Lambda Borrows Another $1 Billion for Nvidia Chips as Central Banks Warn About AI’s Debt Load

Lambda, the Nvidia-backed AI cloud provider, has raised about $1 billion in private, short-dated debt to buy more Nvidia GPUs, its second billion-dollar-scale financing deal in barely three weeks. The raise, arranged by JPMorgan Chase and first reported by Bloomberg, comes as economists at the European Central Bank and the Bank for International Settlements have spent recent months warning that the debt structures underwriting the AI infrastructure buildout are opaque enough to hide how much risk is actually being taken on.

A GPU purchase backed by a lease, not by Lambda’s own books

JPMorgan structured the roughly $1 billion private placement and marketed it to institutional investors, according to Bloomberg’s reporting, independently picked up by TechCrunch, TheNextWeb, and Briefs.co. The proceeds will fund Nvidia GPUs that Lambda will lease to Microsoft, which means the debt is effectively secured by Microsoft’s willingness to keep paying for compute, not by Lambda’s own revenue. Neither Lambda, Nvidia, Microsoft, nor JPMorgan had commented on the terms by publication time; JPMorgan declined to comment, and Bloomberg attributed the deal’s details to people who were not authorized to discuss it publicly.

It is the second time in weeks Lambda has used this playbook. In mid-August, the company closed a separate $926 million loan to fund Nvidia GB300 GPUs for a deployment it is contractually obligated to provide to Nvidia itself, a company that is simultaneously Lambda’s investor, its chip supplier, and now effectively a customer of its own hardware. Lambda also closed a $1 billion secured credit facility in May.

The bigger number: a possible IPO at a $12 billion valuation

The debt deals are unfolding alongside a much larger equity conversation. Lambda is reportedly in talks to raise as much as $3 billion at a valuation of $12 billion or higher, positioning the company for a possible initial public offering as soon as 2027, according to Bloomberg’s reporting as covered by TechFundingNews. The company, founded in San Jose in 2012 by twin brothers Stephen Balaban and Michael Balaban, has already collected multiple term sheets, though people familiar with the talks cautioned that the final size, valuation, and timing could still change. Lambda is expected to generate more than $1.5 billion in revenue in 2026. Its most recent equity round closed in November 2025: more than $1.5 billion led by TWG Global, with Andra Capital, Scott Hassan’s family office SGW, OpenAI co-founder Andrej Karpathy, Cathie Wood’s ARK Invest, and Nvidia among the other backers, at a $5.43 billion post-money valuation, according to PitchBook data cited by TechCrunch.

Lambda is not racing alone. CoreWeave, which went public in 2025, took a $2 billion equity investment from Nvidia in January 2026 and priced an upsized $3.5 billion convertible-notes offering in April. Crusoe closed a $1.38 billion Series E at a $10 billion valuation backed by Mubadala Capital, Valor Equity Partners, and Nvidia, and is now in talks to raise $3 billion at a valuation near $30 billion. Nscale, based in London, is separately preparing a US IPO as early as September on the strength of a $51 billion contracted-revenue backlog.

Europe’s central bank sees the same pattern in Nebius

Lambda’s structure is not unique to it. Amsterdam-based neocloud Nebius used the same approach earlier this year, borrowing $775 million against its own GPUs in its first secured debt raise. Nebius holds a five-year Microsoft contract worth $19.4 billion and has said it could securitize roughly $40 billion of contracts in total, according to TheNextWeb’s reporting. Bloomberg’s own tally, cited by TechCrunch, puts total AI-related debt raised globally in 2026 so far at more than $400 billion.

What the ECB and BIS are actually worried about

The European Central Bank’s May 2026 Financial Stability Review flagged private credit as a sector that could threaten financial stability, warning specifically about its “opaque valuation practices and limited liquidity” and about portfolios concentrated in a handful of US issuers whose valuations track the continued strength of the AI narrative, according to TheNextWeb’s account of the review. ECB Vice-President Luis de Guindos has separately called private credit, a sector estimated at $1.5 trillion to $2 trillion in global assets, a significant emerging risk, noting that funds have piled into a small number of industries, technology among them, in ways that could amplify losses if the sector stumbles, according to CryptoBriefing’s coverage of the same review.

The Bank for International Settlements went further in its Annual Economic Report 2026, released in June. The BIS compared the current AI investment boom to historical episodes like the 1840s British railway bubble and the 2000 dot-com crash, writing that “these episodes ended with an eventual reversal in investment, inducing economy-wide recessions,” according to Fortune’s reporting on the report. The BIS singled out the financing structures behind AI infrastructure specifically, including hyperscalers taking equity stakes in AI labs that then commit to multi-year chip or compute purchases from those same hyperscalers, and data centers leased back from third-party contractors under long-dated contracts with embedded exit clauses. “The terms of such deals are typically poorly disclosed,” the BIS wrote, “with risks of the same asset being pledged multiple times.”

That description sits close to the structure Lambda and Nebius are both now using: debt raised against a single customer’s lease commitment, with the underlying hardware collateral sitting somewhere inside a chain of contracts that is not fully public.

Why it matters

None of this means Lambda’s specific deal is in trouble. Its revenue is real and growing, and its backers include some of the most sophisticated investors in AI. But the pattern the BIS and ECB are both describing, debt secured by a counterparty’s contract rather than by the borrower’s own balance sheet, concentrated in a handful of AI-linked companies, with terms that are not fully public, is exactly what Lambda and Nebius are now both using at billion-dollar scale within weeks of each other. BIS Asia-Pacific representative Zhang Tao told the South China Morning Post that a correction in AI financing could unwind “much faster than previous banking crisis episodes,” largely because so much of it flows through hedge funds and private credit vehicles that carry less regulatory oversight than traditional banks, according to Fortune’s reporting.