Microsoft’s reliance on OpenAI has become a defining feature of its cloud and AI strategy, with OpenAI now accounting for 45 percent of the company’s $625 billion revenue backlog—amounting to $281.3 billion in committed spending over the next several years. This figure, revealed in Microsoft’s second quarter of fiscal 2026 results, underscores how deeply intertwined the two companies have become. The bulk of this backlog stems from a new agreement signed in October, which added $250 billion to the total, significantly boosting Microsoft’s projected future revenue. While OpenAI’s massive investments in infrastructure—through partners like Oracle, CoreWeave, Nvidia, and AMD—and its fundraising from SoftBank suggest it has ample capital, there’s little indication it will diversify away from Azure anytime soon. Unless OpenAI suddenly requires far more GPU or XPU capacity than currently planned, it is likely to continue relying on Microsoft’s cloud for the foreseeable future. This creates a critical dependency: Microsoft is far more dependent on OpenAI than OpenAI is on Microsoft. This dynamic is highlighted by the fact that OpenAI’s portion of the backlog has grown rapidly since Microsoft’s initial $13 billion investment in the startup. Much of that capital, likely in the form of Azure usage fees, was effectively recycled back into Microsoft’s business during the generative AI boom. Though exact figures remain opaque, the scale of this internal flow is substantial. Despite this, Wall Street remains cautious. Microsoft’s capital expenditures hit $37.5 billion in the quarter—up sharply from prior periods—and analysts expect spending between $100 billion and $125 billion in fiscal 2026, a significant increase from the $88.2 billion spent in fiscal 2025. The company aims to double its datacenter capacity by the end of fiscal 2027, which has fueled investor concerns about return on investment. However, the revenue from OpenAI is helping to offset these costs. Microsoft’s CFO Amy Hood noted that the company is using a large portion of its GPU capacity internally—training its own models, running Copilot services, and supporting initiatives like Microsoft Foundry—rather than leasing it out. If not for this internal consumption, Azure revenue growth could have exceeded 40 percent. Still, investors focused on the fact that Microsoft didn’t expand its external cloud offerings as aggressively as expected. Microsoft’s overall revenue rose 16.7 percent to $81.27 billion, driven by a 21.4 percent increase in services and other revenues. The Intelligent Cloud group, which includes Azure and Windows Server, reported $32.91 billion in sales—up 28.8 percent—with operating income of $13.78 billion. Our analysis suggests Azure cloud proper generated $21.95 billion in revenue, representing two-thirds of the Intelligent Cloud segment, and delivered $9.72 billion in operating income—up 44.8 percent. The broader Microsoft Cloud business, including software subscriptions, surpassed $50 billion for the first time, reaching $51.51 billion in sales. Meanwhile, Microsoft’s “real” systems business—combining Windows Server software and Azure infrastructure—generated $22.58 billion in revenue, up 31.9 percent, with operating income of $9.52 billion, up 30.9 percent. While the PC business continues to face headwinds from high memory and storage prices, the cloud and AI segments are proving resilient. Satya Nadella remains optimistic, framing the current moment as the early stages of widespread AI adoption, with transformative economic potential across the tech stack. But for many users—like the author of this piece—AI integration remains invisible or unwanted. The reality is that OpenAI’s spending is the primary driver of Microsoft’s cloud growth, and without it, the momentum would be far weaker.