Why Lambda Secured Funding And Investment From Nvidia

The AI infrastructure market continues to expand as organisations seek computing resources to develop and deploy AI applications.

As a result, global demand for specialised AI hardware has created supply constraints, particularly for Graphics Processing Units (GPUs) – creating opportunities for companies that can acquire and provide access to GPU resources.

This has led to cloud computing providers emerging to fill this gap, offering organisations flexible access to AI infrastructure without requiring substantial capital investment in hardware.

Against this backdrop, Lambda, a cloud computing company that provides hardware and services for AI development, has secured US$480m in Series D funding, with Nvidia among new investors – bringing the company’s total equity to US$863m.

Who is investing in Lambda? 

Lambda, established in 2012 by AI engineers, maintains a client base of over 5,000 customers across manufacturing, financial services and government sectors – and provides both hardware and cloud services, enabling organisations to train, modify and implement AI models.

Lambda's existing investors, including 1517, Crescent Cove and USIT, participated in the current funding round, which the company plans to use the investment to scale both infrastructure and software capabilities.

The funding round was co-led by Andra Capital and SGW, the investment office of Scott Hassan, an early Google investor.

New participants include:

• ARK Invest
• Super Micro
• Pegatron
• Wistron
• Wiwynn
• Andrej Karpathy
• Fincadia Advisors
• G Squared
• In-Q-Tel (IQT)

As a result of the funding, the company's valuation has reached US$2.5bn post-investment, according to sources familiar with the transaction – the investment arriving when demand is increasing for GPUs worldwide.

Open source models reshape Lambda's market position

The AI sector has experienced significant changes in the past year, partly driven by developments in open-source models and improvements in large language model (LLM) reasoning capabilities.

These models, including Llama and DeepSeek-R1, enable organisations to access advanced AI capabilities without developing proprietary systems.

In this context, Lambda operates a cloud platform with more than 25,000 GPUs, which organisations can rent to develop and run AI applications, with an infrastructure that supports both proprietary and open-source AI models, including DeepSeek-R1.

"Lambda is really well positioned as a company to take advantage of open source AI models like DeepSeek-R1 because we have well over 25,000 GPUs on our cloud platform that can be readily repurposed to host these open source models," says Stephen Balaban, Lambda's CEO and Co-founder in an interview.

Infrastructure expansion supports growing enterprise demand

The funding will support Lambda's acquisition of Nvidia's H200 chips, the company's latest GPU designed for AI applications.

Already, Stepehen reports that enterprise customers are pre-purchasing substantial portions of Lambda's H200 capacity before public availability.

Therefore, the investment will also fund expansion of Lambda's cloud platform infrastructure and software development and the company plans to enhance its software tools for AI developers and deploy additional GPUs to meet customer demand.

For now, Lambda will continue development of Lambda Chat, a platform that hosts DeepSeek-R1 and other open source models.

Computing requirements drive infrastructure investment

Recent developments in AI model reasoning have changed the economics of AI deployment. Now, reasoning capabilities allow organisations to improve model output quality by increasing computing power during inference, the process of generating responses from trained models.

Commenting on this transition, Stephen says: "The simple math is this: if you want to 100x the amount of compute required to train Llama (about US$120m), you will need to increase your spend to around US$12bn.

“If you want 100x the amount of compute required to inference Llama, your cost may go from three cents to US$3 –a much more reasonable proposition.”

This means that the change towards inference-time computing from previous approaches that focused on increasing computing power during model training is having implications for infrastructure providers like Lambda that supply computing resources for AI applications.

"We're scaling both infrastructure and software, enabling AI developers to train, fine-tune and deploy models faster and easier than ever.

“We'll build more software tools that delight AI developers and deploy more GPUs to meet the massive customer demand,” Stephen concludes.

Explore the latest edition of Technology Magazine and be part of the conversation at our global conference series, Tech & AI LIVE.

Discover all our upcoming events and secure your tickets today.

Technology Magazine is a BizClik brand