Morgan Stanley: Market Underestimates Microsoft's Azure Acceleration and Copilot Monetization
Taylor Wilson
Morgan Stanley sets a $600 price target for Microsoft, implying roughly 50% upside from the current share price, arguing that Azure's capacity unlock and Copilot's three-engine monetization remain under-reflected in the stock.
Why does Morgan Stanley think Microsoft is undervalued?
Microsoft trades at $402.29, implying a forward P/E of only about 16× on the bank's FY2028 EPS estimate of $23.86.
This means → the market is pricing a tech leader with 20%+ earnings growth as if it were an ordinary compounder.
Applying a 1.2× PEG — a ratio that divides the P/E by the earnings growth rate — Morgan Stanley arrives at a fair value near $600.
The bank did trim its 12–18-month target from $650 to $600, citing gross-margin pressure, higher capex, and rising debt — still bullish, but a notch more cautious.
What was holding Azure back, and what is changing?
Azure's main bottleneck over the past year was not weak demand but constrained supply — not enough GPU capacity to deliver.
CFO Amy Hood disclosed that had all newly onboarded GPUs in Q1 and Q2 been allocated to Azure, growth would have exceeded 40%, not the reported 38% (constant currency).
In plain terms = real Azure demand was stronger than the headline number showed; supply simply couldn't keep up.
As new capacity comes online, Morgan Stanley projects FY2028 and FY2029 Azure and cloud revenue at $214.9 billion and $305.9 billion — 5% and 7.8% above consensus.
What are Copilot's "three engines"?
Morgan Stanley breaks Copilot monetization into three paths: ① direct M365 Copilot seat expansion; ② enterprise migration to the higher-priced E7 subscription; ③ consumption-based billing via AI Agents and workflow automation.
This means → Copilot is not just "one more software add-on" — it lifts average revenue per user (ARPU) from multiple directions at once.
E7 bundles E5, Copilot, and Agent365 together. Morgan Stanley sees this as analogous to the earlier E3-to-E5 upgrade wave, potentially triggering a multi-year enterprise software upgrade cycle.
The bank's latest CIO survey: 47% of enterprises now use E5, 7% use E7; within a year, E7 adoption is expected to reach 21%.
Are enterprises actually willing to pay?
88% of CIOs surveyed plan to deploy M365 Copilot in the next 12 months, up from 80% in the prior survey and 72% the year before that.
This reflects a quarter-over-quarter acceleration in enterprise AI purchasing intent — not a one-off spike.
Morgan Stanley accordingly raises its Copilot revenue forecast: roughly $4.4 billion in FY2026, growing to about $22.5 billion by FY2029.
Gross margins are falling — can profits still grow?
Morgan Stanley cuts its FY2027–FY2029 gross-margin forecasts to 65.7%, 64.4%, and 63.4%, driven by higher AI-related depreciation and front-loaded infrastructure spending.
In plain terms = Microsoft is spending more up front, so the "pure profit" slice of each dollar earned is getting thinner.
Yet the bank argues disciplined opex control offsets the margin drag. Operating-margin forecasts stand at 46.5%, 46.7%, and 47.2% — a gentle upward trend.
Amy Hood said on the April 2026 earnings call: "AI gross margins are better than what we experienced during the cloud transition, and have remained so throughout."
Where could this bull case go wrong?
Azure acceleration depends on new capacity arriving on schedule. Any GPU delivery delays or softer-than-expected demand could push growth below Morgan Stanley's optimistic projection.
The three Copilot engines must fire in parallel: if seat expansion, E7 upgrades, or consumption billing underperform individually, ARPU growth stalls.
This means → Morgan Stanley's logic chain is internally coherent, but every link needs validation in the coming quarters' earnings reports.
Content is for reference only, not financial advice.