1 & 2. How much is AI costing, and is it profitable yet?
The single biggest story in AI is the staggering cost. Microsoft is on track to spend roughly $190 billion on capital expenditures (capex) in calendar year 2026, with a huge portion dedicated to AI data centers, GPUs, and other hardware. This massive
outlay is putting pressure on the company's free cash flow and gross margins, which have dipped due to the investment in AI infrastructure. As for profitability, the answer is complicated. The company's overall AI business reported an annual revenue run rate surpassing $37 billion as of the third quarter of fiscal 2026, a 123% year-over-year increase. However, Microsoft doesn't break out the specific profitability of products like Copilot versus the cost to run them, making a clean ROI calculation difficult for outsiders.
3 & 4. Are customers actually buying AI, and is it making them more productive?
The adoption numbers are impressive. As of mid-2026, Microsoft reported having over 20 million paid seats for its Copilot products, with more than 90% of Fortune 500 companies using the service. This shows strong initial uptake from large enterprises. The productivity gains also appear tangible. Microsoft's own research, supported by external studies like a UK government trial, shows users saving between 14 and 26 minutes per day. Users report completing tasks faster (73% of users), improving work quality (68%), and creating first drafts more quickly (85%). This data provides some of the first concrete evidence that AI tools are delivering measurable value in the workplace.
5 & 6. How is AI affecting the core cloud business, and is demand keeping up?
Far from cannibalizing its existing business, AI appears to be a major growth engine for Microsoft's Azure cloud platform. The company has maintained Azure growth near a blistering 40% rate through fiscal 2026, attributing a significant portion of that to AI services. In fact, demand is so high that Microsoft has reported being 'capacity-constrained,' meaning it can't fulfill all of its customer orders for AI computing power. This suggests the massive capex spending is a direct response to overwhelming demand, not speculative building. Azure remains a key indicator; if its growth slows while spending stays high, investor confidence could waver.
7 & 8. How deeply are employees adopting AI tools, and where are they using them?
While millions of licenses have been sold, adoption within organizations isn't automatic. Some data suggests that just under 36% of employees with access to Copilot use it actively. This highlights a gap between purchasing the technology and integrating it into daily workflows. For those who do use it, the activity is concentrated in communication and document-based apps. Daily use is highest in Microsoft Teams (34%) and Outlook (33%), followed by Word (25%). Adoption in more data-intensive applications like Excel has been slower. This shows AI is currently making its biggest impact on managing emails, meetings, and writing tasks.
9 & 10. What's the status of the OpenAI partnership, and how is regulation playing a role?
Microsoft's partnership with OpenAI remains a cornerstone of its AI strategy and a significant financial asset. The company's large stake in OpenAI has resulted in billions in net investment gains. However, the relationship has evolved, with reports indicating OpenAI is no longer exclusively committed to running on Azure, creating a new dynamic. On the regulatory front, AI companies are facing increasing scrutiny worldwide. While not a major topic in recent earnings calls, the need to build responsible AI and comply with emerging governance rules is a critical, long-term factor that will shape product development and operational costs.
11 & 12. Where is future growth coming from, and is the market getting saturated?
The next wave of growth is expected to come from expanding from large enterprises to smaller businesses and individual 'prosumers,' as well as the development of more sophisticated AI 'agents' that can automate complex workflows. Microsoft's massive backlog of contracted revenue—known as Remaining Performance Obligations—suggests a long runway of committed future spending from customers. The market is far from saturated; in fact, all major cloud providers report that supply, not demand, is the main constraint on growth. The focus is now shifting from simply providing access to models to proving tangible business value and moving from basic assistants to indispensable operational tools.
