The market is too pessimistic about cloud sales after new customer agreements and stronger rental pricing. I reverse my Sell to Buy, with a $185.21 twelve-month target, while allowing for substantial borrowing and employee share issuance.
Key points
- $158 billion of 2027 sales supports today's price under my valuation assumptions.
- $190 billion of 2027 sales is my forecast, replacing $125 billion.
- $35 billion of nine-month cash burn keeps financing central to the call.
The call
The market is too pessimistic about the sales ramp. At my retained revenue multiple, today's price requires about $158 billion of recognized 2027 sales. Management's first-quarter guidance midpoint alone represents a $140 billion annual pace. That is not proof of delivery, but it materially weakens my previous demand objection.
Cloud spending still exceeds the sales it supports- Reported revenue
- Equipment and construction spending
Source: October 20, 2026 earnings release; Q3 2026, $ billions.
I now expect stronger monetization, not merely more installed machines. My 2027 forecast includes $132 billion of cloud sales and $58 billion elsewhere. Cloud assumes average rented capacity of three gigawatts at $44 per watt annually. CoreWeave's reported price increases independently support scarcity; my previous rental-price assumption was too low.
Valuation
| Our October 2027 valuation assumptions | Value |
|---|
| Recognized revenue, calendar 2027 | $190 billion |
| Core business value / revenue | 12 times |
| Integrated engineering and scarcity premium | 2 times revenue |
| Debt less cash and securities | $30 billion |
| Diluted shares | 14.2 billion |
| Twelve-month target | $185.21 a share |
The combined multiple implies 56 times operating profit at an eventual 25% margin, so this is no bargain on mature earnings. The premium rewards launch, connectivity and compute integration; abundant competing capacity would remove it. At $30 per watt annually, my target falls to about $144, versus $203 at $50, supporting Buy rather than Strong Buy.
The cloud rental price decides the upsideOur October 20, 2026 estimates; 3 GW average rented in 2027; target in dollars per share.
My funding model assumes $155 billion of 2027 equipment spending against $85 billion of operating cash inflow. The share estimate adds an allowance for employee awards to September's outstanding shares; it is not the quarterly earnings denominator. No Tesla merger value is included.
Risks
- Customers can leave after ramp periods; watch renewals and rental prices.
- Construction can outrun funding; watch borrowing costs, equipment payments and new shares.
- Deployment can slip; watch energized capacity and repeat Starship flights, not announced projects.
What would change my mind
First-quarter 2027 sales below $33 billion alongside further spending increases would overturn my monetization forecast.