Microsoft: A Deep Value Investment Study

Hunting ten-year five-baggers among great growth stocks — pressing the upside question: "Can it get much bigger?"

• QHow large is its market ceiling? Is it expanding an existing pie, or creating an entirely new market?7/10

  A

  Bottom line first: Microsoft's ceiling is extremely high, but it is mainly expanding and reallocating several already large existing pies: enterprise productivity, cloud infrastructure, developer tools, and enterprise security. It is not creating an entirely new market out of thin air. The only layer that can truly be called a new market is AI agents/Copilot, and today it looks more like an incremental paid layer added on top of existing platforms than a standalone new continent. Measured by Baillie Gifford's LTGG yardstick, the ceiling is large enough to support sustained growth, but it is hard to support the kind of nonlinear expansion where a company grows from a small pond into an ocean, because Microsoft is already a leader in these pools. Further upside depends more on penetration and value per customer than on markets emerging from zero.

  Start with how large these pies are, and where Microsoft sits. Cloud infrastructure is the fastest-growing one: in Q1 2026, global cloud infrastructure services spending grew 35% year over year to about $129 billion, with Amazon, Microsoft, and Google holding roughly 28%, 21%, and 14% shares, respectively. This shows the cloud market itself is still expanding rapidly, but Microsoft is already a firmly established No. 2 giant. It is taking a large fixed slice of a still-growing big pie, not opening a new pie. Enterprise productivity is the opposite: it is a huge but mature pie. By Gartner's measure, Microsoft 365 has about 77% share in SaaS enterprise productivity, making it the absolute leader. The ceiling here lies more in price increases, selling more seats, and adding AI inside the suite than in market expansion.

  The part that can genuinely be called creating a new market is turning AI into a new usage layer that can be priced separately. The data supports the view that a new pie is forming: Microsoft disclosed in FY2026 Q3 that AI annual revenue run rate had exceeded $37 billion, up 123% year over year, and paid Microsoft 365 Copilot seats had exceeded 20 million. But two points need to be stated plainly. First, this $37 billion is still an increment layered on top of the existing Azure and 365 pies. In substance, it is getting existing customers to pay more for new capabilities, closer to raising ARPU in existing markets than creating a wholly independent new market. Second, the report also stresses that the AI profit pool may ultimately flow more to chips, energy, and specialized compute lessors, and Microsoft may not capture all the excess returns from this new market.

  So back to Baillie Gifford's key question: is the market large enough to hold a fivefold return over ten years? Microsoft's addressable market across cloud, productivity, security, developers, and AI agents is at the trillion-dollar level and still growing. Absolute size is not the constraint. The constraint is that Microsoft is already too large. The report gives FY2025 revenue of $281.7 billion and TTM revenue of about $318 billion. On that base, building another Microsoft requires the new AI pie not only to be real, but also to be priced and monetized efficiently under Microsoft's leadership. Overall judgment: the ceiling is high enough that it is not a hard cap on growth, but the main source of growth is expanding and recutting existing pies, with creating a new AI market as a secondary driver. The latter is an upside option, not an already realized fact.

  Jun 11, 2026
• QCan its revenue at least double over the next five years? Will growth be driven mainly by volume, price, or new businesses?5/10

  A

  Bottom line first: the probability that Microsoft's revenue will at least double over the next five years is not high. With current TTM revenue of about $318 billion, doubling means reaching about $636 billion in five years, implying roughly 15% compound growth sustained for five years. That is a very aggressive ask for a $3 trillion giant with revenue already above $280 billion. A more realistic picture is 60%-90% growth over five years, or roughly 10%-14% compound growth, driven jointly by volume, meaning seats and compute consumption, and price, meaning price increases, with AI as a new business providing upside elasticity. But the new business alone is unlikely to double the whole company. In Baillie Gifford terms: Microsoft is an excellent compounding machine, but it is already too large to accelerate easily in a doubling pattern. Doubling is the upper edge of an optimistic scenario, not the neutral expectation.

  Start with the base and recent performance. The report is consistent with official data: FY2025 revenue was $281.7 billion, up 15% year over year, and by FY2026 Q3 quarterly revenue had reached $82.9 billion, up 18% year over year (15% in constant currency). The current company-level growth rate is indeed high, but it is largely pulled up by AI/Azure and comes with a high-base effect. Keeping 15%-18% growth intact for five years is not easy. Historically, Microsoft's company-wide growth has more often fluctuated in the 10%-16% range.

  The three sources of growth do not carry equal weight:

  Volume is the main engine, but it is more about penetration and consumption than new customer count. Microsoft is already a leader in enterprise productivity and cloud, so new growth comes more from existing customers using more. In FY2026 Q3, Azure and other cloud services revenue grew 40% (39% in constant currency), reflecting an explosion in compute consumption. Microsoft 365 commercial seats grew by about 12% year over year in 2025, showing steady seat expansion. The report also points to commercial remaining performance obligation as a very strong forward indicator. Official FY2026 Q3 disclosure showed RPO had reached $627 billion, up 99% year over year (including the large OpenAI contract), with about 25% expected to be recognized over the next 12 months. This large backlog gives decent visibility into volume growth, but it includes a huge Azure commitment related to OpenAI, and the pace of recognition is constrained by capacity, including power and data centers.

