Microsoft Corporation (NASDAQ: MSFT), initiation of coverage, Nov 8, 2026
Microsoft’s cloud keeps growing, but the cost of the AI build leaves little upside
Microsoft is now mainly a cloud company: Azure and Microsoft 365 cloud made up $202bn of its $332bn sales in FY26. We agree with analysts on sales, but expect depreciation from the AI build to cut FY28 profit margins.
- Our rating
- Hold
- 12-month target
- $545.00
- Price, Nov 6, 2026
- $530.61
- Expected return, with dividend
- +4%
By Penmark Team
What Microsoft is now
Two cloud businesses now bring in most of Microsoft’s sales
Azure, billed on usage, and Microsoft 365 cloud, billed per seat a year in advance, made up $202bn of $332bn sales in FY26. Of every $100 of sales, $47 became operating profit.
Azure is the engine. Its sales were $101.9bn in FY26, and we forecast $147.3bn in FY27. By FY28 it would be 42% of all sales, up from 31% in FY26.
FY26 is the year to 30 June 2026. Source: company filings.
Fiscal years end 30 June. Source: company filings, Penmark Research estimates.
What you are really buying
Today’s price pays for analysts’ margins, faster Azure growth and a richer OpenAI stake
Our base value is $489 a share. Moving to the market’s view on margins adds $54, on Azure growth $40 and on the OpenAI stake $13.
That would put the value at $596. A lower required return, which the market accepts for the AI build, brings it back to the $531 price. Margins are the biggest piece, and the one we dispute.
Required return is the equity risk premium in the discount rate; steps applied in sequence to the base case. Source: Penmark Research estimates.
The bull case
Cloud demand still outruns supply, with a backlog of $677bn
Azure grew 48% in the quarter to September 2026, and contracted backlog reached $677bn, up 73% on a year. Microsoft is limited by data-centre capacity, not demand.
In the good case Azure grows 47% in FY27 and 40% in FY28, and the OpenAI stake is valued at $1.25tn. That case is worth $681 a share.
Backlog is commercial remaining performance obligation (RPO). Fiscal years end 30 June. Source: earnings calls and quarterly filings.
The bear case
Depreciation from the AI build more than doubles and squeezes profit margins
About $350bn of spending on data centres and chips across FY26 and FY27, 72% of it short-lived, lifts depreciation from $38.5bn to $82.0bn by FY28.
That cuts our FY28 operating margin to 43.7%, against about 47% in analysts’ forecasts. Our FY28 earnings per share of $21.90 are 9.3% below analysts’ $24.14.
Depreciation and amortisation (D&A). Fiscal years end 30 June. Source: company filings, Penmark Research estimates.
What decides it
The FY28 margin is worth $54 a share and settles the call
The share price has tracked operating margin closely over the past eight quarters. Sales are not the dispute: we are only 1.5% below analysts on FY28 sales.
The gap is cost. At analysts’ margin our base value would be $543, not $489. An FY28 outlook with margin above 46% would move us to the top of the Hold range.
Analysts’ margin is implied by consensus EPS of $24.14 on revenue of $477bn. Fiscal years end 30 June. Source: company filings, consensus aggregators, Penmark Research estimates.
The odds
Downside of $259 outweighs upside of $150, so we stay at Hold
We weight three cases 20%, 50% and 30% to a value of $503, 5.2% below the price. Rolled forward 12 months, that gives our $545 target.
Even the middle case sits below today’s price. Only the good case, with 30% weight, is above it, so the odds do not justify buying.
- Good, 30%Azure grows 47% in FY27 and 40% in FY28, and OpenAI is valued at $1.25tn.
- Middle, 50%Azure grows 44.5% then 35%, and depreciation cuts the FY28 margin to 43.7%.
- Bad, 20%OpenAI trims its commitments, Azure growth slows to 24% in FY28 and costs stay fixed.
Circle size shows how likely we think each case is. Case values are today’s DCF values; the target weights all three and adds 12 months of expected return. Source: Penmark Research.
What would change our mind
An FY28 outlook with margins above 46% would lift our view
Three of the four debates resolve by late January 2027. The OpenAI funding round should close in December, and Microsoft reports Azure growth on 27 January.
The test that matters most comes in July 2027, when Microsoft gives its FY28 outlook. We expect a 43.7% operating margin; analysts expect about 47%.
Our $852bn basis for OpenAI is 32% below the $1.25tn first close, worth $13 a share.
Azure growth is guided at 49% to 50%; we model 44.5% for FY27.
Limits on new data centres would add $3.50 a share in the good case.
We expect a 43.7% operating margin against about 47% for analysts.
Inside the full report
What each page covers after the summary on pages 1 and 2. Each opens the full report at that page.
See all 13 pages
- 3Returns stay above 24% ROE, but margins and FCF are where we differ from consensus: FY28E EPS is 9.3% lower
- 4Four dated events decide the call, and at 27.7x trailing earnings the stock is one deviation below its own average
- 5The stock trades on operating margin (0.64 correlation), and our FY28 margin is 3 points below the Street
- 6The call rests on one debate worth $54 a share: whether depreciation pulls FY28 margin toward 44%
- 7Azure and Microsoft 365 cloud are 61% of revenue and all of the FY27E growth, at a 47% operating margin
- 8The stock trades on the multiple, not on earnings: two drivers, FY28 margin and Azure growth, explain it
- 9We forecast 19% revenue growth in FY27E, in line with guidance; depreciation then cuts margin by 2.7 points in FY28E
- 10Value of $503 a share is 5% below the price: DCF and P/E agree at $489 and $487, and the weighted value sets a $545 target
- 11The next nine months bring five dated tests; the FY28 outlook in July 2027 is the one that can end the Hold
- 12Ownership is index-heavy and the Street is 88% Buy; short interest is falling
- 13Capital allocation has shifted from returns to the build: capex is 81% of operating cash flow in FY27E, up from 27% in FY22A
- 14The model: revenue compounds at 18% a year to FY29E while operating margin falls 4.8 points from FY26A
- 15Reference: every figure traces to the model, with a DCF primary method and a P/E cross-check that agree within 0.5%
Penmark Research. This summary draws on our initiation report on Microsoft Corporation, published Nov 8, 2026. It is not investment advice. Analyst certification, rating definitions and important disclosures are in the full report.