Microsoft Azure Is Growing Faster Than AWS — And Copilot Has Only Just Begun

Microsoft's AI monetisation story is the clearest proof of concept in enterprise technology. In the span of three years, the company transformed OpenAI's research capability into a commercial product line — GitHub Copilot, Microsoft 365 Copilot, Azure OpenAI Service, Copilot Studio — that collectively generated over $10 billion in annualised recurring revenue by mid-2026. No technology company has monetised an external AI partnership as effectively, or as quickly.

The headline number that commands attention is Azure's growth rate. In Q3 FY2026, Azure and other cloud services grew 35% year-over-year. For context, Amazon Web Services grew 17% in the same period. Google Cloud grew 28%. Microsoft has built the fastest-growing major public cloud, and the primary driver is AI workloads — both the training and inference compute consumed by Microsoft's own AI products, and the growing volume of enterprise customers building AI applications on Azure OpenAI Service.

This article makes the case that Microsoft's competitive position in enterprise AI is stronger today than its position in enterprise software was in 2010 — and that the monetisation of Copilot has barely begun.

The OpenAI Partnership: Not a Dependency — a Distribution Engine

Microsoft's $13 billion investment in OpenAI is widely described as a dependency risk. The bears argue that if OpenAI were to produce inferior models, lose key researchers, or pivot away from the Microsoft partnership, MSFT's AI advantage would evaporate. This misunderstands the nature of the relationship.

Microsoft does not merely resell OpenAI's models. It has embedded them into the core of its existing enterprise software stack — Outlook, Word, Excel, PowerPoint, Teams, Dynamics 365 — which is used by approximately 400 million commercial seats globally. The channel advantage is enormous: Microsoft can distribute AI features to 400 million users through a software update, without requiring a single new customer acquisition.

Azure OpenAI Service, which allows enterprises to deploy OpenAI models on private Azure infrastructure, is growing faster than any other Azure service. It enables enterprises to use GPT-4o and o3 models on their proprietary data without that data leaving their Azure tenant — a critical requirement for regulated industries including healthcare, financial services, and government. Microsoft handles the model hosting, the security perimeter, the compliance certifications, and the billing. OpenAI provides the model weights. This division of labour is as much Microsoft's advantage as OpenAI's.

Copilot 365: The $30 Per Seat Per Month Flywheel

Microsoft 365 Copilot costs $30 per user per month, layered on top of a standard M365 Business or Enterprise subscription. At that price point, it is the highest-margin product Microsoft has ever sold — the incremental cost of serving Copilot features over an existing M365 tenant is largely the inference compute cost, which scales at a fraction of the pricing.

Adoption has accelerated significantly since Microsoft simplified the deployment requirements in late 2025, removing the minimum seat requirement that previously blocked smaller enterprise customers. Fortune 500 adoption rates are now consistently above 40% of eligible employees at companies that have deployed Copilot at all, according to Microsoft's own data released at its March 2026 developer conference.

The compound effect is substantial. An enterprise with 10,000 M365 seats that deploys Copilot to 40% of employees generates $1.44 million in incremental annual ARR for Microsoft from that one customer — at near-zero marginal cost of acquisition, since the sales relationship already exists. Multiplied across Microsoft's enterprise base of hundreds of thousands of commercial customers, the Copilot revenue potential is measured in tens of billions of dollars annually at maturity.

The $80 Billion Capex Bet: Building the AI Utility

Microsoft announced plans to invest $80 billion in AI infrastructure in FY2026 — the largest single-year capital expenditure commitment in the company's history. This capital is being deployed across two primary uses: Nvidia B200 GPU clusters for training and inference, and Azure data centre expansion across 60+ countries.

The scale of this investment is both a strength and a risk. On the bullish side, data centre capacity takes 18–24 months from announcement to production, meaning Microsoft's FY2026 capex is building the capacity that will power FY2028 revenue. Enterprises signing multi-year Azure contracts today are locking in against a supply-constrained future, creating pricing power for Microsoft as demand grows faster than capacity in the near term.

On the bearish side, $80 billion in annual capex requires commensurately large revenue growth to maintain returns on invested capital. Microsoft's cloud gross margins (approximately 72%) must sustain the infrastructure investment — and if AI demand growth moderates, the depreciation burden from this capex cycle will weigh on earnings for years.

GitHub Copilot: Capturing Developer Spend

GitHub Copilot, the AI code completion and generation tool embedded in VS Code, JetBrains, and other editors, has over 1.8 million paid subscribers as of Q1 2026, growing 55% year-over-year. The enterprise tier, which includes Copilot Business and Copilot Enterprise, now accounts for the majority of GitHub Copilot revenue and is growing faster than the individual tier.

Beyond direct subscription revenue, GitHub Copilot serves a strategic purpose in Microsoft's competitive moat: it embeds Microsoft tooling into the daily workflow of the world's software developers. A developer who writes code with GitHub Copilot, hosts repositories on GitHub, runs CI/CD pipelines through GitHub Actions, and deploys applications on Azure is deeply integrated into Microsoft's ecosystem — and the switching cost of that integration is prohibitively high for most organisations.

The Activision Wild Card: Gaming as AI Testbed

Microsoft's $75 billion acquisition of Activision Blizzard closed in October 2023 and is now fully consolidated. The gaming division — including Call of Duty, Diablo, Overwatch, World of Warcraft, and the Xbox Game Pass subscription — generated approximately $25 billion in annualised revenue in FY2025.

Beyond revenue, Activision provides Microsoft with something strategically valuable: a real-time AI testing environment. Microsoft is deploying AI-generated NPCs (non-player characters), AI-assisted game design tools, and AI-powered matchmaking at scale across its game titles — generating hundreds of millions of daily AI inference calls that improve model reliability, latency tolerance, and personalisation algorithms. These learnings transfer directly to enterprise Copilot product development.

Conclusion: The Enterprise AI Standard

Every major enterprise technology shift over the past 40 years has produced a dominant platform player: IBM in mainframes, Oracle in databases, SAP in ERP, Salesforce in CRM. The AI era is following the same pattern, and Microsoft is the leading candidate for that role in enterprise AI infrastructure.

The combination of Azure's AI compute scale, Copilot's distribution into 400 million M365 seats, GitHub's developer ecosystem, and the OpenAI model partnership creates a competitive moat that no single competitor can replicate quickly. Google Cloud is the strongest challenger, but lacks Microsoft's enterprise relationships. Amazon AWS is larger but lacks the application layer. OpenAI lacks the distribution.

Microsoft is not just selling AI tools. It is becoming the enterprise operating system of the AI era. For investors with a three-to-five-year horizon, that positioning is worth paying a premium to own.