  Price is a stable secondary engine: rigid, but limited in magnitude. The report notes that part of FY2025 Microsoft 365 Consumer revenue growth came from the January 2025 price increase, and Microsoft has also announced that it will raise prices for commercial and government Microsoft 365 suites starting in July 2026. The ability to keep raising prices without damaging seat growth proves real pricing power, but price increases usually contribute single-digit percentage points to total revenue growth. They are an enhancer, not the main force behind a doubling.

  New businesses, namely AI/Copilot, are the source of elasticity and the swing factor for whether doubling can happen. AI annual run rate has exceeded $37 billion, up 123% year over year, and paid Copilot seats have exceeded 20 million. If AI can continue expanding from $37 billion to the hundreds of billions, on top of Azure compute consumption, then a company-wide doubling becomes possible. But the honest point is that $37 billion is currently only about one-tenth of total revenue. To push the whole base from $318 billion to $636 billion, AI would need to maintain extremely high growth for years and avoid being dragged down by capacity, competition, and customers optimizing their own costs. The report lists all of these as risks that could overturn the thesis.

  Overall: at least doubling over five years is an optimistic-case target, requiring AI monetization to keep exceeding expectations while Azure does not slow. The neutral expectation is 60%-90% growth, led by volume, supported by price, with new businesses providing upside. Baillie Gifford would recognize this as a good business that keeps growing, but it would not pretend that a $3 trillion giant can easily double. What is there is there; what is not is not.

  Jun 11, 2026
• QFive years from now, what will take over as the next growth engine? Does this "second curve" exist today?5/10

  A

  Bottom line first: Microsoft's "second curve" already exists today and is converting at high speed. It is the AI agent layer: Azure AI + Microsoft 365 Copilot + Foundry/developer AI. This is the clearest current handoff engine. But to be honest, this second curve is highly co-sourced with the first curve, namely Azure cloud and the Office suite. It shares the same customers and the same channels. It looks more like an AI-enhanced extension of the first curve than an orthogonal new growth pole capable of independently recreating the company. By Baillie Gifford's yardstick, the successor is already on the field and highly visible, which is Microsoft's strength. But the second curve's independence and regenerating power are weaker than in companies that keep opening entirely new battlefields through founder-led reinvention. Microsoft's next leg is still "old platform + new capability."

  Why the second curve already exists today, and is not a slide deck: it has already appeared in the financials. Microsoft disclosed in FY2026 Q3 that AI annual revenue run rate had exceeded $37 billion, up 123% year over year, paid Microsoft 365 Copilot seats had exceeded 20 million, and Azure's overall growth had been pulled up by AI to 40% (39% in constant currency). The report therefore calls AI the reinvestment entry point for the AI era. A $37 billion annual run rate would be the scale of a large software company if it stood alone. This is a real growing curve, not an imagined one.

  But its co-sourced nature must be understood. That is exactly how it differs from a classic second curve. Microsoft's AI revenue mainly lands in two forms: AI compute and model API consumption on Azure, and Copilot seat add-ons inside the 365 suite. Both are sold to the same enterprise customers already using Azure or Office, through the same Enterprise Agreements, direct sales, and partner channels, and both rely on the switching costs repeatedly emphasized in the report: identity, file formats, workflows, and data residency. In other words, the second curve reuses all the infrastructure and customer relationships of the first curve. That makes deployment fast and visibility high, but it also means it shares the same risks as the first curve: regulatory scrutiny of bundling, customer cost optimization, and profit leakage to chip and energy lessors. It does not diversify those risks.

  How good is the handoff quality? Unit economics are a concern. The report states clearly that AI infrastructure expansion has pushed Microsoft Cloud gross margin down to 69%, while capex in the first nine months of FY2026 surged to $80.1 billion, far above $47.5 billion in the prior-year period. This second curve is powerful on revenue, but it is a capital-heavy second curve, unlike the asset-light, high-cash-conversion software curve from which Microsoft originally emerged. Baillie Gifford would ask: can this new curve eventually turn revenue into distributable shareholder cash flow as efficiently as the old curve did? The report's FCF conversion rate has fallen from about 92% in FY2021 to about 58% TTM, so the answer is still unresolved.

  Are there seeds of the next curve after that? There are signs, but with more uncertainty: gaming after the Activision integration, security driven by the Secure Future Initiative, and LinkedIn, whose FY2025 revenue had exceeded $17.8 billion, are all potential contributors. But the report also notes that gaming is hit-driven and less predictable than enterprise software. These are more like supplementary engines than main engines.

  Overall: the second curve, the AI agent layer, clearly exists today. Its growth is fierce and visibility is high, which is a clear advantage over many companies whose second curves are still only promises. The deductions are that it is co-sourced with the first curve, an extension rather than an orthogonal new continent, and it is capital-heavy with monetization quality still to be proven. The handoff is not in doubt, but it is an extension of the same baton.

