Nvidia, OpenAI Make $100 Billion Data Center Deal

0:00With the highest number of young STEM graduates per capita in the EU, Ireland has the people and skills your company needs to succeed here. IDA Ireland, the National Investment Development Agency, can help you find and nurture the people you need to internationalise and thrive. Our talent is just one of the extraordinary benefits Ireland has to offer. Learn more at idaireland.com. Invest in extraordinary.

0:30The thing about AI for business, it may not automatically fit the way your business works. At IBM, we've seen this firsthand. But by embedding AI across HR, IT, and procurement processes, we've reduced costs by millions, slash repetitive tasks, and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business, IBM. Support for the show comes from public.com. If you're actively involved in your portfolio, you probably catch yourself repeating the same actions.

1:07Buying the dip, manually sweeping idle cash, putting on a hedge. On public, you can now create AI agents that handle all these tasks on your behalf. Just describe what you want to do in plain English. Like, if the VIX hits 25, buy a put option on the S &P 500. Or, if my cash balance goes above$20 ,000, move the excess into my direct index. You approve of the workflow and your agent handles the risk. Monitoring the market, watching for your conditions, and executing your strategies exactly as defined. An investing platform driven by your intent, not just your clicks. You can also get full read and write access to your account via the public API.

1:47Go to public.com slash market and fund your account in five minutes or less. That's public.com slash market. Paid for by Public Investing. Brokered services by Open to the Public Investing, Inc., member FINRA and SIPC. Advisory services by Public Advisors, LLC, SEC Registered Advisor. Complete disclosures available at public.com slash disclosures. Bloomberg Audio Studios. Podcasts. Radio. News.

2:23Welcome back to Bloomberg Tech. Let's have a quick check-in on these markets because we are continuing to see pressure on the tech stocks only by about two-tenths of a percent. And as that 100 coming off of what was a five-tenths of a percent increase the previous day, a lot of macro, a lot of digestion this week as we anticipate, of course, the inflation readings and a lot of Fed speak. We're just questioning, though, perhaps some of the valuations in the AI space. Let's move on to see what's happening on this day. With Micron get earnings after the bell were up more than 2%. This is a stock that has increased 40 % in the month of September alone.

2:55But can it live up to the hopes for revenue increases to also be up some 40 % in their fiscal quarter on which they're reporting? Of course, key is the demand for high bandwidth memory and the use within AI accelerators. Talking of AI accelerators, check out NVIDIA. We're off by 1.8 % after yesterday's rally. Remember that$100 billion investment commitment into OpenAI. If OpenAI in return commits to buying GPUs of the future from NVIDIA, the merry-go-round of money continues. Let's break that all down with Bloomberg Semiconductors reporter Ian King. Just going back to that extraordinary$100 billion commitment, getting equity from OpenAI, but only as they bring on AI accelerators, GPUs from NVIDIA and 10 gigawatts eventually in the next few years, Ian.

3:40are we questioning the way in which money is floating right now? That's a large number and it seems excessive even for NVIDIA but then you sort of take a step back and you realize that this is going to happen in an incremental basis and it's going to be put to work only when the actual hardware is put in place and the actual capacity is put in place so yes, it's a massive headline yes, even for NVIDIA that seems huge but if you take it over a number of years then maybe it isn't that excessive. Maybe it's just what is needed for this market. And NVIDIA continues to be a deeply strategic investor, often putting money to work in the likes of CoreWeaver, N-Scale, or long-term demand for their GPUs.

4:23Ian, talking of demand, we're going to get a bit of a fundamental check on it, right, with Micron's earnings after the bell. Yeah, the Micron story is everybody is focused, as you mentioned, on HBM, high bandwidth memory, one of Ed Ludlow's favorite topics. Everybody wants to see this company converting what they know is strong demand into very rapid growth. Everybody wants to see Micron really having a say in the AI world. And that's what people will be focused on because the market overall for other types of memory maybe isn't that strong, maybe isn't recovering as quickly as some had hoped.

