Is Microsoft Hiding Its True AI Costs? Breaking Down FY27 Q1

AI-generated, human-reviewed.

On Windows Weekly, the latest Microsoft earnings dominated the conversation, revealing that AI infrastructure investments are driving record spending, even as Microsoft reworks how it shares financial data with the world. The episode shed light on why Microsoft changed its business reporting structure, how much it's truly spending on artificial intelligence, and what this means for transparency, the PC market, and tech users in the years ahead.

How Microsoft’s New Business Structure Affects Financial Transparency

The discussion focused on Microsoft's decision to restructure from three to two top-level business units: "Agents and Infra" and "Devices and Consumer" categories. These changes bundle major business segments together, integrating cloud and AI services more tightly.

One concept discussed was how such consolidation can make it harder for industry watchers and investors to track the performance of key areas like Copilot and Windows, as revenues become blended across broader categories. The hosts explained that this approach limits visibility into specific divisions and can make long-term year-over-year comparisons challenging due to the lack of "apples to apples" breakdowns. Several points were raised, including the regulatory requirement that Microsoft provide comparable data for a limited number of past quarters, but after that period, the transparency largely disappears. This has become a common strategy among large tech companies seeking to control investor narratives.

Microsoft’s AI Spending: How Much and Why It Matters

A key theme that emerged was Microsoft’s massive investment in AI infrastructure, with reported capital expenditures (CapEx) reaching $45.2 billion—an increase over the previous year, but still potentially understating the true costs. The discussion explored how Microsoft’s accounting practices, such as spreading out the lifespan of hardware, can underreport current spending, while delays in component deliveries further distort the numbers. The hosts noted that, despite these tactics, Microsoft's CapEx is growing at an unprecedented rate, and that figure was only limited by supply chain issues—not a lack of intent to spend.

Microsoft rivals like Amazon and Google are also spending heavily, with Amazon planning $220 billion in CapEx and Google also incurring significant costs. All three companies are racing to build the AI infrastructure that underpins their cloud and AI offerings.

Where’s the Payoff? Tracking AI Revenue and Product Impact

The conversation highlighted the challenges of connecting AI spend to revenue. For example, Microsoft disclosed $10 billion in quarterly revenue from its partnership with OpenAI, representing over 10% of total revenues. But much of this income cycles back as infrastructure costs, making it hard to clarify profits. Copilot, Microsoft's flagship AI assistant for Microsoft 365, has seen slow uptake, with paid seats representing only a small percentage of the existing customer base 31:49. Notably, Microsoft discontinued reporting paid Copilot subscribers this quarter. It’s a signal, according to the hosts, that initial adoption may have stalled or not met targets.

Another insight was that much of the AI focus is still forward-looking. Many of the anticipated benefits, such as hybrid AI models that blend local and cloud processing, are in early development. So users are still struggling to find compelling everyday use-cases, especially considering the high costs.

The State of the PC Market and Windows

Several points were raised about the impact on Windows and the broader PC market. Revenues from PC makers and Surface dropped 27% year-over-year, highlighting continued challenges due to high component prices and delayed upgrades by businesses and consumers. Despite Windows remaining a pillar for the company, the new business segmentation hides much of its commercial strength within broader cloud and subscription categories, making it hard for the public to gauge the platform's standalone health.

A Window Into Microsoft Spending

• Microsoft restructured its business units to combine cloud, AI, and productivity into "Agents and Infra", making financial performance harder to trace.
• AI infrastructure spending hit $45.2 billion for the year, with actual costs likely higher due to accounting shifts and component delivery lags.
• OpenAI partnerships provided $10 billion in revenue, but the profitability is murky as much is offset by infrastructure costs.
• Microsoft stopped reporting Copilot paid users, suggesting slower than hoped adoption.
• PC market revenues, including Windows, dropped significantly, reflecting hardware pricing pressures and delayed upgrades.
• Rivals like Amazon and Google are spending even more on CapEx, intensifying the race to build AI-powered infrastructure.
• Transparency challenges mean the true health of individual business lines is obscured, especially as AI costs and revenues blend with other operations.

According to the hosts on Windows Weekly, Microsoft’s latest earnings reveal a company deep in transition—spending massively on AI and cloud, but making transparency more difficult for outside observers. For tech users and industry watchers, the big story is that while AI is the focal point for Microsoft’s future, it’s not yet delivering the clear product wins, revenue streams, or transparency that many might expect. Sustained investments and evolving business segmentation mean the full payoff (and the true picture) may remain hidden for several more quarters.

Stay informed and never miss an episode: Subscribe to Windows Weekly at https://twit.tv/shows/windows-weekly/episodes/1008