  Jun 11, 2026
• QWhat is its core competitive advantage? Will this moat widen or narrow over the next three to five years?7/10

  A

  Bottom line first: Microsoft's core competitive advantage is not a single brand or low cost. It is stacked moats: switching costs, network effects, scale, channels, data, and brand lock each other together, binding enterprise customers across products and budget lines for the long term. Over the next three to five years, this moat should broadly stay stable to widen. But part of it is shifting from a pure software moat into a software + compute capex moat. The moat is deeper, but also more expensive. Greater width does not mean short-term shareholder returns improve in tandem. By Baillie Gifford's criteria, Microsoft's moat is both wide and durable. It is a genuinely rare platform-level monopolistic stickiness, and this is its strongest side. The only caution is that the cost of the moat is rising.

  Why the moat is wide: the evidence lies in share and lock-in. Switching cost is the core. The report lists identity and permissions, file formats, email systems, document collaboration, ERP/CRM processes, developer toolchains, cloud architecture, and data residency strategies as factors that jointly raise migration costs. The external result of that stickiness is high share: by Gartner's measure, Microsoft 365 has about 77% share in SaaS enterprise productivity. Scale advantage is also quantifiable: the report says Microsoft operates more than 400 data centers across 70 regions, and the three largest cloud providers together account for 58% of global hyperscale capacity. Network effects show up in LinkedIn, with 1.2 billion members, GitHub, and the Teams/Office collaboration ecosystem. Layered together, these mean a competitor trying to replicate Microsoft's depth from endpoint entry to collaboration, identity security, and cloud infrastructure would need years, tens of billions of dollars in capital, and an extremely strong channel. The report's judgment that the moat is stable to widening is defensible.

  Directional evidence that the moat will widen over the next three to five years: Azure is still taking share in a growing cloud market at a speed above the industry. In FY2026 Q3, Azure grew 40% (39% in constant currency), while the overall cloud infrastructure market grew 35% year over year in Q1 2026. Microsoft's faster-than-market growth implies share expansion. AI further reinforces lock-in: once enterprises embed Copilot (paid seats above 20 million) into daily workflows and place data in Azure, migration costs only rise. Pricing power is the moat's litmus test. Microsoft's ability to raise commercial Microsoft 365 prices starting in July 2026 while maintaining seat growth proves that stickiness is real.

  But the offsetting forces that could narrow the moat or make it more expensive must be stated plainly. This is the dynamic Baillie Gifford would press on for the next three to five years:

  First, the moat is being capitalized. Maintaining Azure's competitive position is no longer as light as writing code. The report shows Microsoft Cloud gross margin has fallen to 69%, while capex in the first nine months of FY2026 reached $80.1 billion. The moat now needs to be filled with real cash on an ongoing basis. The barrier is higher, but returns are diluted.

  Second, AI may reshape the profit pool. The report's strongest bear case is exactly this: the AI platform value chain is still forming, and models, agents, and development frameworks may alter profit allocation. Excess returns may flow more to chips, energy, and specialized compute lessors, or be absorbed by large customers optimizing costs themselves. If so, Microsoft's moat remains wide, but the profits inside the moat become thinner.

  Third, regulation targets the key joints of the moat: bundling. The report notes that the UK CMA has opened an investigation into Microsoft's position in business software, reviewing the bundling of Windows, Word, Excel, Teams, and Copilot and cloud licensing; the US FTC is also asking about its AI, identity, and cloud bundling practices. Bundling is exactly the mechanism that locks the stacked moats into one system. If regulators force unbundling, moat synergies will weaken.

  Overall: the core advantage is platform stickiness from stacked moats. Its width and durability are first-class, and over the next three to five years the direction is continued widening. But this moat is becoming increasingly expensive, and it faces erosion from regulation against bundling and from AI's potential redistribution of the profit pool. The moat itself is not the problem. The question is whether it can remain high-return while becoming deeper. That is exactly why the report holds the valuation rating at Watch.

  Jun 11, 2026
• QIf its core business is disrupted, does it have the genes for self-reinvention? How does it treat mistakes and bad news?6/10

  A

  Bottom line first: Microsoft is one of the few companies with documented evidence of genuinely showing the ability to reinvent itself when its core business is disrupted. Over the past decade, it successfully turned from a desktop software company selling Windows/Office licenses into a cloud + AI + workflow platform company. That is textbook self-reinvention. When facing mistakes and bad news, its posture also leans toward candid disclosure rather than concealment. Both are the adaptability and honest culture that Baillie Gifford values most, and Microsoft scores clearly well on this dimension. The only reservation is that the last successful pivot does not guarantee the next one, because disruption in the AI era takes a different form, and this reinvention is capital-heavy, making the cost of correction higher.

  The gene for self-reinvention is backed by real history, not PR language. The report states clearly that under Satya Nadella, Microsoft moved from a desktop software company to a cloud and platform company, then further into a cloud + AI + workflow platform company. The strategic direction is very clear. This is not a slogan. It is a pivot verified by financial results: in FY2025, revenue was $281.7 billion, up 15% year over year, while Azure revenue exceeded $75 billion and grew 34%; by FY2026 Q3, AI annual run rate had exceeded $37 billion. A company that once treated Windows as its lifeblood was able to move its strategic center to cloud, then to AI, and each move has converted into real revenue. This is evidence of an ability to build anew when the core business is challenged by the times. The "self-renewing" trait Baillie Gifford looks for has been proven by Microsoft through a complete transformation cycle that has already landed.

