AI is not just boosting cloud growth. It is creating a traffic jam.
After the latest round of earnings, Amazon $AMZN ( ▼ 1.56% ) , Alphabet $GOOGL ( ▼ 2.08% ) , and Microsoft $MSFT ( ▲ 1.42% ) revealed massive levels of remaining performance obligations, or RPO. These are signed cloud contracts that have not yet been delivered or recognized as revenue, essentially a waiting line of future business.
Together, the three giants are now sitting on a combined $1.1 trillion in backlog.
AI Workloads Are Filling the Queue
This surge in RPO highlights just how intense demand has become for cloud infrastructure, especially for AI-related workloads. Training models, running inference, and storing the growing flood of data all require enormous compute resources, and customers are locking in capacity years in advance.
The result is revenue that is effectively spoken for, even if it cannot be delivered immediately because of hardware, power, or data center constraints.
Overflow Demand Helps the New Players
That bottleneck is where smaller “neoscalers” come into play. Companies like CoreWeave $CRWV ( ▼ 3.03% ) and Nebius $NBIS ( ▼ 1.89% ) are positioning themselves as alternative providers of AI-heavy compute, stepping in when hyperscalers cannot meet demand fast enough.
CoreWeave itself reported a hefty $55 billion backlog last quarter, showing that even second-tier providers are struggling to keep up with customer appetite.
A Good Problem to Have
In theory, this kind of backlog offers strong visibility into future revenue for the cloud giants. But it also underscores a key tension: demand is racing ahead of supply, forcing companies to pour unprecedented amounts of money into new data centers and hardware.
For investors, that $1.1 trillion figure is both a bullish signal for long-term growth and a reminder of just how much capital will be required to turn those contracts into actual dollars on the income statement.