MSFT is the next ADBE

ADBE trades at ~14x trailing earnings. Two years ago it was 40x+. The market re-rated it from a compounding SaaS monopoly to a mature software business losing share to AI-native competitors. Despite this sub’s affinity, I think MSFT is on the same trajectory, and the math on a sum-of-the-parts basis makes it hard to justify the current ~$2.65T market cap even at a compressed 22x P/E.

-----

THE SOTP PROBLEM

@chat_SBC put it concisely (https://x.com/chat_sbc/status/2037539351115223269?s=46): “MSFT SOTP is basically office = ADBE (9x p/e) + bing/windows = yelp/ibm (5-7x p/e). need big multiple for Azure to make math work.”

Let me put numbers on it.

MSFT FY2025 segment operating income (fiscal year ended June 2025):

Productivity & Business Processes — Revenue: $120.8B — Operating Income: $69.8B

Intelligent Cloud — Revenue: $106.3B — Operating Income: $44.6B

More Personal Computing — Revenue: $54.6B — Operating Income: $14.2B

Total — Revenue: $281.7B — Operating Income: $128.5B

Now value each piece by what it actually looks like competitively:

Productivity & Business Processes (~$69.8B OI): This is Office/M365, LinkedIn, Dynamics. Office is the crown jewel, but Google Workspace now holds ~50% of the productivity software market by domain count vs. M365 at ~45%. Microsoft still dominates enterprise seats (~58% of Fortune 500), but the trend is moving against them, particularly among SMBs and new company formations that default to Google. This segment is growing at 13-16%. Solid, but increasingly resembles ADBE’s trajectory: a legacy monopoly being chipped away at the margins by cloud-native, AI-integrated competitors. At an ADBE-like 9-10x operating income, this is worth ~$630-700B.

More Personal Computing (~$14.2B OI): This is Windows OEM, Xbox, Bing/Search, Surface. Xbox hardware revenue fell 32% YoY in the most recent quarter. Xbox content and services down 5%. Gaming revenue declined 9% overall. Windows OEM is roughly flat. Bing has never achieved meaningful market share. These are IBM/Yelp-tier businesses. Low or no growth, commodity products, losing relevance. At 5-7x operating income: ~$70-100B.

Intelligent Cloud (~$44.6B OI): This is Azure, SQL Server, Windows Server. Azure grew 39% in the most recent quarter, which is strong. But here’s the issue: to get from the ~$700-800B implied by Productivity + Personal Computing to the current $2.65T market cap, you need Azure valued at roughly $1.85-1.95T. That implies Azure alone is being valued at ~42-44x operating income. For a business whose margins are compressing.

THE MARGIN PROBLEM

Azure’s economics are going the wrong direction. Microsoft Cloud gross margin has declined from 72% to 67% over the past year as AI infrastructure scales. Capex in Q2 FY2026 alone was $37.5B. Annualized, that’s $150B, nearly consuming 100% of operating cash flow. FY2025 FCF was $71.6B, already declining 3.3% YoY even as revenue grew 15%. The capex-to-OCF ratio is approaching hyperscaler territory where free cash flow generation is impaired for years to come.

Management explicitly acknowledged on the Q2 earnings call that capex is growing faster than Azure revenue. Investors are right to question ROI.

THE OPENAI LIABILITY

Microsoft’s $625B cloud backlog is ~45% tied to OpenAI. OpenAI is structurally unprofitable. ~33% gross margins, projected $14B in losses in 2026 alone, not expected to be cash-flow positive until 2030. Cumulative losses through 2029 are projected at $115B+. Under the renegotiated partnership, OpenAI pays Microsoft 20% of revenue through 2032, but this is a revenue share on a business that is bleeding cash faster than it grows.

OpenAI sells a commodity (inference), cannot meaningfully monetize the enterprise (only 5% of ChatGPT users pay), and is losing the talent war to Anthropic (engineers are 8x more likely to leave for Anthropic than the reverse). Its GPT-5 launch was poorly received. It is now pivoting to advertising, which is the ultimate admission that the subscription/API model isn’t working.

Microsoft is subsidizing a company that is, in my view, terminally incapable of generating the returns needed to justify the circular economics: Microsoft invests in OpenAI, OpenAI buys Azure, Microsoft counts it as cloud revenue. Strip out the OpenAI-related backlog, and the Azure growth narrative looks very different.

WHY THIS LOOKS LIKE ADBE

ADBE’s re-rating happened when the market realized three things simultaneously:

1. Core monopoly was eroding — Figma, Canva, and AI-native design tools chipped away at Creative Cloud’s moat, just as Google Workspace is chipping away at Office.

MSFT is earlier in this arc but the ingredients are identical:

- Office market share eroding (Google Workspace 50% vs. M365 45% by domains)

- Gaming in structural decline (Xbox hardware -32%, content/services -5%)

- Windows/Bing are zombie businesses (flat to low-single-digit growth)

- AI capex is a margin headwind, not a tailwind, with cloud gross margins down 500bps

- The AI premium on the stock depends entirely on Azure sustaining 35%+ growth AND eventually converting $150B/yr in capex into proportionate FCF, which I think is highly unlikely

VALUATION SUMMARY

Productivity & Business — OI: $69.8B — Multiple: 9-10x — Implied Value: $630-700B

More Personal Computing — OI: $14.2B — Multiple: 5-7x — Implied Value: $70-100B

Intelligent Cloud — OI: $44.6B — Multiple: 20x (generous) — Implied Value: $890B

Total implied value: ~$1.59-1.69T

At 20x operating income for Intelligent Cloud, which is generous for a business with compressing margins and capex that may exceed OCF, the SOTP implies ~$1.6-1.7T, or roughly $215-230/share. That’s 35-40% below the current price.

The market is pricing Azure at 40x+ OI. For that to be justified, Azure needs to sustain 35%+ growth for 5+ years AND margins need to re-expand AND the $150B/yr capex cycle needs to convert into durable FCF. If any one of those assumptions breaks, MSFT follows ADBE down to a multiple that reflects what it actually is: a mature, diversified software conglomerate with one high-growth segment that is spending faster than it earns.

-----

Disclosure: No position in MSFT or ADBE. This is not investment advice.

Edit: I never said MSFT doesn’t have switching costs. Still doesn’t change the fact that Azure is massively overvalued.