  How it treats mistakes and bad news: it leans candid and proactively discloses unfavorable facts. The examples in the report are persuasive, and all can be found in official filings:

  • Microsoft proactively disclosed that AI infrastructure expansion dragged down cloud gross margin (Microsoft Cloud gross margin fell to 69%), without glossing over the share-price-negative fact that heavy investment is compressing current earnings quality.
  • It proactively disclosed the profit-lifting effect of accounting changes: in FY2023, extending the useful lives of servers and network equipment increased operating income by $3.7 billion and net income by $3.0 billion; in FY2026 Q2, the fair value gain on the OpenAI investment increased quarterly net income by about $7.6 billion (that is, 7.6 billion dollars). Spelling out these non-operating disturbances that make P/E look cheaper, rather than letting investors mistake them for better earnings quality, is a sign of honesty.
  • It proactively disclosed the major contingency that the IRS is seeking about $28.9 billion in additional taxes and interest related to transfer pricing.

  The report's assessment is apt: being willing to disclose uncomfortable information does not mean there is no risk, but it is usually more trustworthy than concealment. Microsoft's 2025 annual report also disclosed that Deloitte, after audit, concluded its internal controls were effective in all material respects, further reducing the probability that bad news is being cosmetically treated through accounting.

  It treats security incidents as bad news requiring institutionalized correction. The report mentions Microsoft's concentrated investment through the Secure Future Initiative. This is itself an institutional response to the bad news of past security incidents: elevating a pain point into a company-level priority and allocating resources. The report also soberly notes that Microsoft's own decision to place security at the center shows the risk is real. In other words, the company is not denying the problem; it is acknowledging and investing in it.

  The honest reservation: the self-reinvention gene is strong, but the next disruption may not be easy to handle. This AI transition is capital-heavy. The report shows that FCF conversion has fallen from about 92% to about 58% TTM, and capex in the first nine months of FY2026 was $80.1 billion. If Microsoft bets on the wrong compute or model direction during the correction window, the sunk cost will be far higher than in the software era. In addition, management's alignment with shareholders mainly comes from compensation design rather than high equity ownership (see the management dimension), meaning the drive for self-revolution relies more on professional reputation than personal net worth.

  Overall: self-reinvention genes have been proven by a full transformation and are first-class; treatment of mistakes and bad news is candid disclosure plus institutionalized correction, indicating a healthy culture. This is one of Microsoft's most stable dimensions under the Baillie Gifford framework. The deduction is only that this reinvention is more expensive and the cost of correction is higher, not that the gene itself is defective.

  Jun 11, 2026
• QDoes management, especially the founder, have a long-term horizon and deep alignment with the company? Is it willing to sacrifice current profits for five to ten years out?5/10

  A

  Bottom line first: Microsoft's management has a clear long-term horizon and is plainly willing to sacrifice current profits for five to ten years out. The current aggressive AI capex is the most direct evidence. Its compensation design is also heavily tilted toward long-term value. But on the point Baillie Gifford values most, deep alignment of interests with the company, Microsoft is only moderately above average rather than top-tier. It has no founder, and overall executive ownership is below 1%. Alignment comes more from compensation structure and career/reputational capital than from high-equity personal wealth as in founder-controlled companies. So the answer to this question is: long-term vision and the willingness to trade today for the future are well established, but founder-style deep ownership alignment is absent.

  Long-term horizon has been verified by both strategy and financials. The report notes that under Nadella, the strategy moved from desktop software to cloud, then to cloud + AI + workflow platforms, with a very clear and continuous direction. This is not a temporary slogan: by FY2026 Q3, AI annual run rate exceeded $37 billion and Azure grew 40%, proving that the long-term bets are converting.

  Is it willing to sacrifice current profit for five to ten years out? The answer is clearly yes, and almost aggressive. This is the strongest evidence in this question: Microsoft is sacrificing current free cash flow to bet on long-term AI capacity. Capex in the first nine months of FY2026 had reached $80.1 billion, far above $47.5 billion in the prior-year period, and management publicly guided to roughly $190 billion of capex for calendar year 2026. The direct consequence is the report's recorded decline in FCF conversion from about 92% in FY2021 to about 58% TTM. Management knows this will weigh on current free cash flow and valuation sentiment, yet it continues investing. This is exactly the willingness to tolerate short-term ugliness for the long term that Baillie Gifford appreciates. Of course, the honest caveat is that this sacrifice is a bet. It is betting that AI capacity can be monetized at high returns in the future. The report also lists "AI capex staying high for a long time without incremental gross profit showing up" as a signal that would overturn the thesis. Long-term vision does not mean the bet is automatically right.

  Compensation design is strongly tilted toward long-term value, with proper incentives. The report cites Microsoft's 2025 Proxy Statement: Nadella's equity awards are entirely performance stock awards tied to long-term value creation, with no time-based equity. Together with annual cash incentives, more than 95% of CEO compensation is performance-linked. The company also requires management to hold shares equal to 5 to 15 times annual salary; Nadella's requirement is 15 times, and all named executives complied with the ownership policy in FY2025. This design ties executive pay to long-term shareholder value. The direction is right.

