MSFT at $516: Stronger Earnings, but Is It Better Value?
Same share price.
Stronger earnings.
A harder cash-flow test.
MSFT is back near $516. The investment question is whether a larger earnings base outweighs the growing cost of building the AI economy.
A stock returning to a familiar price can feel like a wasted journey. Investors remember the purchase price, the uncomfortable decline, and the months spent waiting to recover. What they may overlook is the business changing underneath that price.
Using $516.17 as the September 25 closing-price reference, this MSFT stock analysis asks a more useful question than whether Microsoft has recovered: how much earning power does that price now purchase? The reference comes from historical market data; it is not a live quote.
The central thesis is straightforward. Earnings growth can reduce valuation even when a share price makes little net progress. However, a lower earnings multiple does not automatically make an investment safer, especially when its infrastructure requirements are expanding.
1. Stocks can correct through price or time
A price correction happens when the share price falls relative to earnings. A correction through time happens when earnings catch up with a previously demanding valuation. Both can compress the price-to-earnings ratio, but their effect on the investor’s experience is different.
Consider a hypothetical company trading at $500 with annual earnings per share of $20. Its P/E is 25. If EPS rises to $26 while the price remains $500, its P/E becomes 19.23. Earnings increased 30%, while the multiple fell approximately 23.1%.
The distinction matters because price alone is an incomplete measure of value. Paying the same amount for a larger sustainable earnings stream can improve the investment proposition. Yet sustainability is doing essential work in that sentence: an accounting gain or temporary profit surge may not deserve the same multiple as recurring operating earnings.
There is also no promise of a painless holding period. Similar starting and ending prices can conceal a severe intervening drawdown. Inflation, foregone returns elsewhere, taxes, and dividends affect the investor’s outcome. A time correction describes a valuation mechanism, not a guarantee that waiting creates wealth.
2. Microsoft’s earnings base has expanded
| Metric | FY26 Q4 | FY26 full year |
|---|---|---|
| Revenue | $90.007B; +18% | $331.839B; +18% |
| GAAP diluted EPS | $4.81; +32% | $17.95; +32% |
| Adjusted diluted EPS | $4.74 | $17.28 |
Source: Microsoft FY26 Q4 release. Fiscal year ended June 30, 2026. Adjustments address OpenAI investment effects; “adjusted” does not automatically mean every unusual item is excluded.
For valuation, the important exercise is choosing a consistent denominator. Compare GAAP earnings with GAAP earnings, or a clearly defined adjusted measure with the same measure in the previous period. Switching definitions can manufacture an apparent improvement that shareholders cannot actually spend.
Another useful distinction is between total earnings growth and growth per share. Share issuance, employee compensation, and repurchases can change the relationship. Investors own a claim per share, so a growing business is most valuable when the economic benefit reaches that individual claim.
A flat price alongside rising earnings therefore deserves investigation. It does not, by itself, prove the market is wrong. Investors may be discounting slower future growth, greater capital intensity, or a higher required return. The task is to decide whether that discount is proportionate.
3. Azure demand is substantial, but contracts are not cash
Microsoft’s earnings call reported Azure annual revenue above $100 billion, growing 41%; quarterly growth was 43%. Commercial remaining performance obligations reached $678 billion, up 84%, and paid Microsoft 365 Copilot seats exceeded 30 million.
These measures illuminate different stages of commercialization. Revenue reflects recognized activity. RPO describes contracted future obligations. Paid seats indicate distribution and monetization. Treating them as interchangeable would exaggerate how much future business has already become shareholder cash.
A large contract can strengthen visibility while also requiring substantial investment before delivery. Its economic value depends on timing, service costs, customer performance, and the terms governing consumption. A dollar of backlog is not a dollar of profit, and its present value changes with the time needed to collect it.
Similarly, a paid software seat does not reveal the margin on every interaction. Frequent AI usage may improve customer retention while increasing inference costs. Attractive economics require a balance between subscription revenue, utilization, customer value, and the expense of serving that usage.
4. The cash-flow bridge is the real stress test
For this analysis, free cash flow equals operating cash flow minus cash additions to property and equipment. This makes the calculation reproducible and keeps noncash lease additions separate.
| Measure | FY25 | FY26 | Change |
|---|---|---|---|
| Operating cash flow | 136.162 | 182.935 | +34.4% |
| Cash property and equipment additions | 64.551 | 115.948 | +79.6% |
| Calculated free cash flow | 71.611 | 66.987 | −6.5% |
Source: Microsoft’s cash-flow statement; changes and free cash flow calculated from reported figures. Swipe tables horizontally on mobile.
The issue is visible without a complicated model: additional operating cash was more than absorbed by additional cash investment. Earnings and cash available after infrastructure spending can move in opposite directions.
Investment is not inherently bad. Spending today can create profitable capacity tomorrow. The danger is paying for that future capacity as though its eventual utilization, pricing, and useful life were already certain. Servers can become obsolete, customers can renegotiate, and power or construction constraints can postpone revenue.
