Microsoft will report its fiscal fourth-quarter 2026 results after the market closes on July 29, and I believe this may be one of the company’s most intellectually demanding earnings reports in years.
I am not expecting Microsoft to reveal whether artificial intelligence is real, whether enterprises want it or whether the company possesses the technical resources to compete. Those questions have already been answered. Demand is real. Adoption is expanding. Microsoft has assembled one of the most comprehensive AI distribution systems in the world, spanning Azure, GitHub, Microsoft 365, Dynamics, security and its broader developer ecosystem.
The harder question is whether Microsoft can convert extraordinary AI demand into returns that justify an equally extraordinary level of investment.
This distinction matters. The market is no longer evaluating a promising experiment financed from the margins of an enormously profitable software business. Microsoft is redesigning its capital structure, infrastructure footprint, product architecture and commercial model around AI. The company expects approximately $190 billion in capital expenditures during calendar 2026, including about $25 billion associated with higher component prices. Fiscal fourth-quarter capital spending alone is expected to exceed $40 billion. Management also believes capacity will remain constrained through at least the end of 2026. Microsoft FY2026 Q3 earnings call
Those numbers change the analytical framework.
For most of its modern history, Microsoft has been admired for the economic elegance of software: high incremental margins, recurring revenue, modest physical capital requirements and enormous operating leverage. AI introduces a more industrial form of technology economics. Before Microsoft can sell intelligence at scale, it must finance data centers, power infrastructure, networking equipment, CPUs, GPUs, storage systems and increasingly expensive short-lived computing assets.
The software empire is learning to pour concrete.
That does not destroy the investment thesis. It does mean I can no longer judge Microsoft primarily by whether quarterly earnings exceed Wall Street’s estimate by a few cents. I need to understand the relationship between capital deployed, capacity activated, customer usage, revenue recognized and cash ultimately returned to shareholders.
The headline results will matter. The underlying conversion mechanism will matter much more.
The Expectations Entering the Quarter
Microsoft has officially scheduled its fiscal fourth-quarter report for July 29, 2026, with its conference call beginning at 2:30 p.m. Pacific time. Microsoft Investor Relations
Management has guided for total quarterly revenue between $86.7 billion and $87.8 billion, implying year-over-year growth of approximately 13% to 15%. The midpoint is $87.25 billion. Intelligent Cloud revenue is expected to fall between $37.95 billion and $38.25 billion, while Productivity and Business Processes should generate between $37 billion and $37.3 billion. More Personal Computing is expected to produce between $11.75 billion and $12.25 billion.
Microsoft expects Azure revenue to grow between 39% and 40% in constant currency. Microsoft 365 Commercial cloud revenue is expected to grow between 13% and 14% on a reported constant-currency basis, or 15% to 16% after normalizing for a prior-year revenue-recognition benefit. Microsoft Cloud gross margin is expected to decline to approximately 64%, reflecting continued AI investment and increased GitHub Copilot usage. Microsoft FY2026 Q3 earnings call
The broad Wall Street earnings estimate is approximately $4.24 per share, although consensus figures can shift as analysts update models before the report. Kiplinger earnings calendar
At the July 24 close, Microsoft traded at approximately $381.70, with a market capitalization near $2.84 trillion and a trailing price-to-earnings ratio of about 22.7. Those figures make the setup more nuanced than the familiar story of an obviously excellent company priced for perfection.
Microsoft remains an elite business, but the stock has not been behaving like an unquestioned market sovereign. Investors are trying to reconcile exceptional operating growth with accelerating capital requirements, changing competitive relationships and uncertainty over how much economic value AI suppliers will ultimately retain.
That is the debate I expect this report to sharpen.
Azure Is Still the Most Important Number
If I could examine only one operating metric, I would choose Azure growth.
Microsoft reported 40% Azure and other cloud services growth in the fiscal third quarter, or 39% in constant currency. That result exceeded management’s expectations because additional capacity became available earlier than anticipated, permitting Microsoft to satisfy more consumption demand. Management emphasized that customer demand across workloads, regions and customer segments continued to exceed available capacity. Microsoft FY2026 Q3 results
That last point makes Azure growth unusually difficult to interpret.
Ordinarily, when a cloud business grows faster than expected, I infer that customer demand has strengthened. In Microsoft’s present situation, reported growth is the product of demand and infrastructure availability. If demand already exceeds supply, Azure’s quarterly growth rate becomes partly a measurement of how rapidly Microsoft can install equipment, connect power, complete data centers and make computing capacity revenue-ready.
