Microsoft reported Q2 fiscal 2026 revenue of $81.3 billion, topping consensus estimates by roughly $1 billion, and non-GAAP EPS of $2.98, exceeding analyst forecasts by $0.07. However, growth in Azure and other cloud services slowed to 39%, missing the 39.4% StreetAccount consensus. The More Personal Computing segment guided to approximately $12.6 billion in Q3 revenue, below the $13.7 billion consensus, and implied operating margins were narrower than expected. CFO Amy Hood noted that reallocation of recently deployed GPUs for in-house AI workloads constrained cloud sales growth.

On the day of the release, Microsoft shares plunged nearly 10%, the steepest decline since March 2020, erasing about $357 billion in market capitalization and contributing to a 0.7% drop in the Nasdaq Composite. Investors focused on the company’s plan to invest heavily in AI infrastructure, with capital expenditures expected to remain elevated following a record $37.5 billion outlay in the quarter. Market participants questioned whether the long lead times on data-center builds and GPU availability would delay returns on that investment.

Following the sell-off, analysts split on the stock’s outlook. Goldman Sachs maintained a Buy rating, citing Microsoft’s 1-in-1-billion false positive biometric platform and cross-selling potential in security and Entra ID. Barclays acknowledged near-term margin pressure but cited the company’s $3.2 trillion market cap positioning as a defensive AI platform play. Meanwhile, some strategists at Wedbush and BlackRock argued that capex intensity could suppress free cash flow until late FY2027, suggesting investors wait for clearer evidence of AI-driven revenue acceleration.

Management reiterated that in-house AI chips, such as the Maia 200, will augment rather than replace Nvidia and AMD hardware purchases, preserving supply partnerships while improving unit economics. CEO Satya Nadella emphasized that prioritizing internal model development will ultimately raise Azure’s AI usage curve. Microsoft forecasts that AI workloads could represent over 50% of cloud consumption by FY2028, underpinning a multi-year growth runway even if near-term cloud growth moderates.