5:01$11.15 billion is a consensus, up 44 % in terms of revenue for fiscal fourth quarter. Bloomberg's Ian King, as always, a very busy man. We thank him. Look, as we await earnings, interestingly, anxiety around AI funding just continues to be intensifying. The industry, we understand, will need$2 trillion in combined annual revenue to fund the computing power that's needed by 2030. But their revenue is likely to fall short by a tune of$800 billion. It's all according to Bain's annual global technology report. It was released today, Tuesday. Let's bring in Jay Jacobs for more, U.S. head of equity ETFs over at BlackRock, who you really get an analysis of the insatiable demand of investors into the AI data center story.

5:44But are they starting to question ultimately whether we can meet that insatiable demand, whether funding is there in the longer term? No, I think it's quite the opposite. Investor enthusiasm is only growing. If you look at one of our ETFs, BAI, it's an actively managed AI ETF. it's brought in over$5 billion this year and is now the largest AI ETF in the United States. And a lot of the exposure that it is providing to investors is around that digital infrastructure layer, the hardware producers in the semiconductor space, the digital infrastructure owners, some of the data owners that are becoming really valuable in this explosion of AI models.

6:17And so there's a ton of investor enthusiasm and I think there's a lot of expectations that revenues will match the expenditures if not exceed them in the next several years. But how much should we get comfortable with companies like NVIDIA basically plowing their own revenues of today into hopes of revenue of tomorrow, basically funding their end clients? I think it's beyond comfort. This is a necessity for some of the leading technology companies. They have a tremendous moat because they have low cost of capital. They have tremendous access to dollars. And they're able to really kind of leverage their own strengths, which is if you look at a semiconductor company, owning GPUs, owning this manufacturing process is really valuable in the semiconductor space.

6:59That's the shortfall that's needed by large language model developers. So this investment across the industry makes a ton of sense. So going back to your iShares AI Innovation and Tech actively managed ETF, how much are clients begging you for more diversification when it comes to the AI trade? Because we have put so much stock into a few names. There's a lot of ask to go beyond the Mag7. I think people understand the Mag7 story. They understand why these companies are so successful in the AI space and why they're leading it. But they do want to know what's beyond that. What are the other semiconductor companies?

7:30What are some of the data owners? What are some of the companies that aren't necessarily a consensus view as an AI company, but are involved in the space? And so that's where an actively managed fund, I think, can provide a lot of value to investors is by, yes, it's going to have MAC7 exposure. But, you know, beyond two thirds of it is going to look beyond it and provide kind of the next names that are going to rise in this ecosystem. And are those names rising in the United States? We've also just seen President Trump alongside Jensen Huang and many others go to the United Kingdom to talk up an investment there.

7:59Are we seeing more names being added to the ETFs? They're coming from Europe, coming from Asia. It's a global theme, but predominantly it is being led out of the United States today. And that's where we've seen some of these kind of new up-and-coming AI risers largely coming from. We still have a really robust startup environment and venture capital environment around AI. So the U.S. is very much at the center of this ecosystem, but it is a global theme. So the members of the ETF are largely U.S. Are the investors in that ETF getting ever any more global, or is it predominantly U.S. investors that want to stick with this theme?

8:29It's both. We certainly have seen a ton of global appetite, but some of the biggest investors today are coming from the United States. I think AI is not surprising anyone. Everyone knows globally this is the biggest area right now. They want to have access to the best investment tools to get exposure to it. Our ETF has really risen as the number one tool in the US for AI exposure. What about other parts, like the energy side of the equation, other areas that eventually will do well? We're always talking about healthcare predominantly being a user of generative AI and therefore more productive?

8:59I would say the energy space is coming up more and more with clients as they look to the AI trade beyond technology. Kind of what's the next shoe to drop? What's the area of opportunity? We've seen a ton of interest around different energy sources to be able to power the AI trade, whether that's nuclear, whether it's fossil fuels, whether it's renewables, really looking across the entire energy ecosystem. It's a way of looking at utilities from a more forward-looking lens. If you take the utility sector, but say, who's really going to grow here? Who really has this great 10-year trajectory because of AI?