  But deep alignment is discounted. This is the key gap versus founder-led companies. The report gives hard data: as of September 30, 2025, Nadella beneficially owned about 900,600 shares, Amy Hood about 531,100 shares, and directors and executives together about 2,279,600 shares, all below a 1% ownership stake. Against Microsoft's roughly 7.4 billion shares outstanding, this absolute ownership percentage is very low. The report's judgment is accurate: management's alignment with shareholders comes more from compensation design and reputational/career capital than from a high equity stake as in founder-controlled companies. Baillie Gifford likes founders because their net worth is tied to the company's fate and they are willing to make contrarian long-term bets. Microsoft is governed by professional managers. The incentives are good, but the "my net worth is this company" kind of alignment does not exist. This is a structural deduction, not a personal criticism of Nadella.

  Capital allocation's self-hedging also hints at the limit of alignment. The report notes that buybacks are qualified but not outstanding: shares outstanding fell only from 7.519 billion in FY2021 to 7.434 billion in FY2025, a net reduction of about 1.1% over four years, while SBC rose from 6.118 billion to 11.974 billion. A large portion of buybacks offsets dilution from stock-based compensation. This shows that the equity component in executive compensation is large, and indirectly confirms that alignment comes through paid compensation rather than out-of-pocket ownership.

  Overall: long-term horizon is strong; willingness to sacrifice current profit for the long term is strong, with aggressive capex as hard evidence; compensation incentives point in the right direction. But founder-style deep ownership alignment is absent, with executive ownership below 1% and alignment relying on compensation and reputation rather than personal wealth. Under the Baillie Gifford framework, this is a management team with excellent governance but without a founder soul: credible, reliable, and willing to invest for the long term, but missing the layer of personal fortune placed on the table.

  Jun 11, 2026
• QIf it disappeared tomorrow, how much would customers miss it? Is its growth model sustainable, without relying on harm to society or regulation?6/10

  A

  Bottom line first: if Microsoft disappeared tomorrow, global enterprises, governments, and developers would miss it intensely. Its products are deeply embedded in identity, files, email, collaboration, ERP/CRM, developer toolchains, and cloud infrastructure. It is almost impossible to replace as a whole in the short term, so indispensability is extremely high. But on whether its growth model is sustainable and does not rely on social or regulatory harm, the answer needs a clear discount. Microsoft's strongest moat mechanisms, bundling and cloud licensing, are exactly the targets of regulatory scrutiny in the US, UK, and EU. Part of its growth is indeed built on lock-in that makes it difficult for customers to leave, and sustainability carries real regulatory uncertainty. Baillie Gifford would recognize its indispensability as top-tier, but would mark the social/regulatory sustainability of the growth model as yellow, not green.

  Start with indispensability: the degree to which customers would miss it is close to infrastructure-level. The report describes comprehensive lock-in: identity and permissions, file formats, email systems, document collaboration, ERP/CRM processes, developer toolchains, and cloud architecture. Share data confirms this irreplaceability: Microsoft 365 has about 77% share in SaaS enterprise productivity, Windows has about 63.66% desktop operating system share according to the report, and Azure is the world's second-largest cloud. More importantly, customers have mission-level dependence. In FY2026 Q3, commercial remaining performance obligation, or backlog, reached $627 billion, up 99% year over year. This is a quantified expression of how much global organizations have already committed to keep using Microsoft's compute and software in the future. If Microsoft disappeared tomorrow, corporate email, documents, identity login, ERP processes, code hosting, and cloud applications would be disrupted at the same time. Migration would take years and cost enormously. On indispensability, Microsoft is textbook full marks.

  Now look at whether the growth model is sustainable and does not rely on harm to society or regulation. This half deserves an honest deduction. A considerable part of Microsoft's growth engine comes from bundling + lock-in + continuous price increases, and that is triggering regulatory pushback:

  • The report notes that, according to the report, the UK CMA has opened an investigation into Microsoft's position in business software, reviewing bundling of Windows, Word, Excel, Teams, Copilot, and cloud licensing issues; the US FTC is also asking about its AI, identity, and cloud bundling practices; the EU remains focused on cloud practices. Regulators are targeting exactly the mechanism in Microsoft's growth model that locks multiple moats into one system through bundling.
  • Continuous price increases are a double-edged sword. The report records that Microsoft will raise prices for commercial and government Microsoft 365 editions starting in July 2026. The ability to raise prices proves stickiness, but when customers cannot easily leave and must still accept higher prices, this growth model looks more like monetizing market power in the eyes of society and regulators. Sustainability depends on regulatory tolerance.
  • The report's strongest bear case echoes this point: the market prices Microsoft as the default AI winner, but if regulators require material separation, unbundling, or cloud licensing concessions, platform synergy and pricing power would be impaired. This is precisely the risk path in which growth harms competition and is then corrected by regulation.

  Balancing the two sides: why this is still yellow rather than red. On one hand, Microsoft's growth is not purely predatory. It genuinely creates real value in productivity, compute, security, and AI capabilities, and customer dependence is largely the natural result of products that work well and are integrated, not simple coercion. The report also notes very low customer concentration, with no single customer or any single country outside the United States accounting for more than 10% of revenue, so growth does not depend on capturing a few large customers. On the other hand, regulatory scrutiny is ongoing, not a conviction. The UK CMA previously also found that the Microsoft-OpenAI partnership did not constitute a relevant merger situation investigable under merger law, showing that regulation is not one-sided. But scrutiny of bundling and cloud licensing is indeed escalating, and treating it as a material risk to growth sustainability is prudent.

