Report Interpretation

Overview

BofA Global Research maintains a constructive long-term view of Microsoft, citing stronger Azure growth, increasing Copilot adoption and an AI strategy that combines specialized models, model-agnostic software controls and increasingly efficient infrastructure. It reiterates Buy and raises its price objective to $600 from $500.

Core views

Microsoft’s 4Q26 results reinforced BofA’s constructive AI thesis. Azure growth accelerated from 39% in 3Q26 to 43% in 4Q26, and management guided to 45% growth in 1Q27 versus the Street’s initial 40.6% expectation. Paid M365 Copilot seats exceeded 30 million, with net additions more than doubling quarter on quarter, while remaining performance obligations increased 84% year on year to $678 billion. BofA sees the combination of stronger cloud demand, expanding AI adoption and improved capacity availability as evidence that Microsoft can sustain Azure growth and convert contracted backlog into revenue. Its model forecasts Azure growth of 41.8% in FY27E, up from 39.9% in FY26. The report argues that Microsoft’s AI economics improve because it does not rely on one frontier model for every workload. The company is expanding first-party MAI models across reasoning, coding, security, image, voice and transcription, while continuing to use third-party frontier models where appropriate. Specialized models can handle high-volume, product-specific tasks with lower latency, token use and cost. MAI-Code-1-Flash reportedly delivers comparable performance to GPT-5.6 for common Excel tasks at lower cost; the newer MAI-Code-1.1-Flash reportedly uses 25% fewer tokens per task and costs one quarter of its prior version in GitHub Copilot. In security, MAI-Cyber-1-Flash handles up to 90% of MDASH vulnerability-identification and remediation tasks, with the combined system delivering comparable performance at 50% of the cost of leading models. BofA believes this workload routing can improve the economics of enterprise AI applications. Copilot is central to that strategy because BofA views it as a model-independent enterprise orchestration layer. Its harness supplies enterprise context and memory, connects models to data and tools, governs permitted actions and evaluates output quality. This allows tasks to be routed among frontier, specialized and customer-developed models based on quality, latency, cost and compliance. The report therefore argues that Copilot’s value need not depend on Microsoft owning the leading general-purpose model for every task; instead, Microsoft can monetize through applications and consumption while Azure supports the underlying workloads. BofA also identifies four infrastructure levers supporting AI returns. First, engineering and software improvements across the CPU and GPU fleet increased Copilot workload throughput fourfold year to date; with demand exceeding supply, BofA says the additional output was quickly monetized and contributed to Azure’s 4Q outperformance. Second, Microsoft reduced the time to operationalize new GPUs in its largest regions by nearly 50% over the past year, enabling capacity to generate revenue faster. Third, first-party Maia 200 silicon complements NVIDIA and AMD hardware; MAI models co-designed for Maia 200 reportedly deliver 40% better performance per watt. The report notes a potential significant Maia 300 production ramp in 2027 and emphasizes that Maia can support multiple model families, including OpenAI and MAI models. Fourth, Microsoft added about 1GW of capacity in 4Q26 and aims to roughly double total capacity over the next two years. Fairwater’s interconnected AI datacenters are intended to improve workload allocation and utilization; the first Wisconsin facility was completed in June 2026, with a second expected in 2028. BofA believes these model and infrastructure efficiencies can lower AI serving costs, support margins and improve ROIC over time. It expects sustained mid-double-digit growth over the next three years, led by Azure, cloud-based Office 365 and AI solutions, while projecting that operating-expense scale can support roughly 20–50 basis points of annual margin expansion despite a mix shift toward lower-margin Azure and M365. The price objective is raised to $600 from $500 and is based on 28x BofA’s CY27E P/E, versus 24x previously. BofA considers the 28x multiple justified by sustained revenue growth and margin characteristics, despite being above the stated 18–25x peer range.

Analysis framework

The report links quarterly cloud and Copilot indicators to Microsoft’s AI monetization outlook, then examines how specialized models, model routing and infrastructure execution can improve capacity utilization and serving costs. It translates those operating assumptions into a higher P/E valuation multiple and price objective.

Methodology notes

• Valuation methodsP/E and PEG Valuation

  Price-to-earnings valuation

  BofA sets its $600 price objective using a 28x CY27E P/E multiple, increased from 24x, and compares it with an 18–25x peer range.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Microsoft Corporation (MSFT)

  Primary covered company; BofA sees its AI strategy and execution as supporting higher valuation.

  Strengths

  Accelerating Azure growth, more than 30 million paid M365 Copilot seats, diversified MAI models, model-agnostic Copilot orchestration, and expanding AI capacity.

  Weaknesses

  A revenue mix shift toward lower-margin Azure and M365 may offset some scale benefits in the near term.

  Comparison

  The 28x CY27E P/E used for the price objective is above BofA’s stated 18–25x peer-group range.

  Risks

  Near-term gross-margin pressure, faster innovation by AI application and model providers, and cyclical enterprise application spending.

Key data

• Azure growth43% YoY in 4Q26Accelerated from 39% in 3Q26; management guided to 45% for 1Q27.
• Paid M365 Copilot seatsMore than 30 millionNet additions more than doubled quarter on quarter.
• Remaining performance obligations$678 billionRPO increased 84% year on year.
• FY27E Azure growth41.8%BofA forecast, compared with 39.9% in FY26.
• Copilot workload throughput4x year to dateAttributed to CPU/GPU fleet engineering and software optimization.
• New GPU operationalization timeNearly 50% reductionImprovement in Microsoft’s largest regions over the past year.
• Maia 200 efficiency40% better performance per wattReported for MAI models co-designed with Microsoft’s internally developed AI chip.
• Price objective valuation28x CY27E P/ESupports the $600 price objective; the prior multiple was 24x.

Impact & implications

BofA’s higher valuation rests on the view that accelerating Azure growth, Copilot adoption and better AI cost economics can support sustained growth, margins and returns. Faster capacity deployment and more efficient workload routing are presented as mechanisms that make Microsoft’s substantial AI investment more monetizable.

Risks

• Near-term gross-margin pressure could arise as Azure becomes a larger share of total revenue.
• AI application and model providers could innovate faster than Microsoft, potentially limiting its growth profile.
• Enterprise application spending is cyclical because application projects are relatively discretionary.