CoreWeave Revenue Doubles as AI Backlog Hits $99 Billion The Rise of the Infrastructure Titan.

CoreWeave Revenues Double to $2.07 Billion as Backlog Nears $100 Billion Milestone

CoreWeave, the specialized GPU cloud provider, has reported an explosive start to 2026. The company’s first-quarter revenue surged to $2.078 billion, more than doubling compared to the same period last year. However, this rapid scaling came with a net loss of $740 million, widening from the previous year due to significant operational expenditures required to fuel its massive expansion.

The highlight of the report is the company’s staggering backlog, which has ballooned to $99.4 billion, reflecting the insatiable global demand for AI compute power.

Scaling to 8 Gigawatts

CEO Michael Intrator highlighted that CoreWeave’s operational capacity has already surpassed 1 gigawatt across its data center fleet, with a strategic roadmap to hit 8 gigawatts by 2030. Intrator emphasized that the company is financially and structurally prepared for this growth, particularly as the AI market shifts its focus from model training to large-scale inference, which requires even more persistent and distributed compute resources.

Diversifying the Client Portfolio

In a significant move away from customer concentration, CoreWeave has successfully diversified its revenue streams. While Microsoft accounted for 62% of the company’s revenue in 2024, CoreWeave now boasts 10 major clients with contracts exceeding $1 billion each. This shift marks CoreWeave’s transition from a niche provider into a cornerstone of the global AI infrastructure.

The CEO's mention of moving towards inference (real-world AI deployment after training is complete) is a crucial turning point. Training projects are typically short-term, requiring massive computing power. Inference, however, involves millions of users accessing AI simultaneously, demanding 24/7 data centers. This results in CoreWeave's revenue shifting towards more stable recurring revenue.

The reason CoreWeave can afford massive losses yet still expand is its debt-financing model. The company uses high-value, highly-demand NVIDIA GPUs as collateral to borrow billions of dollars to build new data centers. This allows them to expand capacity faster than major cloud providers at certain times.

The 8 gigawatt target by 2030 is enormous (equivalent to the power needed to power several large cities). This reflects that CoreWeave's challenge is no longer just sourcing GPUs, but competing for "electricity" and building complex cooling systems—factors that will determine who truly leads the AI ​​cloud market.

 

Copilot Cowork Adds Mobile Support and Advanced Skills. 

 

Source: CoreWeave 

💬 AI Content Assistant

Ask me anything about this article. No data is stored for your question.

CoreWeave Revenues Double to $2.07 Billion as Backlog Nears $100 Billion Milestone

CoreWeave, the specialized GPU cloud provider, has reported an explosive start to 2026. The company’s first-quarter revenue surged to $2.078 billion, more than doubling compared to the same period last year. However, this rapid scaling came with a net loss of $740 million, widening from the previous year due to significant operational expenditures required to fuel its massive expansion.

The highlight of the report is the company’s staggering backlog, which has ballooned to $99.4 billion, reflecting the insatiable global demand for AI compute power.

Scaling to 8 Gigawatts

CEO Michael Intrator highlighted that CoreWeave’s operational capacity has already surpassed 1 gigawatt across its data center fleet, with a strategic roadmap to hit 8 gigawatts by 2030. Intrator emphasized that the company is financially and structurally prepared for this growth, particularly as the AI market shifts its focus from model training to large-scale inference, which requires even more persistent and distributed compute resources.

Diversifying the Client Portfolio

In a significant move away from customer concentration, CoreWeave has successfully diversified its revenue streams. While Microsoft accounted for 62% of the company’s revenue in 2024, CoreWeave now boasts 10 major clients with contracts exceeding $1 billion each. This shift marks CoreWeave’s transition from a niche provider into a cornerstone of the global AI infrastructure.

The CEO's mention of moving towards inference (real-world AI deployment after training is complete) is a crucial turning point. Training projects are typically short-term, requiring massive computing power. Inference, however, involves millions of users accessing AI simultaneously, demanding 24/7 data centers. This results in CoreWeave's revenue shifting towards more stable recurring revenue.

The reason CoreWeave can afford massive losses yet still expand is its debt-financing model. The company uses high-value, highly-demand NVIDIA GPUs as collateral to borrow billions of dollars to build new data centers. This allows them to expand capacity faster than major cloud providers at certain times.

The 8 gigawatt target by 2030 is enormous (equivalent to the power needed to power several large cities). This reflects that CoreWeave's challenge is no longer just sourcing GPUs, but competing for "electricity" and building complex cooling systems—factors that will determine who truly leads the AI ​​cloud market.

 

Copilot Cowork Adds Mobile Support and Advanced Skills. 

 

Source: CoreWeave