Microsoft is therefore operating under a strange condition: its growth may appear slower than its underlying demand because the company cannot produce enough digital capacity to fulfill all available consumption.
This sounds bullish, and in one sense it is. Businesses generally prefer excess demand to idle capacity. Yet persistent shortages create their own risks. Customers unable to obtain sufficient Microsoft capacity may shift workloads elsewhere, delay deployments, build more internally or distribute demand across competing platforms. Scarcity validates the product while simultaneously creating an opportunity for competitors.
For the fourth quarter, I will compare reported Azure growth with the 39% to 40% constant-currency guidance range. A result above 40% would be encouraging, especially if management attributes the upside to both capacity additions and broad demand. A result around 39% would be acceptable if forward guidance supports the promised acceleration in the second half of calendar 2026.
A material miss would require careful interpretation.
If growth falls short because a data center opened later than expected, I would treat the problem as operational and potentially temporary. If Microsoft describes weaker consumption, customer optimization, pricing pressure or slower AI deployment, I would view the issue as more fundamental.
The language surrounding the number will be nearly as important as the number itself.
I will listen for whether management continues to say that demand “exceeds available capacity.” I will also look for any changes in the duration, geography or workload composition of those constraints. If the bottleneck is narrowing, Azure should eventually recognize revenue that current infrastructure cannot serve. If management becomes less emphatic about demand, the market may conclude that supply and demand are moving toward balance for less flattering reasons.
Capacity Is Becoming a Revenue Variable
Microsoft has turned infrastructure execution into part of its quarterly earnings algorithm.
During the third quarter, the company added another gigawatt of capacity. It said its Wisconsin Fairwater data center came online six weeks early, allowing revenue to be recognized sooner. Microsoft also reported that it had reduced the time required to move newly delivered GPUs into production by nearly 20% since the beginning of the year. Inference throughput across its most-used Copilot models improved 40% through hardware and software optimization. Microsoft FY2026 Q3 earnings call
These may sound like technical footnotes. I see them as direct economic indicators.
If Microsoft can reduce the time between receiving a GPU and generating revenue from it, capital turns faster. If inference throughput improves, the company can serve more demand from the same installed equipment. If custom accelerators lower the cost per token, gross margins can improve even while usage expands.
A useful way to think about AI infrastructure is that every unit of computing capacity has three relevant dimensions: acquisition cost, productive utilization and economic yield. Investors naturally focus on the first because Microsoft’s spending is visible and enormous. The long-term investment outcome will depend on the second and third.
A GPU sitting idle is an expensive object.
A GPU operating at high utilization while supporting profitable Azure consumption or a valuable first-party application is productive capital.
A GPU serving heavily subsidized usage with limited pricing power may generate revenue without generating an adequate return.
Microsoft does not disclose enough information for outsiders to calculate a clean return on invested capital for its AI infrastructure. I therefore have to assemble indirect evidence: Azure acceleration, Copilot usage, cloud gross margins, free cash flow, customer commitments, capacity constraints, inference efficiency and management’s discussion of pricing.
The next report should tell me whether these variables are beginning to move in the right sequence.
The $190 Billion Question
Microsoft expects approximately $190 billion of capital expenditures in calendar 2026. For perspective, that is larger than the annual economic output of many countries and exceeds the market capitalization of most publicly traded companies.
I do not object to large spending simply because it is large. If Microsoft sees demand that can generate attractive risk-adjusted returns, refusing to invest would be the greater strategic mistake. A dominant platform can lose its position by protecting near-term margins while a new computing paradigm forms around it.
The question is whether capital expenditures are scaling faster than the evidence supporting their returns.
In the fiscal third quarter, Microsoft reported $31.9 billion of capital expenditures. Roughly two-thirds went toward short-lived assets, primarily GPUs and CPUs. The remaining third supported longer-lived assets expected to generate value over 15 years or more. Cash paid for property and equipment was $30.9 billion. Free cash flow fell to $15.8 billion despite operating cash flow of $46.7 billion, largely because infrastructure spending absorbed so much cash. Microsoft FY2026 Q3 earnings call
This composition matters.
