The stock needs roughly $7 of FY28 earnings to justify today’s price at a mid-teens multiple. Cloud can get there if commerce holds; I set the target at 18 times that earnings.
Key points
- $7.24: my FY28 earnings per ADS if cloud keeps compounding and commerce profits hold.
- 18 times: the multiple I apply to that figure, producing a $130 target.
- $44.7 billion: free-cash-flow outflow in the June quarter as servers were bought ahead of demand.
The call
The $106 price requires about $6.60 of earnings twelve months from now at 16 times, or a faster recovery in cash flow than the last two years have shown. Operating cash flow already fell from $25 billion in FY24 to $11 billion in FY26 while capex jumped.
I believe FY28 earnings can reach $7.24 per ADS because external cloud grew 45% in the June quarter and AI products are 35% of that mix. Commerce EBITDA was stable even as customer-management revenue fell 7%. The market treats the cash drain as permanent; I treat it as a three-year payback on servers.
Revenue ($ billions) and my FY27-28 estimatesCompany filings through FY26; my estimates thereafter. As of Oct 8, 2026.
Valuation
| Item | Figure |
|---|
| FY28 net income (my estimate) | $18.0 billion |
| ADS count (Sept issued shares / 8) | 2.49 billion |
| FY28 EPS | $7.24 |
| Multiple | 18 times |
| 12-month target | $130 |
Eighteen times is a discount to a US hyperscaler because of VIE risk and China consumer cyclicality, but above the mid-teens the stock has traded while cash was leaving. Upside if cloud margins keep expanding; downside if the $380 billion infrastructure plan stays cash-negative past FY28. That skew supports a buy, not a strong buy.
Risks
- Commerce transaction weakness spreads and customer-management revenue stays negative after the subsidy restatement.
- Server payback stretches past three years if utilization or chip supply disappoints.
- Further share issuance or a larger fine/probe cuts the ADS count or cash.
What would change my mind
A second consecutive quarter of negative operating cash flow with cloud growth below 30% would force the multiple down.