Lambda, the AI Developer Cloud, raised a $480 million Series D to expand its AI cloud platform. The round included participation from Andra Capital, SGW, Andrej Karpathy, ARK Invest, Fincadia Advisors, G Squared, In-Q-Tel (IQT), KHK & Partners, and NVIDIA. Strategic investment also came from Pegatron, Supermicro, Wistron, and Wiwynn. Total equity capital raised to date is $863 million. Lambda’s mission is to provide ubiquitous and affordable access to computation. The company provides infrastructure, cloud services and software for the fine-tuning, training, and inferencing of AI models. Lambda also closed a $1.008 billion senior secured, fixed-rate loan facility on October 1. The $1.008 billion senior secured, fixed-rate loan facility, which Lambda closed on October 1, uses delayed drawdowns tied to cluster commissioning to fund GPU infrastructure for three committed customer deployments serving two investment-grade counterparties. This financing carries a 6.78% fixed interest rate and reaches final maturity on May 30, 2033. The loan is secured by the funded equipment and contracted cash flows. This transaction follows a $926 million senior secured term loan B facility priced on August 12. The facility supported GPU infrastructure for an investment-grade customer and carried a Moody’s Baa2 rating. Pricing tightened to SOFR plus 3.00%. I see Lambda using these specific credit lines to scale its on-demand cloud without relying on general corporate credit.
CoreWeave faces earnings pressure
CoreWeave maintains a massive presence in the AI infrastructure space, but its stock fell 14% after it reported a second-quarter adjusted loss of 27 cents per share. This miss exceeded analyst expectations by 6 cents. Even though revenue tripled year-over-year to $1.21 billion, the market reacted to the bottom line. Morgan Stanley analysts pointed to delayed capital expenditures in the fourth quarter as a reason for uncertainty. I find the pressure on CoreWeave’s cash flow evident as it generated $3.7 billion in operating cash flow against $14.1 billion in capital spending during the first half of the year. The company’s revenue in 2024 was eight times the prior year. CoreWeave manages its debt through large facilities, including an $8.5 billion delayed-draw term loan on March 31, 2026. This facility received an A3 rating from Moody’s and an A (low) rating from Morningstar DBRS. The floating-rate tranche carried a pricing of SOFR plus 2.25%. CoreWeave also secured a $3.1 billion facility in May 2026 and a $2.6 billion facility in August 2026. Nvidia has a commercial interest in expanding the model, as financing helps customers acquire its systems without paying the full cost upfront. It acquired Weights and Biases for $1.4 billion during the quarter. You know the players, so I will focus on the capital.
Pricing and service competition
| GPU Model | CoreWeave Price | Lambda Labs Price |
|---|---|---|
| A100 80GB | $1.21 – $2.21 /hr | $1.10 – $1.29 /hr |
| A100 40GB | $1.21 /hr | $1.10 /hr |
| H100 SXM5 80GB | $2.49 /hr | $2.49 /hr |
| H200 141GB | $2.58 /hr | Not Available |
| B200 180GB | $4.26 /hr | Not Available |
| GH200 96GB | $6.50 /hr | Not Available |
| L40 40GB | $0.78 /hr | Not Available |
| L40S 48GB | $0.98 /hr | Not Available |
CoreWeave wins enterprise customers by beating Lambda Labs on support, cluster availability at scale, and flexibility. A healthcare IT director noted that CoreWeave delivered capacity in weeks where AWS required months, and that the company beat Lambda Labs specifically on enterprise support, cluster availability at scale, and flexibility. Will the current high cost of capital stall the expansion of these neoclouds? Lambda Labs remains cheaper for A100s, with prices between $1.10 and $1.29 per hour compared to CoreWeave’s $1.21 to $2.21 range. For the H200 141GB, CoreWeave charges $2.58 per hour, while the price for a B200 180GB is $4.26 per hour. A healthcare buyer placed more than half of their spending on CoreWeave, while another technology buyer placed 10% to 25% of their GPU spending there. CoreWeave reported that 98% of its revenue came from committed take-or-pay contracts. This revenue includes a massive $30.1 billion backlog. Lambda Labs offers a startup program with $5,000 to $25,000 in credits, whereas CoreWeave reserves credits for larger venture-backed startups. Lambda provides built-in Jupyter Notebooks, while CoreWeave uses Argo.