Data center land, buildings and power infrastructure can support multiple generations of computing equipment. Their useful economic lives may extend across a decade or more. GPUs and other accelerators become obsolete much faster. Microsoft may depreciate them over an accounting schedule, but technological obsolescence does not always wait politely for the accounting department.
The investment case therefore depends partly on Microsoft earning sufficient returns before current equipment loses its competitive usefulness.
I will watch three aspects of capital spending.
First, does the fiscal fourth-quarter number remain near the “over $40 billion” expectation, or does it move materially higher? An increase caused by timing or finance leases is different from an increase caused by another upward revision to structural spending requirements.
Second, does management maintain the $190 billion calendar-year outlook? A higher forecast could be interpreted positively if demand and Azure growth accelerate proportionally. Higher spending without improved revenue expectations would place more pressure on free cash flow and future depreciation expense.
Third, I want greater clarity on the division between short-lived equipment and long-lived infrastructure. The larger the share directed toward short-lived assets, the faster Microsoft must monetize that capacity to preserve attractive returns.
Capital spending is not an expense on the current income statement, which makes earnings appear more insulated than cash flow. Depreciation arrives gradually. Cash leaves much sooner.
That timing difference is becoming central to Microsoft’s valuation.
Free Cash Flow Deserves More Attention Than EPS
Quarterly earnings per share remains important, but I believe free cash flow is becoming the more revealing measure of Microsoft’s economic transition.
A software company can report strong operating income while capital intensity quietly rewrites the quality of those earnings. If the business requires increasingly large investments merely to sustain competitive growth, less cash remains available for repurchases, dividends, acquisitions or balance-sheet accumulation.
This does not mean earnings are false. It means the cash cost of producing future earnings has changed.
Microsoft’s fiscal third-quarter operating cash flow increased 26% to $46.7 billion, an excellent result. Free cash flow was only $15.8 billion after capital investment. The company still returned $10.2 billion to shareholders through dividends and repurchases, but the distance between accounting profit and immediately distributable cash has widened.
I will examine whether fiscal fourth-quarter operating cash flow can absorb the expected increase in capital spending. Seasonal collection patterns may create volatility, so I would not judge the thesis from one quarter alone. I will instead look at the trailing twelve-month relationship among operating cash flow, capital expenditures and free cash flow.
If Azure growth accelerates while free-cash-flow margins stabilize, Microsoft will demonstrate that its infrastructure expansion is beginning to generate operating leverage.
If revenue remains strong while free cash flow continues to compress, the company may still be building a valuable platform, but shareholders will need to wait longer for the economic harvest.
Patience can be rational.
It should never be automatic.
Copilot Must Become More Than an Adoption Story
The Copilot narrative is moving from product launch to economic proof.
Microsoft reported more than 20 million paid Microsoft 365 Copilot seats in the fiscal third quarter. Copilot queries per user increased almost 20% sequentially, while monthly active usage of Microsoft’s first-party agents rose sixfold from the beginning of the fiscal year. Management said weekly Copilot engagement had reached a level comparable with Outlook. GitHub Copilot was used by nearly 140,000 organizations, and enterprise subscribers had nearly tripled from the prior year. Microsoft FY2026 Q3 earnings call
Those are impressive indicators, but adoption metrics can create a comforting illusion of precision.
I want to know whether customers are renewing, expanding and using Copilot intensively enough to support durable pricing. A paid seat can represent deep workflow integration, or it can represent a corporate experiment purchased by an innovation budget and forgotten by half the employees assigned to it.
Usage intensity helps separate those outcomes.
Microsoft’s most important strategic insight may be that AI monetization will not remain purely seat-based. The company increasingly describes a hybrid model combining licenses with consumption. A business may pay for employee access and then purchase additional usage as agents perform more work. GitHub Copilot has already moved further toward usage-based pricing, and Dynamics customers are buying consumption credits for automated service functions.
This model could materially expand Microsoft’s addressable market. Traditional software pricing is limited by the number of human users. Agentic pricing can be tied to the volume of work performed, tasks completed or economic value created.
It could also make revenue less predictable.
Seat-based subscriptions provide recurring revenue with relatively stable gross margins. Consumption revenue fluctuates with usage and carries direct computing costs. Microsoft must price each workload so that the value delivered to customers exceeds the price while the price exceeds the cost of inference, infrastructure and product development by a satisfactory margin.
That sounds obvious. In AI economics, it is the entire game.