The quarterly picture was stronger: $55.441 billion of operating cash less $35.802 billion of cash equipment additions produced $19.639 billion of free cash flow. Management separately discussed approximately $41 billion of capital expenditure including finance leases. Those are different measures, not conflicting versions of the same cash payment. Source: earnings call.
Nor must Azure’s percentage growth exceed CapEx percentage growth for free cash flow to recover. The relevant arithmetic concerns absolute dollars across the company: the increase in operating cash must exceed the increase in cash investment. Margin, working capital, software revenue, and lease payments all influence that outcome.
5. Peer comparisons need matching definitions
| Company | Free cash flow | Definition to notice |
|---|---|---|
| Microsoft | $19.639B | Operating cash less cash equipment additions |
| Alphabet | −$5.855B | $39.069B operating cash less $44.924B CapEx |
| Meta | $0.784B | Company measure also deducts finance lease principal |
Sources: Microsoft, Alphabet SEC earnings exhibit, and Meta investor relations. Accounting periods match; FCF definitions are not identical.
The comparison supports a narrow conclusion: Microsoft retained a larger quarterly cash surplus under these measures. It does not establish permanent superiority. Construction schedules, supplier payments, working capital, and investment cycles can produce substantial quarter-to-quarter differences.
It is also too simple to explain the entire gap as subscriptions versus advertising. Revenue mix matters, but infrastructure intensity, payment timing, and financial definitions matter too. A durable comparison requires several periods and consistent treatment of leases, rather than selecting one favorable quarter.
Shareholder distributions provide another perspective. Microsoft’s September dividend announcement increased the quarterly payment from $0.91 to $0.98, approximately 7.7%, rounded by the company to 8%. That supports an income discussion, but a dividend increase alone cannot establish that AI investment earns an adequate return. Official announcement.
6. Maia 200 could improve unit economics
Microsoft introduced Maia 200 as an inference accelerator. Its published specifications include more than 140 billion transistors, 216GB of HBM3e memory, and over 10 petaflops at FP4 precision. Microsoft claims roughly 30% better performance per dollar than the latest-generation hardware in its fleet. These are company specifications and claims, not an independent universal benchmark. Official Maia 200 announcement.
For investors, the interesting variable is useful work delivered per dollar, rather than an isolated peak-compute number. A specialized accelerator can make sense when workloads are predictable and software can be optimized around them. The economic payoff weakens if migration costs, poor utilization, or limited workload compatibility offset the hardware advantage.
There is an important mathematical distinction: 30% better performance per dollar does not mean costs fall 30%. Under a simplified equal-output comparison, cost becomes one divided by 1.30, implying approximately 23.1% less spending. Real deployments add memory, networking, power, software, and operational expenses.
Lower unit costs can also stimulate more usage. That may help revenue and customer value while keeping total spending elevated. Consequently, a cheaper inference request need not produce an immediate decline in the total infrastructure budget.
Custom silicon is therefore one route toward better cash economics, alongside higher utilization, improved software, pricing, and product mix. It should not be treated as the only solution or as automatic proof that competing accelerators lose their role.
7. Is MSFT better value at the same price?
At the reference price, dividing by FY26 GAAP EPS produces approximately 28.8 times earnings. Using the company’s adjusted EPS instead produces approximately 29.9 times. These are backward-looking calculations, not forward estimates or a declaration of fair value.
A forward P/E requires an explicitly dated earnings forecast. Different forecast periods, accounting adjustments, or analyst sets can produce different answers. A lower forward multiple usually reflects a higher forecast denominator, but it does not independently prove the quality or durability of that forecast.
The most useful valuation question is what operating outcome the price requires. If future earnings grow but investors accept a lower multiple, returns can disappoint. Conversely, modest growth combined with stable valuation and cash distributions can still generate a positive result.
| EPS growth assumption | Ending P/E | Implied price change | Lesson |
|---|---|---|---|
| +20% | 30× | +20% | Stable valuation passes earnings growth through |
| +20% | 25× | 0% | Multiple compression absorbs growth |
| +10% | 24× | −12% | Growing earnings do not prevent losses |
Illustrations only, excluding dividends, taxes, and trading costs. Formula: (1 + EPS growth) × ending P/E ÷ starting P/E − 1. No scenario probabilities are assigned.
The bull case is improving cash conversion as new capacity fills and unit economics strengthen. The base case is continued earnings growth with investment absorbing much of the cash benefit. The bear case combines weaker monetization with persistent infrastructure commitments and multiple compression.
These scenarios are more useful than labeling Microsoft simply cheap or expensive. Each identifies evidence that could change the conclusion, while avoiding the false precision of an unsupported price target.
Separate a good company from a good entry price
An investment can fail to meet expectations even when the company executes well. Suppose an investor expects rapid earnings growth for several years and pays a premium that assumes nearly flawless delivery. If results are merely good, the market may lower the multiple. The business improves, but the investor’s starting expectations were too demanding.
