NVIDIA is about to spend $750 billion on AI. Critics are calling it a bubble
NPR's Ayesha Rascoe asks Bloomberg AI infrastructure reporter Dina Bass about concerns over circular financing in the AI sector.
Transcript
AYESHA RASCOE, HOST:
Seven hundred and fifty billion dollars. That's what reporters from The Information, The Wall Street Journal and Bloomberg say chipmaking giant Nvidia is thinking of spending on data centers to strengthen its supply chain. And it's not just the price tag that's raising eyebrows. Dina Bass is one of the reporters with the story. She covers AI infrastructure for Bloomberg News, and she joins us now. Welcome to the program.
DINA BASS: Thanks for having me.
RASCOE: What other companies are involved in these talks, and how big are they as players in the AI sector?
BASS: So let's start with Nvidia. They are the largest by far. They have these graphics chips that, you know, years ago they used and still do for video games and things like that. It turned out to be really important for training AI models. So it's kind of the workhorse chip in the AI space. And they are, you know, kind of using their funding and their money, their investments and their customer relationships to continue to juice demand for that.
And so, you know, in the past month alone, Nvidia has unveiled a partnership with South Korean conglomerate SK Group, which means those companies will be doing more than 500 billion in business with each other. And SK owns a company called SK Hynix, which makes the memory chips whose products are critical for NVIDIA's AI chips. And also, in the last couple of days, we've reported - our sources tell us that Nvidia is in talks with OpenAI to help fund both a lease on a massive data center in Ohio and also putting those - the Nvidia chips in there and, you know, possibly other data centers.
RASCOE: So who are the people kind of calling out this kind of circular financing? And explain it for the layperson - why this is circular financing.
BASS: I mean, you've had, you know, Michael Burry, who's kind of famous for calling out, you know, what ended up being the - sort of the mortgage crisis. You know, a lot of reporters, a - you know, investors are looking at this as well. And the concern here ends up being - it's not just Nvidia. Nvidia's rival AMD is investing 5 billion in OpenAI's rival Anthropic. And Anthropic is then going to use AMD chips. Anthropic also has taken huge - billions of dollars of investments from Google and Amazon and agreed to use billions of dollars of their chips. And in some cases, what we're literally seeing amounts to a company giving another company money to buy its product.
There's also concern that with all of the interlocking relationships, any impairment to one company risks kind of spreading a contagion throughout the whole, you know, ecosystem of all of these companies. And, you know, one of the wrinkles - you know, you have to consider that an increasing number of financial firms, so your Apollos, your Blackstones and many more, are now helping finance the, you know, billions of dollars of these large deals for data centers and chips as well.
RASCOE: Yeah. So then you have financial companies also exposed.
BASS: Exactly. You know, in a nutshell, the question becomes, if there's a point where the music stops on the AI expansion, you know, who ends up left without a chair? Who's left holding the bag?
RASCOE: Well, how has Nvidia responded to this criticism?
BASS: Look, Jensen Huang hates the term circular financing. I have personally, as have other reporters, asked him about it. You know, when I asked Jensen this question a couple months back on the day he signed an investment and sales deal with AI cloud CoreWeave, he felt that, you know, they are investing. They are juicing sales. But they are not, in his view, literally funding their own sales.
RASCOE: This isn't just a concern about shareholders or workers at these companies. It's about how important these companies are to the entire U..S economy. And I should note that Anthropic, Google, Amazon and Apollo are NPR funders, just to make a note of that. But if we are seeing a bubble, like the critics say, what would be the best parallel from recent memory? Like, the dotcom bust or the meltdown over mortgage-backed securities?
BASS: I covered the dotcom bubble very early in my career at Bloomberg, and there were a lot of ideas and companies that came about in that era that ended up being right. People ended up doing exactly those things. They are now major parts of, you know, the technology that we use every day - things like grocery delivery, things like location-based services, some early mobile things. But the companies behind them did not always end up being the ones that led the way for that.
