Nvidia Strikes $1.5B Deal to Rent Back Own AI Chips from Lambda

TLDRs;

• Nvidia invests $1.5 billion to lease 18,000 AI-equipped GPU servers from Lambda.
• Lambda’s GPU leasing deal positions it for an upcoming IPO and growth in AI cloud.
• Nvidia supports smaller cloud providers to compete with Amazon, Microsoft, and Google.
• Lambda expects cloud revenue to exceed $1 billion by 2026, targeting major AI developers.

Nvidia has signed a $1.5 billion deal to lease GPU servers equipped with its own chips from Lambda, a small cloud service provider preparing for an IPO.

The agreement includes a $1.3 billion, four-year lease for 10,000 servers and an additional $200 million contract for 8,000 servers, making Nvidia Lambda’s largest customer to date.

Media reports indicate that Nvidia, already an investor and supplier to Lambda, is strategically supporting smaller cloud service providers to compete with established giants like Amazon, Microsoft, and Google.

This marks a continuation of Nvidia’s efforts to expand its influence in the AI cloud ecosystem, a strategy previously seen in deals with CoreWeave, another emerging GPU cloud provider.

AI Chip Shortage Spurs Strategic Rentals

The deal underscores the severe shortage of high-performance AI chips, even affecting Nvidia, the leading chip designer globally. Lambda has been leasing Nvidia GPUs to research institutions and AI developers, creating a lucrative rental market.

By leasing its own chips back from Lambda, Nvidia ensures access to critical GPU resources while simultaneously supporting a growing AI infrastructure network.

Insiders report that Nvidia researchers will also use the leased servers, reflecting a circular model where Nvidia acts as a supplier, investor, and customer. Analysts suggest this model strengthens Nvidia’s market position while promoting smaller cloud players that rely heavily on Nvidia’s technology.

Small Cloud Providers Gain Market Boost

Lambda’s business model involves leasing server space equipped with Nvidia GPUs to customers, including major players such as Amazon and Microsoft.

While these large companies primarily use Lambda’s resources internally, Nvidia’s substantial lease will allow Lambda to expand its cloud offerings and attract new AI clients.

Executives at Lambda also revealed that the deal will support its DGX Cloud platform, where GPUs are rented to companies developing AI applications. Lambda’s recent collaboration with image generation startup Midjourney demonstrates the company’s ability to convert users from other AI chip providers, such as Google, to Nvidia GPUs.

Lambda Eyes Rapid Growth Ahead of IPO

Lambda expects cloud revenue to exceed $1 billion by 2026, with long-term targets surpassing $20 billion by 2030. The company aims to expand computing power to nearly 3 gigawatts by 2030, a significant increase from 47 megawatts in the second quarter, although specific expansion plans remain undisclosed.

The Nvidia lease deal, codenamed ‘Project Comet,’ provides both financial momentum and credibility ahead of Lambda’s anticipated public listing.

Industry analysts view the agreement as a strategic example of Nvidia’s long-term plan to support smaller cloud service providers ensuring loyalty and safeguards for Nvidia’s core GPU business, especially as larger clients like Microsoft, Amazon, and Google develop their own AI chips to reduce dependency.

The post Nvidia Strikes $1.5B Deal to Rent Back Own AI Chips from Lambda appeared first on CoinCentral.