9:31It's a more thematic cut of that space. We've just been coming out of the United Nations General Assembly. It does feel as though fossil fuels very much have to be put back on the map by President Trump, at least. Is that something you're seeing echoed by investors? They want to be recommitting once again after years of perhaps rebalancing, moving towards renewables, looking at cleaner forms of energy? It's across the entire energy value chain. I think you have to look at there are going to be energy fuels that are going to be really necessary. There's going to be energy producers that are really necessary, and there's energy distribution that's going to be really necessary in this AI theme.

10:01And so I think people are looking very broadly at where is this going to come from. Every now and then, the market goes through moments of anxiety, whether it's about the fact that maybe we've got this revenue mismatch when it comes to compute funding, whether it's about actual productivity gains with the MIT report that came out last month. Are you seeing any anxiety pullback in actual fund flows wanting to go into AI more broadly? Like I said, our AI fund flows have been accelerating recently, so we're not seeing a lot of anxiety. Look, this is a long-term theme. This is a bet for the next 5, 10, and beyond years.

10:33I'm sure within that time frame, there will be bouts of anxiety. There will be questions around what the valuation should be. But if you take a long-term view of artificial intelligence, we believe this is going to be one of the most powerful drivers of the markets for the next decade and beyond. Okay. And more broadly, when we're thinking about fundamentals, how much do earnings matter for your portfolio managers, but more broadly for the investor sentiment too? So they matter, but when you're looking over the long term, it's more about what does the earnings say about the direction of travel rather than what's happening today.

11:02There's not a ton of concern about what is your revenue today if you're a large language model developer. What we prefer to see is, have you gotten more paying subscribers? Have you developed the best model? Are you starting to see more penetration in the corporate world where you're going to have more ability to spend on these technologies. The direction of travel is what we're really looking at in the earnings. Jay Jacobs, it's always great to have you here. Thanks for coming in the studio. BlackRock's Jay Jacobs there. Meanwhile, I want to talk other areas of AI. Now, the application of it, Snowflake, alongside industry leaders and key partners, just unveiled a new initiative aiming to set a new industry standard for AI interoperability.

11:41They call it the Open Semantic Interchange, and it will be open-sourced in an effort to remove what they call data islands and accelerate AI for all. Sridhar Ramaswamy, I'm pleased to say, is with our Snowflake CEO. Set the context of the problem here. Sridhar, how difficult is it to get access to all data out there at the moment to make generative AI become this reality? Well, great to be here, Caroline. AI has a lot of promise for making data very broadly available to everyone in the enterprise, including business users, using natural language, which all of us like doing. Well, the problem often is that the meaning behind the data is in multiple places.

12:24It can be in databases. It can be in business intelligent tools. This just makes developing AI solutions on top of this data very hard. And what we announced today with a consortium of industry folks is an open standard for exchanging this semantic information. And what that means is this is going to enhance interoperability between different folks, between Salesforce, DBT Labs, BlackRock, that we would like to see happen. It can help accelerate value that is going to be gotten from AI. It also just simplifies things for everybody that's involved. And that's why we are super excited about working with these giant industry players to announce this new interchange format.

13:07Yeah, a universal translator is kind of a nice way of talking about it with Snowflake, the Salesforce, DBT Labs, BlackRock, ThoughtSpot. We're geeky on this program. Can you tell us about how difficult the technology exchange is going to be here? What actually you need to do beneath the surface? Yeah, it's a standard. What this means is that people that want to be part of the standard or want to confirm to the standard just are going to make it easy to send and receive this semantic information. This is, you know, this is business stuff, stuff like how is revenue defined, how is profit defined, how is EBITDA defined, and so on.

13:45So we intentionally wanted it to be a light lift for each of these players. But what happens collectively is that all of a sudden, this semantic information is free to move about. It just makes value creation, especially with AI, just a whole lot faster. And it's a big step forward for the industry. And we expect more and more folks in the data industry to be a part of this standard. Interesting. So that's for the data industry. Going to the large language model industry in and of itself, Anthropic seem to have seen this issue with data and interoperability as well. They've been seeing silos. They introduced something called model context protocol.