  Overall: "how much would customers miss it tomorrow" equals intensely, with infrastructure-level indispensability. "Is the growth model sustainable and not harmful to society or regulation" equals partly sustainable but carrying significant regulatory uncertainty, because the strongest moat mechanisms, bundling/lock-in/price increases, are exactly the regulatory target. Baillie Gifford's judgment would be: full marks for indispensability, yellow light for social/regulatory sustainability. This business is too important, which is exactly why it cannot escape the counterforce that greater power brings tighter regulation.

  Jun 11, 2026
• QWhat are the unit economics of this business, including gross margin and incremental returns? Do they improve or deteriorate with scale? Where does the money it earns go?8/10

  A

  Bottom line first: Microsoft's unit economics are outstanding overall. Gross margin has long been around 68%-69%, operating margin has risen to about 47%, and ROIC is about 27%, which is classic high-quality software economics. But an ongoing structural change must be stated honestly: as AI data center capex explodes, its incremental returns are worsening. Free cash flow conversion has fallen from about 92% in FY2021 to about 58% TTM. In other words, overall margins are still improving as the company scales, but the incremental economics of how much distributable shareholder cash each new dollar of investment produces are deteriorating. The money it earns is mainly going into AI compute infrastructure, dividends, and buybacks, with capex already the absolute largest use. By Baillie Gifford's yardstick, the business quality itself is top-tier, but the marginal economics of scaling have hit an inflection point, shifting from asset-light high conversion toward heavy assets and discounted conversion. This is the key variable to watch now.

  Start with static unit economics: excellent and stable. The report is consistent with official data: in FY2025, revenue was $281.7 billion and operating income was $128.5 billion (operating margin about 45.6%), while gross margin has long stayed around 68%-69%. By FY2026 Q3, quarterly operating income was $38.4 billion, up 20%, and TTM operating margin had risen further to about 46.8%. The report cites a third-party measure of ROIC of about 27.24%. That means each dollar of invested capital generates about 27 cents of after-tax operating return, a very high return on capital. Compared cross-sectionally, Microsoft's earnings multiple is not the most expensive, but its unit profitability is firmly first-class.

  Do unit economics improve or deteriorate with scale? The answer differs by layer:

  First, reported margins still improve with scale. Operating margin rose from 41.6% in FY2021 to 46.8% TTM, showing that software scale effects are still being released. More seats and more cloud consumption spread fixed R&D, lifting margins. This layer is positive scale economics: the larger it gets, the more profitable it becomes.

  Second, incremental cash returns are deteriorating with scale. This is the key honest point. The table in the report gives FCF/net income conversion: about 91.6% in FY2021, about 84.0% in FY2024, about 70.3% in FY2025, and about 58.2% TTM. Official data confirms it: for FY2025, free cash flow was about $7.16 billion (should be $71.6 billion), while operating cash flow was about $136.2 billion, and the gap was widened by capex. The reason is not deterioration in the core business, but the fact that incremental AI infrastructure investment is consuming incremental cash. The report calls this "real profit being absorbed by reinvestment." In other words, each additional block of AI/cloud business now requires heavy asset investment upfront, and the immediate cash content of incremental returns is weaker than in past pure software expansion. This is why the report prefers a range estimate for Owner Earnings, conservatively about $110 billion-$112 billion, rather than a pseudo-precise number.

  The gross margin structure is also softening. The report states clearly that in FY2025, Microsoft Cloud gross margin had fallen to 69%, and Intelligent Cloud cost grew 36% year over year, mainly due to AI infrastructure expansion. The fastest-growing AI line is precisely the line with relatively lower gross margin and the most capital intensity. The gross margin quality of incremental revenue is below traditional Office/Windows licensing.

  Where does the money go? Capex is already the absolute largest use. There are three destinations: 1. AI/cloud capex, which had already reached $80.1 billion in the first nine months of FY2026, far above $47.5 billion in the prior-year period, making it the largest use; 2. dividends, with FY2025 dividends of $24.677 billion; 3. buybacks, though the report says buybacks are qualified but not outstanding. Net share count fell only about 1.1% over four years, and a considerable portion offset dilution from SBC, which rose from 6.118 billion to 11.974 billion. In other words, more and more cash is flowing into heavy assets that buy future capacity, while the relative share available for direct shareholder returns is being compressed.

  Overall: unit economics quality is top-tier, with high gross margin, high operating margin, and high ROIC. Incremental economics at scale have reached an inflection point: FCF conversion has fallen from about 92% to about 58%, and gross margin structure is being pulled down by capital-heavy AI, so the immediate cash return on each new block of business is worsening. The money earned mainly goes into AI infrastructure. Baillie Gifford's conclusion would be: this remains an excellent business, but it is transitioning from an asset-light cash machine into a capital-heavy business whose returns need time to be realized. Whether incremental returns recover as AI monetization scales is the swing factor for the future quality of this business.

