October 6, 2026
Lambda prepares pre-IPO round, aiming to raise up to $4 billion
Lambda is raising up to $4 billion ahead of a planned IPO in 2027, according to WSJ reporting on Oct 6.
From June–September 2026, Lambda increased its order book from $15 billion to $50 billion. Anthropic commitments account for $35 billion of that total.
Lambda rents out GPUs for AI workloads. Orders for future computing capacity are not yet recognized revenue, but they show the scale of capacity the company must deliver to customers.
Lambda's financing terms
New round. Blackstone and Coatue Management are leading a funding round of up to $4 billion at a pre-money valuation of $14.5 billion. Nvidia is already among the company's investors.
The previous Series E, announced on Nov 18, 2025, raised over $1.5 billion for Lambda. TWG Global led the round, with USIT among the participants.
Loan for clusters. On Oct 1, 2026, Lambda closed a separate $1.008 billion loan at a fixed annual interest rate of 6.78%. Final repayment is due on May 30, 2033.
The loan finances GPU infrastructure for three agreed deployments serving two customers. Lambda receives the funds as the clusters go live.
Michel Combes has led the company since May 2026. Co-founder Stephen Balaban moved to the chief technology officer role.
What you can rent for your project
Lambda offers hourly GPU rentals: you can rent a single machine to work with a model. As of Oct 6, 2026, the company lists the following prices per GPU, with taxes charged separately:
| GPU | Memory | Hourly price | | --- | --- | --- | | H100 PCIe | 80 GB | $3.29 | | H100 SXM | 80 GB | $4.29 | | B200 SXM6 | 180 GB | $6.99 |
To connect to a machine, add an SSH key to your workspace and run `ssh -i '<SSH-KEY-FILE-PATH>' ubuntu@<INSTANCE-IP>`. Open JupyterLab through Instances → Cloud IDE → Launch.
Original source: [TechCrunch, Oct 6, 2026](https://techcrunch.com/2026/10/06/ai-computing-startup-lambda-to-raise-4b-ahead-of-planned-ipo). The funding terms and order book figures come from WSJ, the loan and leadership changes from Lambda's blog, and pricing and connection instructions from Lambda's materials as of Oct 6.
By 2030, Lambda plans to bring its computing capacity under management to 3 GW.