METAL LAB

Lambda raises another $1 billion in private debt to buy NVIDIA chips

JPMorgan arranges short-term private notes so Lambda can secure GPUs to lease to Microsoft

Summary

  • AI cloud company Lambda has raised $1 billion in private short-term debt to fund purchases of NVIDIA AI chips.
  • The chips will be leased to Microsoft, marking the company's third loan-style financing deal this year.
  • Bloomberg tallies show AI-related debt raised by banks and tech firms worldwide has now topped $400 billion this year.

Buying NVIDIA chips with $1 billion in private notes

Lambda has issued a new $1 billion round of private short-term notes to fund its purchase of NVIDIA AI chips. According to Bloomberg, the deal was arranged by JPMorgan Chase, and the chips it buys will be leased out to Microsoft. Structuring the notes with a short maturity signals the plan behind it: deploy the chips quickly, start collecting lease revenue right away, and use that cash to pay down the debt.

An arrow runs from a heavy circle labeled private debt to a central node labeled Lambda, and from Lambda a solid line runs to a diamond labeled GPU chips. Between the GPU chips and a large circle labeled MS, a bidirectional dotted line hints at a cycle: chips are leased out, lease payments flow back, and that revenue repays the debt.An arrow runs from a heavy circle labeled private debt to a central node labeled Lambda, and from Lambda a solid line runs to a diamond labeled GPU chips. Between the GPU chips and a large circle labeled MS, a bidirectional dotted line hints at a cycle: chips are leased out, lease payments flow back, and that revenue repays the debt.

This is the latest in a string of debt deals Lambda has lined up over the past several months to fund GPU infrastructure earmarked for specific customers. Back in May, the company closed a $1 billion secured credit facility, and this week it announced a $926 million loan to buy NVIDIA's latest GB300 GPUs — a batch that, notably, is being supplied under contract to NVIDIA itself.

이미지: TechCrunch AI
이미지: TechCrunch AI

What Lambda does

Lambda is an AI cloud company that buys up compute chips like NVIDIA GPUs and rents them out to businesses. It doesn't build chatbots or image generators itself; instead, it supplies the hardware other companies need to train and run those models. In this business, how fast you can secure GPUs translates directly into revenue — but the catch is that buying up an entire generation of NVIDIA's latest chips takes more cash than most startups can generate on their own. That's why neocloud companies like Lambda have leaned on bonds and loans to buy chips upfront, then use the lease income that follows to pay the debt back down — a repeating cycle.

This latest $1 billion raise comes as Lambda is reportedly negotiating a $3 billion pre-IPO funding round. Back in November of last year, the company raised $1.5 billion from venture capital investors at a valuation of $5.43 billion. The fact that it's running short-term debt and equity raises in parallel suggests Lambda is keeping multiple funding channels open at once so it doesn't fall behind in the race to secure GPUs.

이미지: TechCrunch AI
이미지: TechCrunch AI

AI infrastructure growth, powered by debt

Lambda isn't alone in this approach. Bloomberg's tally shows that AI-related debt raised globally by banks and tech companies has already surpassed $400 billion this year. With GPU demand outstripping supply, venture capital alone can no longer cover the scale of funding needed — debt has effectively become the default tool for scaling AI infrastructure.

A similar pattern shows up in Together AI's deployment of 10,000 B300 GPUs at India's largest AI factory, underscoring how building GPU clusters has stopped being a one-off investment and turned into a recurring cycle of large-scale fundraising. Microsoft, for its part, is increasingly leasing GPU capacity from neoclouds like Lambda alongside its own data center investments — a workaround, reportedly, that lets major cloud providers keep pace with infrastructure demand even as chip shortages persist.

이미지: TechCrunch AI

Comments