Microsoft, the Bill for Artificial Intelligence
Microsoft closed June down 19%, its worst month since 2000, and is now trading around $387, approximately 24% below its year-opening levels. The drag is not coming from operating results, which remain solid: revenues are up 18% and Azure up 40%. The market is repricing the approximately $190 billion the group expects to invest in 2026 and the pressure this places on cash flow. At roughly 20 times forward earnings, the stock trades below its historical averages, with the heaviest trading volume concentrated at the $373 level, where it found initial support.
Key data
| Indicator | Value | Reading |
|---|---|---|
| Market capitalisation | ~$2,900bn | Fourth largest company in the world by value |
| Quarterly revenues (FY26 Q3) | $82.9bn | +18% year-on-year |
| Operating margin | 46,3% | Expected to decline towards ~44% in Q4 |
| Azure growth | +40% | Accelerating from +39% |
| AI revenue (run rate) | $37bn | +123% year-on-year |
| 2026 capital expenditure | ~$190bn | +61%, approximately $38bn above expectations |
| Free cash flow (quarterly) | $15.8bn | Down from $20.3bn a year ago |
| Annual dividend | 3,64 $ | Yield ~0.9%; 16 consecutive increases |
1. Executive summary
The Microsoft case revolves around the price being paid for future growth. The group continues to expand revenues and earnings at a double-digit pace, with cloud and artificial intelligence as its engines; the market, however, has shifted its focus to the cost of fuelling that race: approximately $190 billion in capital investment in 2026, which is compressing near-term free cash flow. The result is the sharpest correction in a quarter of a century, with the stock falling from above $470 to $387.
On the technical side, the heaviest trading volume has clustered around the $373 level, which now serves as initial support. On the valuation side, the multiple has compressed to approximately 20 times forward earnings, against a five-year average close to 32: a discount that reflects genuine concerns about margins and returns on AI investment, not a deterioration of the underlying business. The next verdict arrives with the 28 July earnings release.
2. The signal: the chart and volume nodes
The VolumeLeaders chart shows where the heaviest trading volume has occurred, rather than the day-to-day price action. Each flagged level indicates a price point at which, over recent months, a significant notional value changed hands: a trace of where institutional size has been active, rather than a price target. Those levels tend to act as areas of support or resistance when price returns to them.
The greatest concentration — approximately $48 billion in notional value — sits at $373: this is the threshold at which the stock found its first support after sliding to a low of around $350 at the end of June. Further overhead remain the nodes at $415, $435 and $450, all above current levels.
In the chart below we observe the stock's trajectory since the start of June and, on the right, the volume profile by price level: the red bar at $373 is by far the densest.
3. The company by segment
Microsoft reports its financials across three segments. In the third quarter of fiscal year 2026, Intelligent Cloud, which encompasses Azure, generated $34.7 billion, up 30%. Productivity and Business Processes, which includes Microsoft 365, Copilot and LinkedIn, contributed $35.0 billion, up 17%. More Personal Computing, which covers Windows, devices and gaming, came in at $13.2 billion, down a modest 1%.
The centre of gravity has thus shifted decisively towards the cloud and recurring services, while the segment most closely tied to the personal computer cycle remains the laggard. In the chart below, we examine the relative weight of the three segments and their growth trajectory.
Profitability remains robust: operating income for the quarter came in at $38.4 billion, up 20%, with an operating margin of 46.3%. It is here, however, that the first cause for concern emerges, as infrastructure expenditure is beginning to weigh on margins.
4. The AI engine and the data centre bill
The growth of cloud and AI carries a physical price tag: servers, chips and data centres. In the image below, we look inside a data centre — the infrastructure underpinning Azure and artificial intelligence services, which is absorbing an ever-growing share of the group's capital expenditure.
The AI business is running at an annualised rate of $37 billion, more than doubling over twelve months, and paid seats for Microsoft 365 Copilot have exceeded 20 million, up from 15 million in January. The crux of the matter is spending. Microsoft has guided for capital investments of approximately $190 billion in 2026, an increase of more than 60% and roughly $38 billion above analyst expectations, with a single quarter already exceeding $40 billion.
Microsoft is not alone in this race: in the chart below, we examine capital expenditure across the major technology platforms, providing useful context for the scale of the phenomenon.