The reverse can also occur. A company with visible challenges may generate attractive returns if the price already discounts an outcome worse than the one eventually delivered. This is why business quality and investment attractiveness should be assessed separately. Microsoft’s competitive position is relevant, but it cannot eliminate the need to examine the assumptions embedded in the purchase price.
A useful practice is to write down the conditions required for the investment to work before considering a transaction. These might include sustained customer spending, improving cash conversion, and infrastructure returns above the cost of capital. Then identify which evidence would weaken those conditions. That turns a broad technology narrative into a thesis that can actually be tested.
Distinguish capacity expansion from economic value creation
Installed computing capacity is an input, not the final result. A new facility can increase available supply while initially depressing utilization and cash returns. Later, the same facility may become highly productive as customers ramp usage. Evaluating this sequence requires patience, but patience should be tied to measurable progress rather than an indefinitely extended deadline.
Consider a hypothetical project costing $10 billion. Whether it creates value depends on the future cash it generates after operating costs and ongoing investment, the length of that cash stream, and the return investors require. A large revenue number cannot answer the question without those other inputs. Nor does a fully occupied facility necessarily earn attractive returns if prices are too low.
For a diversified business, precise project-level economics may not be publicly available. Investors can still look for consistency across aggregate indicators: improving margins, stronger operating cash, disciplined investment, and manageable commitments. No single quarter settles the issue, but a persistent combination of those signals is more informative than one spectacular adoption statistic.
That framework also clarifies what would justify a higher valuation. The market needs evidence that growth is becoming more economically productive, not simply more capital intensive. If stronger earnings increasingly translate into distributable cash, the case improves. If they repeatedly require larger commitments with uncertain payback, a lower P/E may be an appropriate discount rather than a missed opportunity.
8. Follow the business, then use the chart
A chart can help organize timing and risk, but familiar price levels do not validate intrinsic value. Any claim about support, resistance, or moving averages needs dated price data and a stated calculation window. This article therefore does not present unverified trading zones as established signals.
A practical earnings review should focus on five connected tests:
- Demand quality: Separate recognized revenue from contracted obligations and assess how broad the customer base is.
- Monetization: Look beyond adoption totals to retention, pricing, usage costs, and evidence that customers receive measurable value.
- Cash conversion: Reconcile operating cash, equipment payments, and lease obligations across several quarters.
- Capital productivity: Ask whether deployed capacity generates enough incremental cash to justify its cost.
- Per-share outcomes: Track diluted earnings, distributions, and dilution together rather than assuming corporate growth automatically reaches shareholders.
The strongest confirmation would be improving cash generation after investment without sacrificing the competitiveness of the business. The clearest challenge would be an expanding gap between impressive adoption statistics and weak economic returns.
Microsoft illustrates why an unchanged price deserves a fresh analysis. A larger earnings base can improve relative value, but the final investment judgment still depends on the price paid for future cash, not merely the growth already reported.
Frequently asked questions
What does a correction through time mean?
It means earnings rise while the share price stays broadly unchanged, reducing the earnings multiple. It describes valuation adjustment, not guaranteed investment returns.
Does stronger EPS make Microsoft stock low risk?
No. Earnings quality, capital requirements, competition, interest rates, and the starting valuation still matter. A lower multiple can coexist with higher business risk.
Why focus on free cash flow instead of earnings alone?
Infrastructure spending can consume cash before its cost fully appears in earnings through depreciation. Reading both measures helps distinguish accounting profitability from cash remaining after investment.
Is RPO equivalent to future guaranteed profit?
No. Contracted obligations still require delivery and carry costs, timing considerations, and counterparty exposure. Revenue recognition and cash collection are separate steps.
Can Maia 200 solve the AI spending problem?
It may improve economics for suitable workloads, but deployment costs and utilization determine the realized benefit. More efficient hardware is one input into returns, not a complete investment thesis.
Should investors buy simply because the price has recovered?
A recovery is not sufficient evidence. Evaluate sustainable per-share earnings, cash conversion, and valuation against your own return requirements and capacity to absorb losses.
Sources and methodology
Company disclosures establish reported facts; calculations and investment interpretations are this article’s analysis. Forecasts and hardware performance claims retain the uncertainty of their underlying assumptions.
- Microsoft FY26 Q4 earnings release and financial statements — earnings, adjustments, and cash-flow inputs.
- Microsoft FY26 Q4 earnings call — operating metrics and investment guidance.
- Microsoft Maia 200 announcement — technical specifications and performance claims.
- Microsoft September dividend announcement — declared payment increase.
- Alphabet Q2 2026 earnings exhibit filed with the SEC — peer cash-flow inputs.
- Meta Q2 2026 earnings release — peer free cash flow and definition.
- Yahoo Finance MSFT historical prices — market-data reference; secondary source, not a company disclosure.
Disclaimer: This article is for education and general information, not personalized investment advice. Figures are dated and may change. Investors should assess their own circumstances and risks before making decisions.