Lambda Labs Series E rerating

Lambda’s Series E clearing above the talk track shows that investors were no longer valuing it as a smaller backup GPU cloud, they were paying for a scarce supplier with real hyperscaler scale. By November 2025, Lambda had reached a $5.9B post money valuation after stepping up from a $2.5B Series D in February 2025, while revenue had climbed from roughly $425M annualized at the end of 2024 to $760M by the end of 2025, giving investors evidence that demand and utilization were outrunning earlier $4B to $5B expectations.1 2 3 4

• The rerating matched a business model shift. Lambda moved away from selling physical GPU boxes and self serve side products, and concentrated capacity on large reserved clusters and multi year contracts, which made its revenue look more like infrastructure backlog and less like spot compute sales.2 3
• Customer evidence helps explain why pricing power improved. At Iambic, Lambda won training workloads because it could offer cheaper H100 hours, stronger InfiniBand cluster specs, and more willingness to customize the setup than AWS or Oracle. That makes the product more than raw GPU rental, it is reserved training infrastructure tuned to how labs actually train models.4
• The closest public comp sharpened the valuation reset. CoreWeave had already shown that giant contracts can compound into cheaper financing, more GPU purchases, and faster scale. Lambda was still smaller, but at 7.8x revenue versus CoreWeave around 12x, the Series E price still left room for investors betting Lambda could keep moving upmarket.2 5

The next leg is about whether Lambda can keep converting developer goodwill into giant contracted clusters. If it keeps landing Microsoft, Nvidia, frontier lab, and enterprise scale deals, valuation will track its ability to secure power, debt, and chips, because in GPU cloud the winners are increasingly the companies that can lock supply first and monetize it second.2 4 5