I will listen for evidence that paid Copilot seats continued to accelerate sequentially, as management projected. I also want disclosure around customer expansion, usage depth, renewal behavior and contribution to average revenue per user.
A rising seat count proves distribution.
Sustained high-value usage proves a business.
The AI Revenue Run Rate Needs Context
Microsoft said its AI business surpassed a $37 billion annual revenue run rate in the third quarter, growing 123% from the previous year. That is already larger than many Fortune 500 companies. Microsoft FY2026 Q3 results
The figure confirms that Microsoft has built a substantial AI operation rather than attaching fashionable language to traditional cloud revenue. Yet annualized run rates can be misunderstood. They extrapolate current-period performance and do not necessarily equal revenue already recognized across an entire year.
I will look for an updated number and, more importantly, the composition behind it.
How much comes from Azure AI services?
How much comes from Microsoft 365 Copilot, GitHub Copilot, security products and Dynamics agents?
How much reflects inference usage with high incremental costs?
How much is connected to large strategic counterparties rather than diversified enterprise customers?
Microsoft may not provide a complete breakdown, but even incremental disclosure would help investors evaluate the durability and margin profile of the AI business.
A $50 billion AI revenue run rate supported by broad, expanding enterprise usage would be profoundly different from the same figure concentrated among a small number of capital-intensive relationships.
Revenue scale matters.
Revenue quality determines value.
The Backlog Is Enormous—and More Complicated Than It Appears
Microsoft’s commercial remaining performance obligation reached $627 billion in the third quarter, up 99% year over year. That figure is astonishing, but it requires adjustment. The increase was heavily influenced by OpenAI commitments. Excluding OpenAI, commercial RPO grew 26%, which remains excellent. Approximately 25% of total RPO is expected to convert into revenue over the following 12 months, while the weighted average duration is approximately two and a half years. Microsoft FY2026 Q3 earnings call
I do not treat RPO as equivalent to guaranteed revenue.
Backlog duration matters. Customer concentration matters. Contract structure matters. Capacity availability matters. The timing of recognition matters. A multiyear commitment can provide excellent visibility while still requiring Microsoft to spend heavily before delivering the contracted services.
The OpenAI component deserves particular attention because Microsoft’s relationship with OpenAI combines strategic partnership, investment exposure, intellectual-property access, cloud demand and commercial commitments. That makes the economic relationship difficult to compare with an ordinary customer contract.
I will separate reported bookings and RPO growth from the figures excluding OpenAI. The adjusted numbers offer a cleaner view of demand across Microsoft’s wider enterprise base.
If commercial bookings accelerate broadly while RPO conversion remains healthy, I will view that as evidence that AI demand extends beyond a few headline partnerships.
If the reported backlog expands while underlying annuity bookings soften, the surface number may conceal a less balanced demand picture.
Gross Margin Is the Bridge Between Excitement and Economics
Microsoft’s fiscal third-quarter company gross margin was 68%, down from the prior year because of AI infrastructure investment and expanding AI product usage. Microsoft Cloud gross margin was 66%. Management expects the latter to fall to approximately 64% in the fourth quarter. Microsoft FY2026 Q3 earnings call
That decline is not automatically alarming.
New infrastructure platforms often experience margin pressure during rapid expansion. Capacity is installed before it reaches optimal utilization. Software must be optimized. Custom silicon takes time to deploy. Pricing models evolve. Early product usage may be subsidized to stimulate adoption.
The investment thesis assumes that efficiency eventually catches up.
Microsoft has several levers. It can increase utilization, improve inference throughput, route workloads to lower-cost models, deploy custom accelerators, renegotiate component economics, adjust pricing and shift customers toward consumption structures that better reflect computational cost.
The company’s Maia 200 accelerator reportedly offers more than a 30% improvement in tokens per dollar relative to the latest silicon in its fleet. Microsoft also achieved the previously mentioned 40% improvement in inference throughput for heavily used Copilot models. Those gains suggest the company is attacking AI costs across the entire stack rather than accepting vendor economics as fixed.
I will watch whether cloud gross margin lands near 64% and how management frames the next fiscal year. Stabilization would suggest that utilization and efficiency gains are beginning to offset infrastructure costs. Continued deterioration could still be acceptable if Azure and Copilot growth accelerate substantially, but the burden of proof would rise.
An AI business can grow quickly while destroying value if every additional dollar of revenue requires too much incremental capital and computation.