The blanket is too short: to avoid falling behind in the AI race, the group is compressing near-term cash generation — the very metric that has historically justified its valuation premium. Added to this is the pressure on gross margins, which fell in the quarter to 67.6%, their lowest level since 2022, due to depreciation charges related to new data centres.
5. Price structure
Translated into price levels, the technical picture is clear. The $373 mark, the heaviest volume node, serves as the first support: as long as it holds, the stock remains in a stabilisation phase following the decline. The late-June low of approximately $350 is the next reference point, and a break below it would reopen a bearish leg.
To the upside, the nodes at $415 and $435 represent the first areas of resistance. These are levels at which many investors bought at higher prices and are now sitting on losses: a return of the price towards those areas tends to attract supply from holders looking to exit at breakeven. They should be read as levels the market is watching, not as infallible barriers.
6. Valuation and peer comparison
At approximately 20 times forward earnings, Microsoft trades at a significant discount to its own five-year average of around 32 times, and below the technology sector average. In comparison with the other major players in the space, the multiple remains among the most modest: on a trailing twelve-month basis, Microsoft trades at around 23 times earnings, below Alphabet (approximately 28), Amazon (approximately 29) and Nvidia (approximately 30).
On the shareholder returns front, the policy remains generous: in fiscal year 2025, Microsoft returned approximately $42.5 billion through dividends and buybacks, marking the sixteenth consecutive annual dividend increase and a new share repurchase programme. The stock is therefore pricing in a more cautious scenario than the operational trajectory would suggest. A compressed multiple, however, is not in itself a catalyst: for the valuation to re-expand, AI investments will need to begin showing measurable returns in cash flow, not just in revenues.
7. The moment: the 28 July earnings release
The next catalyst has a date. Microsoft is expected to publish its fourth fiscal quarter results around 28 July. The consensus points to earnings per share of $4.33 and revenues of approximately $89.4 billion. For the full fiscal year 2026, estimates converge on revenues of between $324 billion and $327 billion and earnings per share of between $16.46 and $17.10.
Market participants will be focused on three things: the resilience of Azure's growth, the capital expenditure guidance for the new fiscal year, and the first tangible signals on the return from AI spending. With a stock that has already corrected significantly, a classic buy the rumour, sell the news dynamic cannot be ruled out: part of the recent recovery may already be pricing in a solid earnings report, leaving the share price exposed to profit-taking should the numbers merely meet expectations.
8. Investment view
| Area | Valuation | Rationale | What to monitor |
|---|---|---|---|
| Cloud & AI growth | Constructive | Azure +40%, AI run rate $37bn | Azure momentum, contract backlog |
| Cash & investments | Risk | $190bn capex, declining free cash flow | Free cash flow, FY27 capex guidance |
| Valuation | Optionality | ~20x vs. 5-year average ~32x | Returns on AI investments |
| Margins | Neutral | Op. margin 46.3% but expected ~44% | Gross margin and depreciation |
| Shareholder returns | Constructive | ~$42.5bn returned in FY25 | Dividend and buyback |
9. Scenario map
| Scenario | Indicative probability | Conditions | Price implication |
|---|---|---|---|
| Base | ~55% | Azure holds, capex confirmed, margins in slight decline | Consolidation between $373 and $420 |
| Optimistic | ~25% | First measurable AI returns, stable margins | Multiple re-expansion towards $470 and above |
| Risk | ~20% | Margins eroded by competition, capex still rising | Break below $350, return to lows |
Probabilities are qualitative editorial estimates, not forecasts: they serve to rank scenarios, not to quantify their outcome.
Principal sources consulted
- Chart and volume levels: VolumeLeaders.com (MSFT, data as of 6 July 2026).
- FY26 Q3 results and segments: Microsoft Investor Relations; Futurum; GeekWire (April 2026).
- Capex, margins and cash flow: CNBC, Yahoo Finance, Qz, The Motley Fool (June–July 2026).
- Valuation, multiples and peer comparison: StockAnalysis.com, Macrotrends, GuruFocus, FinanceCharts (July 2026).
- Dividend and shareholder returns: Forbes, Kiplinger, Nasdaq (2025–2026).
- Earnings date: TipRanks, MarketBeat, Nasdaq (MSFT earnings calendar).
- Images: Wikimedia Commons (portrait of Satya Nadella; data centre, BalticServers.com, CC BY-SA 3.0).
The Financial Spectator · TFS Focus