Off-Prem

PaaS + IaaS

Lambda borrows half a billion bucks to grow its GPU cloud

Will buy tens of thousands of Nvidia's prized accelerators, which will be collateral for the loan


Lambda Labs, operator of a GPU-infused cloud, on Thursday revealed it has secured a $500 million loan to fund the expansion of its accelerators-as-a-service offering.

The deal – backed by Macquarie Group and Industrial Development Funding – will support the deployment of "tens of thousands" of Nvidia's fastest accelerators, which in turn will serve as collateral for the loan.

GPUs of the sort Lambda employs cost $30,000 to $40,000 a piece and are in incredibly short supply.

Lambda, which has worked on GPU systems for over a decade, has offered cloud services for the past six years. The company explained that the "asset-based" nature of the deal will allow it to rent out the accelerators to customers without long-term contracts.

Such contracts are typically the way to rent resources at the lowest possible rate. For example, on-demand pricing for Amazon's P5 instances, which include eight H100 accelerators, is $98.32 an hour, or about $12.29 per GPU. But if you commit to a three-year lease, Amazon will cut that down to $43.16/hr.

By contrast, it appears Lambda's strategy is to capitalize on demand from smaller AI startups hesitant to sign a long-term lease. Lambda lists a config comparable to AWS at $27.92 per hour.

The half billion dollars in debt financing comes in addition to the $320 million in Series C funding Lambda raked in back in February. At the time, Lambda reported the funding would go toward additional Nvidia H100, H200 and GH200 superchip capacity, along with an upgrade to its InfiniBand network.

Nvidia's H200, due out in Q2, is particularly notable as its larger HBM3e memory capacity – 141GB vs 80GB on the H100 – and faster memory bandwidth mean customers can run larger models on fewer GPUs.

If any of this sounds familiar, that's because Lambda is far from the only GPU cloud provider looking to cash in on the AI hype. Over the past few years we've seen bit barns like CoreWeave and Voltage Park appear and offer datacenters bursting with 10,000 or more GPUs available for rent at prices below those charged by larger cloud providers.

Many of these deployments have been funded by massive debt. For example, thanks to a $2.3 billion loan from Magnetar Capital and Blackstore, CoreWeave last northern summer expanded its datacenter footprint. ®

Send us news
8 Comments

Nvidia admits Blackwell defect, but Jensen Huang pledges Q4 shipments as promised

The setback won't stop us from banking billions, CFO insists

Buying a PC for local AI? These are the specs that actually matter

If you guessed TOPS and FLOPS, that's only half right

DoJ reportedly advances Nvidia antitrust probe

Uncle Sam apparently worried GPU giant may be punishing customers who shop around

Nvidia's growth slows to a mere 122 percent but it’s still topping expectations

Still growing in China, ramping Hopper prods and predicting Blackwell billions soon

Tenstorrent's Blackhole chips boast 768 RISC-V cores and almost as many FLOPS

Shove 32 of 'em in a box and you've got nearly 24 petaFLOPS of FP8 perf

Benchmarks show even an old Nvidia RTX 3090 is enough to serve LLMs to thousands

For 100 concurrent users, the card delivered 12.88 tokens per second—just slightly faster than average human reading speed

AMD's Victor Peng: AI thirst for power underscores the need for efficient silicon

Moore's Law may be running out of steam, but there are still knobs to turn and levers to pull

All that new AI-fueled datacenter space? Yeah, that's mostly ours – cloud giants

Construction surged 70% ... with 80% already snapped up

LiquidStack says its new CDU can chill more than 1MW of AI compute

So what’s that good for? Like eight of Nvidia’s NVL-72s?

Copper's reach is shrinking so Broadcom is strapping optics directly to GPUs

What good is going fast if you can't get past the next rack?

AI's thirst for water is alarming, but may solve itself

Its energy addiction, on the other hand, only seems to get worse

Canadian artist wants Anthropic AI lawsuit corrected

Tim Boucher objects to the mischaracterization of his work in authors' copyright claim