Spending big on AI but still not enough power? Microsoft's cloud revenue unexpectedly slowed last quarter, dropping over 6% in after-hours trading | Earnings Report Insights

The "artificial intelligence (AI) leader" Microsoft reported total revenue and profit for the last quarter that exceeded expectations, but capital expenditures accelerated growth. However, the revenue from Microsoft's cloud business, which directly benefits from AI technology applications, fell short of expectations, with growth not only failing to accelerate but also slightly slowing compared to the previous quarter. After the earnings report was released, Microsoft's stock price fluctuated downward, at one point dropping over 6%.

The rise of Chinese AI newcomer DeepSeek has recently disrupted the common perception in Silicon Valley that developing AI models requires massive investments in the most powerful and expensive chips. In the face of DeepSeek's impact, Microsoft now not only has to prove that its AI efforts can increase company revenue but also that the substantial investments in AI infrastructure are necessary. The latest earnings report failed to reassure increasingly skeptical investors regarding such investments by Microsoft.

When announcing the earnings report, after U.S. stock market hours on Wednesday, Microsoft stated that users could use DeepSeek's inference model R1 AI services on the Microsoft cloud platform Azure AI Foundry that day. The R1 version on Azure AI Foundry "has undergone rigorous red team and security assessments," including "automated evaluations of model behavior and extensive security reviews to mitigate potential risks."

On January 29, Wednesday, after U.S. stock market hours, Microsoft released financial data for the second fiscal quarter of fiscal year 2025 (referred to as the fourth quarter) ending December 31, 2024.

1) Key Financial Data

Revenue: Fourth quarter operating revenue was $69.63 billion, a year-on-year increase of 12%, compared to analyst expectations of $68.92 billion, with a year-on-year increase of 16% in the third quarter.

EPS: Fourth quarter diluted earnings per share (EPS) was $3.23, a year-on-year increase of 10%, compared to analyst expectations of $3.10, with a year-on-year increase of 10% in the third quarter.

Operating Profit: Fourth quarter operating profit was $31.7 billion, a year-on-year increase of 17%, compared to a year-on-year increase of 14% in the third quarter.

Net Profit: Fourth quarter net profit was $24.1 billion, a year-on-year increase of 10%, compared to a year-on-year increase of 11% in the third quarter.

Capital Expenditures: Including assets obtained through finance leases, total capital expenditures in the third quarter were $22.6 billion, a year-on-year increase of 96.5%, compared to a 78.6% increase in the third quarter; among these, cash expenditures related to real estate and equipment were $15.8 billion, a year-on-year increase of approximately 63%, compared to a 50.7% increase in the third quarter.

2) Segment Business Data

Commercial Cloud: Including products such as Office and Azure, the total revenue from Microsoft's commercial cloud business in the fourth quarter was $40.9 billion, a year-on-year increase of 21%, compared to analyst expectations of $41.1 billion, with a year-on-year increase of 22% in the third quarter.

Intelligent Cloud: Including Azure public cloud, Windows Server, voice recognition software Nuance, and GitHub's intelligent cloud business segment, fourth quarter revenue was $25.5 billion, a year-on-year increase of 19%, compared to a year-on-year increase of 20% in the third quarter.

Productivity and Business Processes: Including Microsoft 365 Copilot AI tools and other Office software, the productivity and business processes segment's revenue in the fourth quarter was $29.4 billion, a year-on-year increase of 14%, compared to a year-on-year increase of 11% in the third quarterMore Personal Computing: The personal computing business segment, which includes Windows operating system, Surface hardware, Xbox consoles, and video game company Activision Blizzard, reported revenue of $14.7 billion in the fourth quarter, showing almost no growth, compared to a 17% year-on-year growth in the third quarter.

After the earnings report was released, Microsoft's stock price, which had fallen over 1%, plummeted in after-hours trading, dropping more than 5% before narrowing its decline by more than half, briefly falling less than 1%, and then the decline expanded again, at one point dropping over 6%.

Growth in Microsoft Cloud Revenue, Including Azure, Slows; AI Business Annual Revenue Exceeds $13 Billion

The earnings report showed that Microsoft's revenue and profit in the fourth quarter maintained double-digit growth, although revenue growth slowed compared to the third quarter, it still exceeded analysts' expectations. What surprised Wall Street was that the total revenue growth of Microsoft's commercial cloud in the fourth quarter slightly slowed to 21%, compared to an expected flat growth of 22% from the third quarter.

The intelligent cloud business segment, including server products and Azure, also saw a slight slowdown in revenue growth in the fourth quarter, with a year-on-year growth rate down 1 percentage point to 19%. Revenue from Azure and other cloud services grew 31% in the fourth quarter, also lower than the 33% growth in the third quarter.

However, Microsoft revealed that in the fourth quarter, 13 percentage points of Azure's growth came from AI product contributions, higher than the analysts' expected 12.2 percentage points.

Microsoft CEO Satya Nadella highlighted the application of AI in business. He stated:

“We are innovating across the technology landscape, helping customers fully realize the return on their AI investments and seize the tremendous opportunities of the future. Our AI business annual revenue has exceeded $13 billion, a year-on-year growth of 175%.”

CEO: DeepSeek Has "Real Innovation"; Cost Reductions Will Make AI More Common, Good News for Microsoft

During the earnings call, CEO Nadella mentioned that DeepSeek's R1 has been integrated into Microsoft's platform Azure AI Foundry and the model catalog of GitHub, stating that developers will soon be able to run DeepSeek's R1 streamlined model locally on Copilot + PC, as well as run it in the vast GPU ecosystem on Windows.

When analysts asked about how the cost reductions brought by DeepSeek would affect the AI industry's Scaling Law, Nadella commented that DeepSeek "has some real innovations." He believes that the decline in AI costs is a trend.

Nadella responded that the development of AI is not much different from conventional computing cycles. Based on the Scaling Law, pre-training and inference time calculations accumulate continuously, and all software is like thisOver the years, we have seen significant improvements in the efficiency of AI training and inference. In terms of inference, we typically see the cost-performance ratio of each generation of hardware improve by more than 2 times, and the cost-performance ratio of each generation of models improve by more than 10 times.

When the Token price drops, the cost of inference computation decreases, which means people can consume more, and more Apps will be developed. Model optimization means that AI will become more ubiquitous, so for mega enterprises like Microsoft and PC platform providers like Microsoft, "this is all good news."