the "new story" of AI.

In the fourth fiscal quarter, Microsoft's more personal computing business (Windows, devices, games and search advertising) revenue was US$13.91 billion, a year-on-year decrease of 4%. Among them, Windows operating system revenue fell 12% year-on-year; hardware device revenue fell 20% year-on-year; Windows commercial product and cloud service revenue rose 2% year-on-year; search and news advertising service revenue rose 8% year-on-year.

Source: Microsoft

It can be seen that the sluggish performance of the PC market continues.

Data disclosed by Canalys shows that in Q2 of 2023, global PC shipments will be 62.1 million units, a year-on-year decrease of 11.5%. Behind this is mainly factors such as the sharp drop in consumer demand, which has caused the PC industry chain to be hit hard, and Microsoft, which is sitting on Windows systems and Surface terminals, is no exception.

Source: Canalys

However, it is worth noting that Microsoft's search and news advertising service revenue is still growing, up 8% year-on-year. This is largely thanks to Microsoft's integration of ChatGPT in Bing. Similarweb revealed that after adding AI features, Bing saw a 15.8% increase in page views, while Google saw a nearly 1% drop.

In fact, in addition to the search engine, Microsoft's productivity and business process department (including Office productivity software, LinkedIn and Dynamics) and the two major business lines of the intelligent cloud department have also launched a series of layouts around the concept of AI.

For example, on July 18, 2023, at the Microsoft Inspire event, Microsoft stated that Microsoft 365 Copilot will provide AI services to commercial subscribers at an additional $30 per user per month. After the press conference, Microsoft's stock price rose by 3.98%, and the stock price hit a high of $366.78 per share.

Source: Microsoft

Earlier in March 2023, Microsoft also launched the Azure OpenAI service in the global version of Azure services to provide enterprise-level services guaranteed by SLA and help enterprise customers fully explore the huge potential of combining AI with business scenarios.

Based on these actions, the two major business lines of Microsoft's productivity and business process department and the intelligent cloud department have significant growth potential, making a greater contribution to the growth of overall performance.

In the fourth fiscal quarter, Microsoft's productivity and business process revenue was US$18.29 billion, a year-on-year increase of 10%, and operating profit was US$9.052 billion, a year-on-year increase of 25.1%. Office commercial products and cloud service revenue increased by 12% year-on-year, and the number of subscribers of the personal version of 365 increased to 67 million; LinkedIn revenue increased by 5% year-on-year.

During the same period, the revenue of Microsoft's intelligent cloud department was 23.993 billion US dollars, a year-on-year increase of 15%; operating profit was 10.526 billion US dollars, a year-on-year increase of 19.5%. Among them, revenue from server products and cloud services increased by 17% year-on-year.

Compared with companies such as Google, Microsoft has a comparative advantage in the AI business. On the one hand, it is a shareholder of OpenAI, which can directly apply mature ChatGPT products at the product level; on the other hand, it is the exclusive cloud service provider of ChatGPT, ChatGPT The powerful strength also provides an endorsement for Microsoft's intelligent cloud Azure to capture customers.

In this regard, Nancy Tengler, chief investment officer of Laffer Tengler Investments, an investment consulting firm, once said, "Almost everyone believes that generative artificial intelligence will be the key to the next few decades, and I think Microsoft is far ahead of others. Front". But being a pioneer also means being the first to meet the test.

02 The commercialization process of AI is not ideal, how to balance the huge cost is a difficult problem

Although AI technology is very imaginative, there are also obvious problems, that is, it requires huge investment of funds and computing power to have a prominent competitive advantage.

Taking ChatGPT as an example, the reason why it can become the benchmark product of AI large-scale models is mainly due to the "burning money" regardless of cost.

Tianyancha shows that since 2015, OpenAI has completed multiple rounds of financing, with a cumulative financing amount of more than 1.5 billion US dollars, which is mainly used to develop artificial intelligence capabilities and train machine learning models. However, due to the high cost of training, it is difficult to contain losses. In 2022, OpenAI's losses will double to 545 million US dollars.

In the fourth fiscal quarter, Microsoft’s capital expenditures also reached US$10.7 billion, a year-on-year increase of 37.18%, a new high in a single quarter since 2016. In the analyst conference call, Microsoft CFO Amy Hood also stated that as Microsoft builds new data centers to support AI and pushes up costs, the company's capital expenditures will increase quarter by quarter in fiscal year 2024.

