Microsoft shares slide as OpenAI exposure and Azure capacity worries overshadow earnings beat

Microsoft

• Microsoft shares fell over 6 percent despite strong quarterly results
• Investors worry about Azure growth, high AI spending, and reliance on OpenAI
• OpenAI now accounts for 45 percent of Microsoft's $625B cloud backlog

Did our AI summary help?

Shares of Microsoft dropped more than 6 percent in after-hours trading on January 28, even after the company posted better-than-expected results for its second quarter. The reaction reflected investor unease about the outlook for Azure, mounting capital expenditure, and Microsoft’s increasing exposure to OpenAI, rather than the company’s near-term financial performance.

According to reporting by Business Insider, analysts and investors homed in on signs that Azure growth may be moderating just as Microsoft ramps up spending on AI infrastructure. Those concerns were amplified by disclosures around the scale of Microsoft’s future cloud commitments and how heavily they are tied to OpenAI, the maker of ChatGPT.

This was also Microsoft’s first earnings report since OpenAI completed a major restructuring and updated its partnership agreement with Microsoft, which owns roughly 27 percent of the AI company. While the revised structure was meant to clarify governance and long-term collaboration, it has also sharpened investor focus on how dependent Microsoft’s cloud business has become on a single, compute-hungry customer.

During the earnings call, analysts pressed CEO Satya Nadella and CFO Amy Hood on Azure’s growth trajectory. Microsoft disclosed that its remaining performance obligations, effectively its commercial bookings backlog, surged 110 percent year over year to $625 billion. Of that total, OpenAI now accounts for about 45 percent, a figure that raised eyebrows on Wall Street.

The scale of that concentration has become a source of concern. While OpenAI’s commitments provide long-term revenue visibility, analysts worry that such a large share of Azure demand tied to a single partner increases execution risk. This is especially relevant as Microsoft continues to face constraints on cloud capacity, particularly for advanced AI workloads that rely on scarce GPUs.

Nadella sought to reassure investors that Microsoft is not optimising its business around Azure alone. He emphasised that while winning Azure customers remains critical, it cannot come at the expense of other growth engines. He pointed to Microsoft 365, GitHub, and Dragon Copilot as incremental businesses with their own large total addressable markets, arguing that the company does not want to maximise just one segment while neglecting others.

Hood reinforced that message by outlining the trade-offs involved in allocating new hardware capacity. She said Microsoft must carefully decide how newly available GPUs and CPUs are divided between first-party products such as Copilot, internal research and development, and external Azure customers. Only the remaining capacity, she noted, can then be directed towards meeting growing Azure demand. For investors, that balancing act underscored how tight AI infrastructure has become, even for one of the world’s largest technology companies.

Microsoft is not alone in facing these pressures. OpenAI executives have repeatedly said that limited computing power has slowed progress across both product development and research. Although OpenAI has pledged to spend as much as $250 billion on Azure services over time, its leadership has acknowledged that a lack of available compute has forced difficult prioritisation decisions.

The market reaction suggests investors are increasingly sensitive to the risks associated with AI-led growth. While Microsoft’s earnings highlighted strong execution and expanding demand, the sell-off indicates that expectations for Azure and AI may now be so high that any sign of constraint, concentration, or slowing momentum is quickly punished.

For Microsoft, the challenge going forward will be convincing investors that its massive investments in AI infrastructure can support both OpenAI and a broad base of enterprise customers, without creating bottlenecks or overreliance on a single partner. Until that confidence is restored, strong quarterly numbers alone may not be enough to steady the stock.