And you end up with a similar - you know, potentially similar situation in the AI era. People in this space will tell you AI is real. These are ideas that are happening. These are legitimate and important technologies that are going to change significant things about the way our world operates. But that doesn't necessarily translate into the notion that the companies that are currently leading each of those spaces - are currently raising a lot of money to compete in each of those spaces - those may not be the companies who, five or 10 years from now, are, you know, the ones that end up benefiting from them.
RASCOE: That's Dina Bass of Bloomberg News. Thank you so much for being with us.
BASS: Thank you for having me.
(SOUNDBITE OF MUSIC)
Copyright © 2026 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.
Accuracy and availability of NPR transcripts may vary. Transcript text may be revised to correct errors or match updates to audio. Audio on npr.org may be edited after its original broadcast or publication. The authoritative record of NPR’s programming is the audio record.
Read the full story at NPR ↗
NVIDIA is in talks to invest substantial capital in data center infrastructure, with reporting indicating planned spending around $750 billion. The company has established partnerships involving significant purchasing commitments—SK Group committing over $500 billion in business with NVIDIA, and discussions with OpenAI regarding data center leases and chip deployment. Similar patterns exist across the AI sector: AMD is investing in Anthropic, which plans to use AMD chips; Anthropic receives funding from Google and Amazon while committing to use their chips. Critics have noted these interconnected arrangements where companies essentially fund other companies to purchase their products. Financial firms including Apollo and Blackstone now finance many of these large infrastructure deals, creating broader systemic exposure. NVIDIA leadership rejects the "circular financing" framing, characterizing their activities as legitimate market investment. Analysts note that while AI itself represents genuine technological development with real applications, historical precedent suggests current industry leaders may not necessarily be the long-term beneficiaries of the sector's evolution.
Read the full story at NPR ↗
NPR's Ayesha Rascoe asks Bloomberg AI infrastructure reporter Dina Bass about concerns over circular financing in the AI sector.
Transcript
AYESHA RASCOE, HOST:
Seven hundred and fifty billion dollars. That's what reporters from The Information, The Wall Street Journal and Bloomberg say chipmaking giant Nvidia is thinking of spending on data centers to strengthen its supply chain. And it's not just the price tag that's raising eyebrows. Dina Bass is one of the reporters with the story. She covers AI infrastructure for Bloomberg News, and she joins us now. Welcome to the program.
DINA BASS: Thanks for having me.
RASCOE: What other companies are involved in these talks, and how big are they as players in the AI sector?
BASS: So let's start with Nvidia. They are the largest by far. They have these graphics chips that, you know, years ago they used and still do for video games and things like that. It turned out to be really important for training AI models. So it's kind of the workhorse chip in the AI space. And they are, you know, kind of using their funding and their money, their investments and their customer relationships to continue to juice demand for that.
And so, you know, in the past month alone, Nvidia has unveiled a partnership with South Korean conglomerate SK Group, which means those companies will be doing more than 500 billion in business with each other. And SK owns a company called SK Hynix, which makes the memory chips whose products are critical for NVIDIA's AI chips. And also, in the last couple of days, we've reported - our sources tell us that Nvidia is in talks with OpenAI to help fund both a lease on a massive data center in Ohio and also putting those - the Nvidia chips in there and, you know, possibly other data centers.
RASCOE: So who are the people kind of calling out this kind of circular financing? And explain it for the layperson - why this is circular financing.
BASS: I mean, you've had, you know, Michael Burry, who's kind of famous for calling out, you know, what ended up being the - sort of the mortgage crisis. You know, a lot of reporters, a - you know, investors are looking at this as well. And the concern here ends up being - it's not just Nvidia. Nvidia's rival AMD is investing 5 billion in OpenAI's rival Anthropic. And Anthropic is then going to use AMD chips. Anthropic also has taken huge - billions of dollars of investments from Google and Amazon and agreed to use billions of dollars of their chips. And in some cases, what we're literally seeing amounts to a company giving another company money to buy its product.