In fact, it is almost a common phenomenon in the technology industry to overweight AI and lead to rising costs. On the analyst conference call, Google President, Chief Investment Officer and CFO Ruth Porat also stated, "We expect the company's investment in technology infrastructure to increase in the second half of 2023, consistent with previous guidance." , that is, the company's capital expenditure in 2023 will be higher than that in 2022".

In order to balance the high capital expenditure, Microsoft has actively explored the commercialization of AI, but at present, these commercialization actions of Microsoft have not yet shown due imagination.

For example, for its own Windows, Microsoft and other trump products, Microsoft launched the Copilot function. According to Microsoft's presentation, Copilot can be called a powerful AI assistant. Whether it is controlling computer settings, optimizing articles, or designing PPTs, it can be done with one click.

However, according to user experience, the performance of Windows Copilot is not satisfactory. "The only functions currently are New Bing and some basic system function calls."

Therefore, looking at Microsoft 365 Copilot's additional monthly pricing of $30 per user, it seems that there is not enough confidence to drive consumer consumption. Technology investor Paul Meeks believes that "Copilot pricing is too aggressive and may exceed the budget of many ordinary users."

Coincidentally, although Bing has captured many users with the gimmick of ChatGPT before, as the novelty gradually recedes, Bing's influence has been greatly reduced. According to data disclosed by SimilarWeb, in June 2023, in the US iOS market, the download volume of Bing dropped by 38%.

Reflected in the financial report, the growth rate of Microsoft's search and news advertising revenue has dropped from 10% in the third fiscal quarter to 8% in the fourth quarter. Correspondingly, Google’s advertising business has swept away the decline in the previous two quarters. In the second quarter, advertising business revenue was 58.143 billion US dollars, a year-on-year increase of 3.3%, which was 2.1% higher than market expectations.

The above problems essentially reflect that the current AI product experience is not perfect, and it is difficult to make a differentiated advantage to continue to attract C-end users. From the B-end point of view, under the background of the existing Internet model innovation plan, AI It provides more room for imagination for most enterprises, and not all enterprises have high-performance computing resources, and their investment in AI is likely to last for a long time.

Therefore, the most imaginative commercialization scenario of Microsoft's AI business may still be a "shovel seller" who provides AI technology to developers or enterprise customers.

Previously, at the Build conference, Microsoft launched several cloud service products such as Azure OpenAI Service and Azure AI Studio to assist in the deployment or development of AI functions.

This is already the consensus of most players in the industry. In April 2023, Ali released its own large-scale model product "Tongyi Qianwen". Subsequently, Alibaba Cloud launched the "Tongyi Qianwen Partner Program" to provide Qianwen partners with technology, services and product support in the field of large models. In March 2023, in an interview with 36 Krypton, Robin Li, the founder of Baidu, also said to the outside world, "The emergence of a large language model is a game changer for cloud computing, and it will change the rules of the game for cloud computing."

However, facing the B-end market, the development of Microsoft's AI business has also been challenged, which can be seen from the slowdown in Microsoft's cloud revenue growth.

In the fourth fiscal quarter, Microsoft’s intelligent cloud business revenue was US$23.99 billion, a year-on-year increase of 15%, of which Azure and other cloud service businesses increased by 26%, compared with 31% in the previous quarter. Looking ahead to the first quarter of fiscal 2024, Microsoft expects Azure revenue growth to slow further at 25% to 26%.

Source: Microsoft

The big background is that many companies are stepping up cost reduction and efficiency enhancement, and optimizing cloud service usage, thus reducing demand. In addition, for most enterprise customers, although the capabilities of large models have indeed made great progress compared with traditional artificial assistants, it is still unclear how large models can bring higher benefits to their own platforms when empowering consumers .

Taking search as an example, Alphabet chairman Hennessy once pointed out that the search cost based on a large language model may be 10 times that of a standard keyword search. If higher efficiency and accuracy cannot be provided, then search engine companies and users actually prefer keyword searches. This is the key reason for the ebb and flow of Bing and Google.

Overall, although from the perspective of revenue generation, Microsoft's performance in the fourth quarter of fiscal year 2023 is quite impressive, it is worth noting that Microsoft is currently in the "AI shift" period, from the perspective of "new stories" Looking at it, the growth rate of Microsoft's specific business is actually not ideal.

This is largely due to the gap between Microsoft's AI concept and reality. Whether it is Copilot for the C-side or Azure OpenAI Service for the B-side, it is difficult to bring users and customers an experience upgrade.

However, it cannot be denied that large models will most likely reshape the Internet and the technology industry in the future. As a pioneer in the AI track, what Microsoft needs to do may be to continue to improve AI technology and wait for the day when AI technology has revolutionary capabilities.