14:23That's, again, open standard. Shrida, how is that different? This is one layer below that. What a model context protocol does is it creates a standard way, for example, to talk to an AI agent. But on the other hand, even to create the AI agent, you need to make sense of the underlying data. it can be the case that you have data sitting in one place while the definition for what the data means is in a different place. This makes it easier to create agents, which can then take advantage of protocols like model control protocol, MCP that you mentioned, to make agent interoperability work a whole lot better.

15:03So you need all of these layers for AI value to get created. And we need the agents to get better for the productivity to really result. Should I ask you, like after months of worry, perhaps about MIT reports saying that ultimately the productivity gains aren't being seen yet, what are you seeing with your clients using your own offerings and products? Yeah, we are seeing a lot of productivity gain in specific areas. And I can speak even personally to examples like my sales data is just a whole lot easier for me to get to with Snowflake Intelligence. BlackRock, one of the customers in the OSI standard, also is using our AI products to create what's called a Customer 360.

15:45So when you call BlackRock, the person on the other side has access to all of the relevant information about you. It's things like that that are creating value. And what's unique about Snowflake is we let our customers create AI products without a ton of investment on their part. We make it much easier to iterate and create these products and use them and see for themselves what's creating value before turning on the spigot and having these products exposed more and more. And in areas like coding agents, the productivity gains that we are seeing is truly remarkable. And I think that's an area where you will continue to see a ton of adoption across the board in enterprises.

16:28You know, funnily or strangely, it turns out that coding is one of the most interesting and important applications of language models. Shudar, we've seen the adoption of your products. It's shown up in your earnings. People have seen really the growth that you're talking about. For that, you need sales and marketing people to go out and sell your products. But that's where your talent addition has been. But I go to the news of the day, for example, around H-1B visas. How much is that becoming an issue for you with talent? Is that something that you're worried about, access to the right people here in the United States?

16:59Yeah, it is an issue. Roughly less than 10 % of our workforce is on H-1B visas. You know, I'm an immigrant, and I think there are a ton of folks like me that come in, create massive amounts of value, and then become proud citizens. We don't see it as an immediate problem, but we do think that tech workers from around the world have a lot to add to the tech industry. we are watching and working as closely with the administration as we can to make sure that this is more of a win-win for everybody involved. Trita Ramaswamy, so thank you for that answer. And talking us through your announcement today on AI interoperability, we appreciate it, the CEO of Snowflake.

17:44Coming up, Michelle Guider of Purdue's Kraut Institute for Tech Diplomacy. We're going to break down the state of U.S.-China relations. This is Bloomberg Tech. with the highest number of young stem graduates per capita in the eu ireland has the people and skills your company needs to succeed here ida ireland the national investment development agency can help you find and nurture the people you need to internationalize and thrive our talent is just one of the extraordinary benefits ireland has to offer learn more at idaireland.com invest in extraordinary

18:24Support for the show comes from public.com. If you're actively involved in your portfolio, you probably catch yourself repeating the same actions. Buying the dip, manually sweeping idle cash, putting on a hedge. On public, you can now create AI agents that handle all these tasks on your behalf. Just describe what you want to do in plain English. Like, if the VIX hits 25, buy a put option on the S &P 500. or if my cash balance goes above$20 ,000, move the excess into my direct index. You approve the workflow and your agent handles the rest. Monitoring the market, watching for your conditions and executing your strategies exactly as defined.

19:03An investing platform driven by your intent, not just your clicks. You can also get full read and write access to your account via the public API. Go to public.com slash market and fund your account in five minutes or less. That's public.com slash market. Paid for by Public Investing. Brokered services by Open to the Public Investing, Inc., Member FINRA, and SIPC. Advisory services by Public Advisors, LLC, SEC Registered Advisor. Complete disclosures available at public.com slash disclosures. The thing about AI for business, it may not automatically fit the way your business works. At IBM, we've seen this firsthand.