Gross margin is where the romance of artificial intelligence meets the arithmetic of running it.
Microsoft 365 Is the Quietly Crucial Business
Azure receives most of the attention, but Microsoft 365 may ultimately determine whether Microsoft captures disproportionate AI value.
Azure allows Microsoft to sell the infrastructure underlying the AI economy. Microsoft 365 allows it to embed AI directly into the daily work of hundreds of millions of people. The second opportunity may offer greater differentiation because Microsoft controls the application, identity, organizational data, workflow and distribution relationship.
Fiscal third-quarter Microsoft 365 Commercial cloud revenue grew 19%, or 15% in constant currency. Paid commercial seats grew 6%, while average revenue per user benefited from E5 subscriptions and Copilot. Management expects normalized constant-currency growth of 15% to 16% in the fourth quarter.
I will focus on the source of growth.
If seat growth remains around the mid-single digits while revenue expands in the mid-teens, pricing and product mix are doing substantial work. That can be attractive if customers willingly pay more because the product creates more value. It becomes less attractive if growth depends primarily on price increases without corresponding engagement or retention.
Copilot offers Microsoft a route to reaccelerate average revenue per user across an enormous installed base. The opportunity is mathematically powerful. Even modest incremental revenue per seat can become significant when multiplied across large enterprise populations.
However, the company must prove that Copilot is becoming embedded in workflows rather than remaining an optional executive curiosity.
I want to see adoption move beyond technology companies and early enthusiasts. I want evidence from regulated industries, smaller businesses, frontline workers and operational departments. The true prize is not convincing chief information officers to run pilots. It is becoming sufficiently useful that removing Copilot would create organizational resistance.
Software becomes durable when customers complain about the price but refuse to operate without it.
Do Not Ignore the Consumer Weakness
More Personal Computing will probably remain the least celebrated segment in the report.
Management expects revenue between $11.75 billion and $12.25 billion. Windows OEM and Devices revenue is expected to decline in the mid-to-high teens. Xbox content and services revenue is expected to decline in the low teens, while hardware revenue should also fall. Search advertising excluding traffic-acquisition costs is expected to grow in the high single digits.
These businesses are no longer the primary drivers of Microsoft’s valuation, but I would not dismiss them.
The expected Windows weakness reflects difficult comparisons, elevated channel inventory, higher memory prices and a weaker PC market. Some of those factors may be temporary. Gaming is more strategically uncomfortable. Microsoft has spent heavily to assemble a vast gaming ecosystem, yet Xbox performance remains uneven, price changes have frustrated users and the company is still working to reconcile hardware, subscriptions, cloud delivery and cross-platform distribution.
I do not expect one quarter to resolve the gaming thesis.
I will look for evidence that management has a coherent economic objective beyond reporting engagement records. Monthly active users and streaming hours are useful, but investors ultimately need revenue, margin and durable customer value.
Consumer weakness is unlikely to overwhelm Azure and Microsoft 365 strength. It can still reduce the quality of total-company growth and expose areas where Microsoft’s execution is less formidable than its cloud reputation suggests.
Guidance Will Move the Stock
The fiscal fourth-quarter results will describe the past. Microsoft’s guidance for fiscal 2027 will determine how investors value the future.
I will be listening for five commitments.
First, does management still expect Azure growth to accelerate modestly during the second half of calendar 2026?
Second, will capacity constraints persist through the calendar year, and are they expected to ease during fiscal 2027?
Third, what level of capital expenditure should investors expect after calendar 2026? The market needs to understand whether $190 billion represents a peak, a plateau or another step in a rising sequence.
Fourth, how will depreciation and finance-lease interest affect margins and earnings as recent investments move through the financial statements?
Fifth, can Microsoft maintain operating leverage while absorbing AI infrastructure costs?
Management previously indicated that full-year fiscal 2026 operating margin should increase approximately one percentage point despite AI investment and roughly $900 million of fourth-quarter costs associated with a voluntary retirement program. That is remarkable cost discipline. The question is whether the same discipline can continue as depreciation increases and the company maintains mid-to-high-single-digit operating expense growth.
Strong guidance would combine Azure acceleration, broad Copilot adoption, manageable margin pressure and confidence in capital returns.
Weak guidance would contain some combination of slower cloud growth, persistent infrastructure inflation, deeper margin compression and ambiguous AI monetization.
