Skip to content
NVIDIA's $600B AI Data Center Initiative artwork

NVIDIA's $600B AI Data Center Initiative

Acquired AI

August 17, 2026

In this episode, we analyze NVIDIA's bold backing of a groundbreaking AI data center for OpenAI. Additionally, we look at Anthropic's remarkable revenue doubling.

TopicsTechnology

SPEAKER_1 (0:00)

Nvidia is backing a 4.25 gigawatt Ohio AI data center for OpenAI in a $600 billion compute deal. Anthropic's Q2 revenue hit $11.5 billion. That is a 14x jump just coming up ahead of their IPO. SpaceX just closed the $60 billion cursor acquisition, and that is gonna bring all of that AI coding capability into their GPU empire.

GROK, not the same GROK from SpaceX, but GROQ that Nvidia kind of Acquired a while back, they just raised $350 million at a $3.5 billion valuation, and this is a pivot from chips. Now they're gonna be doing NeoCloud. This is actually an interesting story. OpenAI is publishing emails and iMessages to counter Apple's trade secret lawsuit, so there is tons of drama going on right there. If you have a task that you find yourself doing over and over again that uses AI, and you find yourself kind of wasting time in the glue layer in the middle, copying and pasting, I'd love for you to go check out the AI Box Builder. It's aibox.ai/builder. You can describe what you want, and we'll automatically link the AI models together for you, and we will build a working workflow in minutes. You do not have to know how to code. It just automatically builds a workflow for you that you can use over and over again. So you're building these AI workflows, and you're doing it all visually. It's one of the best ways that I like to use an AI to help me do repeated tasks and save me a ton of time in the middle.

Nvidia is backing a 4.25 gigawatt AI factory. Like I mentioned, this is going to be happening in Ohio, and it's happening for OpenAI.

They've committed their own money to help secure the land, the power in the buildings. OpenAI has locked in $600 billion worth of Nvidia chips through 2030 So obviously, this is Nvidia's biggest customer, and it's in their best interest to finance OpenAI and their ability to keep building these. A lot of people criticize it as the circular financing of the AI, the biggest AI companies, where OpenAI is going to pay them $600 billion to buy their chips, and then Nvidia is just going to straight up back them with their own balance sheet to go and buy all of the land, all the data centers, all of the stuff they need. So this is definitely Nvidia's biggest single customer, but each hardware generation deployed at this site in particular could use 1.5 million Nvidia GPUs, and it can generate about 150 billion to 200 billion dollars in revenue, with multiple refresh cycles expected over 20 years, right? Because when you buy GPUs from Nvidia, when OpenAI is doing this, they're not buying, it's not a one-time deal, but the GPUs get old, their new ones come out, they're going to refresh them. So this is a site that is going to make a lot of money for Nvidia.

Nvidia is guaranteeing lease and power payments for about 4 gigawatts over two decades. This is definitely different from their traditional model, where customers handle all of the infrastructure themselves. Nvidia just gives them the chips, and they're going to go run it. Now they're actually helping customers finance their own infrastructure. This particular Ohio campus has room to expand to 7 gigawatts total, so right now they're kind of getting this all ready to go for about four, but it can expand to 7 gigawatts total. And OpenAI's commitment could grow from 12 gigawatts to 16 gigawatts if the full site is built out. So there is a lot of room to grow here. Anthropic has just reported $11.5 billion in Q2 revenue. That is 14x jump over a year ago where they were at. And this is their first profitable quarter before they file to go IPO. And the interesting thing is I think a lot of people know that they're kind of in this race with OpeningAI to IPO, but there's other people in the race too that are getting ready to have these big kind of blockbuster IPOs. DeepSeek is getting ready to IPO. There's a bunch of other people, and there's a lot of money that needs to be raised right now for AI infrastructure, I think, for Cloud specifically. OpeningAI has a lot of that kind of funding locked in. They've done a lot of deals with big firms like Nvidia, like we were just previously talking about, but this is something that Anthropic is getting into for the first time. A lot of the infrastructure level of this, so they really kind of do need this money sooner than later. But it's incredible just how much they're able to generate. Their Q2 revenue more than doubled. So like I mentioned, it's way up, but Q1 was $4.73 billion, and that is, so now they're more than double in Q2 of this year. So they're 14x from last year, but just from last quarter, they're more than double. So that is more than Anthropic's entire $10 billion in revenue that they made last year. They were able to make that already, which is wild. So a lot of these coding tools are driving the growth there. If you're just kind of looking at what segment is using the usage, you can see that there's a lot of cloud integration into a lot of products. I mean, even Cursor, which SpaceX just acquired, is driving a ton of the usage, which is GitHub alternative, and it has a huge chunk of the market. Anthropic's annualized run rate hit $47 billion in May, which is above OpenAI's $40 billion run rate. So Anthropic is basically making more money than OpenAI is right now, which is kind of crazy. And of course, there's tons of money in the kind of public markets for AI. Specifically, 2026 IPOs have raised $256.4 billion. That's the most that has actually been raised since 2021

8 more minutes of transcript below

Thousands of transcripts fetched by people building searchable podcast archives

Fetch the whole transcript

The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000783948427