  Jun 11, 2026
• QWhat conditions must all hold for it to rise fivefold over ten years? Are those conditions realistic? What expectations are embedded in today's share price?3/10

  A

  Bottom line first: for Microsoft to rise fivefold over ten years, a fairly demanding set of conditions must all hold at once. Annualized share price return would need to be about 17.5%, and market cap would need to grow from about $3 trillion today to about $15 trillion. That requires revenue/Owner Earnings to roughly triple or quadruple over ten years, AI capex to convert efficiently into high-return cash flow, and valuation multiples not to contract. Each condition on its own is not absurd, but having all of them hold together is unrealistic. In substance, it requires Microsoft to keep compounding at medium-high speed from a $3 trillion base while fully releasing the free cash flow currently being held down. Today's roughly $402 share price implies exactly that the market has already priced Microsoft as the default big winner of the AI era, leaving most of the room for a fivefold return already consumed. In Baillie Gifford terms: a fivefold return over ten years is the ceiling of an optimistic scenario, not the neutral expectation. The optimistic expectations embedded in today's price are the source of insufficient margin of safety.

  First translate "fivefold" into hard conditions. Microsoft is currently around $402 per share, with market cap of about $2.96-$3.0 trillion. A fivefold rise over ten years means market cap reaching about $15 trillion and annualized return of about 17.5%. To support that result, the following conditions must broadly hold at the same time:

  1. Fundamentals must triple to quadruple. If valuation multiples are roughly unchanged ten years from now, a fivefold market cap requires Owner Earnings/revenue to rise roughly fivefold as well. Even allowing some room for multiple expansion, fundamentals still need to triple or quadruple. But the report and official data show Microsoft FY2025 revenue was already $281.7 billion, with TTM around $318 billion. Tripling or quadrupling over ten years from this base means long-term compound growth of 12%-15%. Possible, but difficult.

  2. AI capex must monetize efficiently into shareholder cash. This is the tightest link. The report shows FCF conversion has fallen from about 92% in FY2021 to about 58% TTM, capex in the first nine months of FY2026 was $80.1 billion, and CY2026 guidance is about $190 billion. For fivefold to work, the AI business with a $37 billion annual run rate (up 123% year over year) must keep expanding exponentially, while gross margin recovers and free cash flow conversion climbs back from 58% to high levels, rather than being permanently consumed by ongoing capex.

  3. Azure must keep at least matching cloud industry growth, and regulation must not force unbundling. The report lists both "Azure growth consistently falling behind the industry" and "regulation forcing unbundling or major concessions on cloud licensing" as signals that would overturn the thesis. For a fivefold return, these downside scenarios cannot happen.

  4. Valuation multiples must not contract. The report gives current P/FCF of about 43 times and P/E of about 25 times. If multiples fall over ten years to the low-to-mid 20 times range, or lower, for mature platform companies, as the report's repricing scenario imagines, even fundamental growth would not be enough for a fivefold return.

  Are these conditions realistic? One by one, they are credible; in combination, they are not. Each condition has a real basis when viewed alone: Microsoft has AI tailwinds, Azure is indeed growing at 40%, and the moat is genuinely wide. But the Baillie Gifford framework asks whether they can all hold simultaneously: fundamentals triple or quadruple + capex monetizes efficiently + Azure does not slow + regulation does not unbundle + multiples do not compress. The joint probability that all five go right is not high. The report's own optimistic scenario gives only 11%-14% annualized return, or about 3-3.7 times over ten years, which still falls short of fivefold. That means the report is internally saying even its most optimistic valuation scenario struggles to support fivefold. The honest conclusion: a fivefold return over ten years is an upside tail, not the center.

  What expectations are embedded in today's share price? It already discounts the "default AI winner" story. This is the key question. The report's judgment is direct: the current price is closer to a fair holding price than a purchase price with margin of safety, and the strongest bear argument is that the market is pricing Microsoft as the default winner of nearly all AI benefits before cash flow has proven it. Quantitatively:

  • Using the report's roughly $110 billion Owner Earnings estimate, the market is assigning about 28 times conservative Owner Earnings; using strict free cash flow, it is about 43 times P/FCF. The latter is far above mature software companies, showing that the price already embeds optimistic assumptions about fast AI monetization.
  • The report's fair intrinsic value range is $390-$470, and the current roughly $402 price sits in the lower part of that range. That means market pricing broadly recognizes fair value, with almost no discount reserved for a future fivefold return. Put differently, if you buy today, the price you pay already embeds the expectation that Microsoft will win AI and monetize the investment efficiently. If it only delivers the neutral scenario, the report estimates future ten-year annualized return at only 7%-9%.
  • The report therefore sets the required margin of safety at a 20%-30% discount to fair value, with an ideal buying range of $300-$360. The current price is outside that range, showing that the optimistic expectations embedded in the price have removed the margin of safety.

  Overall: a fivefold return over ten years requires five conditions to hold at once: fundamentals triple or quadruple, AI capex monetizes efficiently, Azure does not slow, regulation does not unbundle, and multiples do not compress. Individually credible, jointly unrealistic. Even the report's most optimistic scenario reaches only about 3-3.7 times. Today's roughly $402 share price implies the optimistic expectation of Microsoft as the default AI winner with efficient monetization, and is already priced near fair value, leaving almost no discount for fivefold. Baillie Gifford-style judgment: good company, but the current price already reflects most good news in advance. Fivefold is a luxury hope, not a reasonable expectation.

