![\"In](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/07\/1782936854_494_0x0.jpg\")
<\/p>\n<\/p>\n
GERMANY – 2026\/06\/10: In this photo illustration, the logo of productivity software Microsoft 365 is displayed on a smartphone in front of abstract background on computer screen. (Photo Illustration by Timon Schneider\/SOPA Images\/LightRocket via Getty Images)<\/p>\n
SOPA Images\/LightRocket via Getty Images<\/p>\n
This article was written by Doug Nathman, with research by his team at Trefis.<\/p>\n
A new business within Microsoft (MSFT)<\/a> has quietly grown to massive scale. The company’s AI segment has crossed a $37 billion annual run rate, expanding at an astonishing 123%. This new growth engine already represents a meaningful portion of Microsoft Cloud, which itself exceeded $54 billion in quarterly revenue. This is not a future promise; it is a current reality.<\/p>\n This rapid growth is why the upside case is centered on revenue. The Intelligent Cloud segment, which houses these AI services, grew 30% to become a $34.7 billion quarterly business. Continued compounding from this base is the main driver of the stock’s potential upside.<\/p>\n That is the story. The question is whether it is strong enough to drive meaningful upside from here, or whether today\u2019s price already reflects most of that optimism. Yes, but with caveats. A conservative 3-year scenario points to roughly 49%. Revenue compounding does the heavy lifting, while the multiple barely changes.<\/p>\n Here is the operational picture behind the math:<\/p>\n How Compounding Builds The Upside<\/p>\n Revenue compounds at 15.2% annually, lifting the top line from $318.3B to $486.5B over three years. That is a step down from the LTM 17.9% pace, since today’s acceleration is unlikely to extrapolate cleanly over a full three-year period.<\/p>\n Margins ease from 39.3% to 38.3% as today’s LTM level gives back slightly toward the longer-run average. Together, that lifts earnings from $125.2B to roughly $186.1B, a 49% increase.<\/p>\n The model assumes a constant price-to-earnings (P\/E) multiple of 21.9x, implying that earnings growth alone drives the projected valuation gain. Applying that multiple to higher earnings puts the stock near $547.83, with a market cap of $4.1T versus $2.7T today. That is roughly 49% above where the stock trades now.<\/p>\n Has revenue compounding been the lever behind MSFT’s recent move? See the lever breakdown<\/a>.<\/p>\n What Could Accelerate The Top Line<\/p>\n The next leg of growth could come from a fundamental shift in the business model, as management explained that per-user businesses will become both per-user and usage-based. With Microsoft 365 Copilot seat additions already up 250% year-over-year, layering consumption on top of this adoption curve creates a new, unmodeled revenue opportunity.<\/p>\n What Could Slow It Down<\/p>\n The main concern raised on the call is the sheer scale of investment needed to support this growth. Management expects to invest roughly $190 billion in capital expenditures in calendar year 2026 alone. This spending pace creates what one analyst described as a disconnect, making investors nervous about the timing of the return.<\/p>\n Is The Compounding Real?<\/p>\n For this case to play out, revenue needs to keep compounding near 15.2%, a step down from today’s 17.9% but still clearly positive. The multiple is not being asked to do anything dramatic, which makes the case more defensible. The projected margin also sits at or near the 3-year peak, so any move back toward the longer-run average would make the rest of the math more difficult.<\/p>\n While the shift to usage-based pricing provides a clear revenue catalyst, the planned $190 billion capital investment creates meaningful near-term risk.<\/p>\n Should You Invest In Microsoft?<\/p>\n A careful 3-year case on a single company is still a concentrated bet, as historical volatility<\/a> across past market crises shows. Investors who build analyses like this around individual positions often want the same framework applied across a diversified book, partly for discipline and partly because even the cleanest single-stock thesis can break for reasons the math does not capture.<\/p>\n