What Would Make Me More Bullish
I would become more constructive if Microsoft delivers Azure growth above 40% in constant currency, maintains broad-based capacity-constrained demand and guides toward further acceleration.
I would also want paid Microsoft 365 Copilot seats to rise meaningfully above 20 million, accompanied by stronger usage and clear average-revenue-per-user benefits. A substantial increase in the AI revenue run rate would help, particularly if management offers more insight into its diversification and margins.
Cloud gross margin near or above the 64% outlook would indicate that economics are tracking the investment plan. Strong operating cash generation would help offset the enormous capital-spending burden.
The most bullish outcome would be evidence that Microsoft’s AI flywheel is working across layers: infrastructure capacity enables Azure consumption, Azure supports application development, applications drive Copilot usage, usage produces data and workflow integration, and that integration increases customer dependence on Microsoft’s platform.
That would make the capital spending look less like an arms race and more like the construction of a self-reinforcing economic system.
What Would Concern Me
I would become more cautious if Azure misses guidance and management’s explanation shifts from insufficient supply toward softer demand.
A major increase in capital-spending expectations without corresponding revenue acceleration would also concern me. The market can tolerate extraordinary investment when demand is visibly outrunning capacity. It will become less forgiving if spending continues to rise while growth plateaus.
I would watch closely for further cloud gross-margin deterioration. Some pressure is expected, but a persistent decline without a credible path toward stabilization would suggest that AI economics are less attractive than the revenue narrative implies.
Weak Copilot seat additions, declining engagement or evasive discussion of renewals would raise questions about enterprise willingness to move from experimentation to durable deployment.
I would also be cautious if commercial bookings excluding OpenAI weaken materially. Microsoft’s AI story is strongest when it rests on diversified enterprise demand rather than a few massive strategic relationships.
Finally, I would pay attention to management’s tone. Executives often reveal changing conditions through emphasis before the numbers make the change obvious. Repeated references to “long-term opportunity” can be perfectly reasonable, but they sometimes appear more frequently when near-term economics have become harder to defend.
My Investment View Before the Report
At approximately $381.70 and roughly 22.7 times trailing earnings as of July 24, Microsoft does not appear extravagantly valued relative to its business quality and growth. The stock’s multiple reflects meaningful skepticism compared with the premium often granted to a company of Microsoft’s durability.
I understand the skepticism.
Microsoft is attempting one of the largest capital reallocations in corporate history. It is moving from a predominantly asset-light model toward a hybrid business that combines exceptional software economics with the physical demands of global AI infrastructure. The company may create extraordinary value from this transition. It may also discover that portions of the AI stack commoditize faster than expected, allowing customers and chip suppliers to capture more of the economics.
My current view is cautiously bullish.
Microsoft possesses advantages that are difficult to replicate: enterprise distribution, cloud scale, developer relationships, identity infrastructure, proprietary applications, enormous cash generation and the ability to spread AI costs across multiple products. It can monetize the same underlying infrastructure through Azure services, GitHub, Microsoft 365, security, Dynamics and consumer applications. Few competitors have that breadth.
Yet I would not buy the stock merely because Microsoft is “an AI leader.” Leadership is strategically meaningful but financially incomplete. I need evidence that each additional wave of investment generates durable incremental cash flows at attractive returns.
That is what I will be looking for on July 29.
I will read past the earnings beat or miss. I will ignore the first emotional after-hours move. I will resist the ritual in which investors celebrate a three-cent EPS surprise while overlooking a multibillion-dollar change in capital intensity.
The central question is larger:
Is Microsoft building AI infrastructure faster because profitable demand is compounding, or is it spending faster because competition makes restraint impossible?
The answer may not be delivered in a single quarter. This earnings report should move us closer.
Microsoft does not need to prove that it can generate growth. It needs to prove that the growth emerging from this unprecedented investment cycle will retain the economic characteristics that made Microsoft one of the greatest businesses in modern capitalism.
If Azure accelerates, Copilot deepens, margins stabilize and free cash flow begins to absorb the spending, I will view the current valuation as compelling.
If capital requirements continue to expand faster than the monetization evidence, I will become more patient.
Microsoft is building the machinery through which much of the world may consume artificial intelligence. On earnings night, I will be watching whether that machinery is becoming an economic engine—or simply the most expensive construction project in the history of software.
Comments
Post a Comment