  Jun 11, 2026
• QWhy has the market not realized all this yet? Is it failing to understand it, dismissing it, or not looking far enough? What will become the narrative inflection point?3/10

  A

  Bottom line first: Microsoft has almost none of the traditional perception gap where the market fails to understand, dismisses, or cannot look far enough. Quite the opposite: it is one of the most thoroughly researched, institutionally covered, and fully priced companies in the world. The market not only understands it; it may well view it too favorably and has already discounted AI benefits into the share price in advance. So this question has to be answered in reverse for Microsoft: the real disagreement is not that the market underestimates it, but whether the market's optimistic assumptions about AI monetization will fail. The possible narrative inflection point is not a sudden market recognition of some positive fact, but the moment when the gap between AI capex and cash flow returns is either disproven or confirmed. By Baillie Gifford's yardstick, it seeks great growth stocks the market has not yet recognized. Microsoft's problem is the opposite: it has been fully recognized, the perception gap is tiny, and that is the source of the lack of valuation margin of safety.

  Why there is no underappreciated perception gap: the evidence lies in valuation and coverage. The report's core judgment is that the current price looks more like a fair-to-slightly-expensive price for a high-quality asset than an obvious bargain with margin of safety. Quantitatively, Microsoft is currently around $402 with market cap near $3 trillion, corresponding to P/FCF of about 43 times and P/E of about 25 times. This is a level that prices optimistic expectations quite fully. It is not a neglected, dust-covered stock. The report's fair value range is $390-$470, and the current price sits inside the range, meaning market pricing roughly matches intrinsic value, with no meaningful "doesn't understand" discount. Microsoft is heavily covered by global sell-side firms, buy-side investors, and index funds. Information is almost fully efficient. The three classic perception gaps of not understanding, dismissing, or not looking far enough are all hard to establish here.

  If one insists on finding subtle things the market may not fully see, the direction is two-sided, and not necessarily bullish:

  • The bearish side, where the market may be underestimating risk: the report repeatedly stresses that reported net income has been beautified by one-off items. In FY2026 Q2, the fair value gain on the OpenAI investment increased quarterly net income by about $7.6 billion (that is, 7.6 billion dollars), and the FY2023 extension of equipment useful lives lifted profit. Investors who look only at static P/E can easily overestimate true earnings quality. In other words, the market may not fully appreciate the water content in current earnings and capex's consumption of cash flow, with FCF conversion having fallen to about 58% TTM. This is the market possibly missing risk, not missing value.
  • The bullish side, weaker: the report acknowledges that if AI monetization is faster than expected and Azure continues growing at 40%, the optimistic scenario still has room. But even then, the report's optimistic annualized return is only 11%-14%. The market broadly knows this upside. It is not an unrecognized perception gap.

  What will become the narrative inflection point? This is Baillie Gifford's key follow-up question. Since there is no positive inflection point of the market discovering an undervaluation, Microsoft's narrative inflection is more likely to be the critical event where optimistic assumptions are confirmed or falsified. The report has essentially listed them all:

  1. Convergence or deterioration in the gap between AI capex and cash flow. This is the No. 1 inflection point. If capex remains high in the coming years, with CY2026 guidance around $190 billion, while free cash flow expands alongside AI monetization and FCF conversion recovers from 58%, the narrative will shift positively to "heavy investment is beginning to harvest." Conversely, if capex stays elevated but incremental gross profit does not appear, the narrative will reverse to "good company, bad price." The report explicitly lists "AI capex continuing to rise sharply without incremental gross profit showing up" as a signal for reassessment.

  2. Whether Azure growth keeps its industry lead. The report lists "Azure growth staying materially below industry growth" as a fact that would overturn the thesis. If it happens, the foundation of the "default AI winner" narrative would shake. This is the most dangerous negative inflection point.

  3. Regulatory decisions. The report notes that the UK CMA has opened an investigation into bundling and cloud licensing across Windows, Word, Excel, Teams, and Copilot, and that the US FTC is also asking questions. If regulators force unbundling or major cloud licensing concessions, the narrative of platform synergy + pricing power will be directly rewritten.

  4. The sustainability of OpenAI ecosystem and Copilot monetization. Copilot paid seats have exceeded 20 million. If ARPU, retention, and incremental gross profit prove durable, the narrative turns more bullish. If revenue growth does not translate into cash improvement, the narrative turns bearish.

  Overall: Microsoft does not have a traditional positive perception gap where the market does not understand, dismisses, or cannot look far enough. It has been researched too thoroughly and priced too fully, and the market may even be too optimistic. The Baillie Gifford framework should be answered honestly in reverse here: this is not a dusty pearl waiting to be discovered, but a bright gem already held high. The very small perception gap is exactly why it lacks margin of safety. The true narrative inflection point is not the discovery of good news, but whether AI capex can monetize efficiently, Azure can maintain its lead, and regulation lands. Those moments of confirming or falsifying optimistic assumptions are the real dividing line for Microsoft's future narrative.

  Jun 11, 2026