There's also concern that with all of the interlocking relationships, any impairment to one company risks kind of spreading a contagion throughout the whole, you know, ecosystem of all of these companies. And, you know, one of the wrinkles - you know, you have to consider that an increasing number of financial firms, so your Apollos, your Blackstones and many more, are now helping finance the, you know, billions of dollars of these large deals for data centers and chips as well.
RASCOE: Yeah. So then you have financial companies also exposed.
BASS: Exactly. You know, in a nutshell, the question becomes, if there's a point where the music stops on the AI expansion, you know, who ends up left without a chair? Who's left holding the bag?
RASCOE: Well, how has Nvidia responded to this criticism?
BASS: Look, Jensen Huang hates the term circular financing. I have personally, as have other reporters, asked him about it. You know, when I asked Jensen this question a couple months back on the day he signed an investment and sales deal with AI cloud CoreWeave, he felt that, you know, they are investing. They are juicing sales. But they are not, in his view, literally funding their own sales.
RASCOE: This isn't just a concern about shareholders or workers at these companies. It's about how important these companies are to the entire U..S economy. And I should note that Anthropic, Google, Amazon and Apollo are NPR funders, just to make a note of that. But if we are seeing a bubble, like the critics say, what would be the best parallel from recent memory? Like, the dotcom bust or the meltdown over mortgage-backed securities?
BASS: I covered the dotcom bubble very early in my career at Bloomberg, and there were a lot of ideas and companies that came about in that era that ended up being right. People ended up doing exactly those things. They are now major parts of, you know, the technology that we use every day - things like grocery delivery, things like location-based services, some early mobile things. But the companies behind them did not always end up being the ones that led the way for that.
And you end up with a similar - you know, potentially similar situation in the AI era. People in this space will tell you AI is real. These are ideas that are happening. These are legitimate and important technologies that are going to change significant things about the way our world operates. But that doesn't necessarily translate into the notion that the companies that are currently leading each of those spaces - are currently raising a lot of money to compete in each of those spaces - those may not be the companies who, five or 10 years from now, are, you know, the ones that end up benefiting from them.
RASCOE: That's Dina Bass of Bloomberg News. Thank you so much for being with us.
BASS: Thank you for having me.
(SOUNDBITE OF MUSIC)
Copyright © 2026 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.
Accuracy and availability of NPR transcripts may vary. Transcript text may be revised to correct errors or match updates to audio. Audio on npr.org may be edited after its original broadcast or publication. The authoritative record of NPR’s programming is the audio record.
Read the full story at NPR ↗
NVIDIA is the largest chipmaker for AI, producing graphics chips essential for training AI models NVIDIA has partnerships with SK Group valued at over $500 billion in business commitments NVIDIA is in discussions with OpenAI regarding funding for a data center lease in Ohio and chip deployment AMD is investing $5 billion in Anthropic, which has agreed to use AMD chips Anthropic has received billions in investments from Google and Amazon while committing to use their chips Apollo, Blackstone and other financial firms are financing major data center and chip deals in the AI sector Critics including investor Michael Burry characterize these arrangements as "circular financing" The interlocking relationships create systemic risk where problems at one company could spread contagion throughout the ecosystem NVIDIA's leadership rejects the term "circular financing" and contends they are making legitimate investments rather than funding their own sales AI represents genuine and transformative technology with legitimate applications Current market leaders in AI may not be the companies that ultimately benefit from the sector five to ten years forward, similar to patterns from the dotcom era
Read the full story at NPR ↗
- NVIDIA is reportedly planning to spend approximately $750 billion on data centers to support AI chip production and supply chain control
- Critics including investor Michael Burry have identified "circular financing" patterns where chipmakers fund AI companies that then purchase their products
- Financial institutions including Apollo and Blackstone are increasingly financing large data center and chip deals, expanding systemic risk exposure
- NVIDIA executives dispute the "circular financing" characterization, arguing they are making legitimate investments rather than funding their own sales
- Tech industry leaders maintain AI development is genuine and transformative, though historical parallels suggest current market leaders may not ultimately dominate the sector