19:42But by embedding AI across HR, IT, and procurement processes, we've reduced costs by millions, slash repetitive tasks, and freed thousands of hours for strategic work. Now we're helping companies get smarter by putting AI where it actually pays off, deep in the work that moves the business. Let's create smarter business, IBM.

20:10Welcome back to Bloomberg Tech. A quick check on these markets, which are pretty sanguine at the moment as we have a big week ahead in terms of inflation data and Fed speed. NASDAQ 100 just down about two-tenths of a percent, having been on a record-high streak across most of the benchmarks. We pause for breath on equities more broadly. NVIDIA in the red, which, of course, tugs down in large part the rest of the index. But we want to set a global context here because the movement being made geopolitically at what's in the eye of the investor right now. We've just had the United Nations General Assembly speech by the United States president.

20:42Really big claims and blasting the United Nations more broadly. But all of this in a context of geopolitical relationships, perhaps the US and China, for example. We have been understanding they're nearing an agreement to hive off the US operations and social media platform TikTok to a consortium that includes software giant Oracle. We want to just give you the global context right now with Bloomberg senior tech editor, Mike Shepard standing by. You've just been sat tuned in to President Trump where he talked about a whole wealth of issues that he sees right now. Any of them catch your attention?

21:15And certainly for our context, Carol, on this program, what really caught my ear was what he was talking about in terms of investment and what he's been trying to attract to the U.S. When you think about all the announcements that we've seen related to artificial intelligence, And more recently, as you just alluded to, this push to try to reach a deal with China on TikTok. The president was talking about$17 trillion in announced investment pledges so far this year. That number bears some skepticism, but there is some kernel of truth in it. Of course, when you towed up the pledges that we have seen from NVIDIA, from OpenAI, from SoftBank and others to build plants here in the U.S.

21:54And TSMC, we can't leave them out either. And all that fits into the picture that the president is trying to portray of the U.S. on the move during this address to the U.N. General Assembly. He talked about this golden age of America, and he defended his tariff program as well and some of the trade agreements that he's reached with the U.K., the E.U., and other nations. And he even talked about trying to meet with Brazil's leader next week to try to sort out some of the trade and other differences with that trading partner as well. So there's a lot happening here on the international trade and dealmaking front.

22:29And it had a market reaction, certainly in the Brazilian real gaining in the session, Mike. What's interesting is he did go on to chastise China when it comes to the Russian relationship that they hold. Anything that could put at risk any sort of ultimate deal around TikTok? Well, it's a great question because, you know, for the president, the TikTok deal is of paramount importance. He's talked about it for more than a year. It was something that came up on the campaign trail. It's something that he credits with helping him in his comeback victory to the White House. And yet it has been a very difficult victory for him to bring home, in part because China is reluctant to sell this precious asset from Beijing-based ByteDance Limited.

23:14it. China is reluctant also to see ByteDance let go of in any way, shape or form the algorithm that powers TikTok and that has really been the secret sauce to the app and its appeal to young users. Now, the question in this deal that is taking shape is perhaps there could be a way for a copy to be handed to the new U.S. venture that would be led by American investors and have six of seven board seats occupied by Americans. Now, would China go along with that? Oracle would have a role with the new venture in retraining this new model. But will China go along? We have not heard enough from China yet about whether they will approve this deal in full.

23:57On Friday, we heard Donald Trump talk about how Xi Jinping, his counterpart hours earlier in a conversation they had had on Friday, really blessing the deal. But China's foreign ministry was far more circumspect in its statements. So we'll have to see where this goes and whether some of that leverage on Russian oil purchases may factor in into the conversation here. Remurg's Mike Shepard, thorough wrap up. Thank you very much indeed. Now let's discuss this more broadly with Michelle Geiner. She is the CEO of the Krak Institute for Tech Diplomacy at Purdue, used to serve as Assistant Secretary of State for Global Public Affairs under the first Trump administration.