Nvidia Is Expanding Its Role in the AI Craze

[00:00:00] [SPEAKER_00]: First, the bad news. SAP Business AI won't generate amusing holiday cards, but it will

[00:00:05] [SPEAKER_00]: personalize career paths for your people and let you know which suppliers are best so

[00:00:09] [SPEAKER_00]: you can be ready for the next opportunity. Revolutionary technology, real world results.

[00:00:13] [SPEAKER_00]: That's SAP Business AI.

[00:00:18] [SPEAKER_03]: Welcome to Tech News Briefing, its Wednesday, September 4th. I'm Zoe Thomas for The Wall Street Journal.

[00:00:25] [SPEAKER_03]: Technology has come for so many things in our lives. Now we could help deal with the stink

[00:00:31] [SPEAKER_03]: from your kitchen trash can. But for a hefty price, we'll tell you more.

[00:00:37] [SPEAKER_03]: And then, Nvidia already dominates the market for chips powering the artificial intelligence

[00:00:43] [SPEAKER_03]: boom. Now the company is playing a growing role in designing AI data centers. Our reporter

[00:00:49] [SPEAKER_03]: is a fish will join us to explain what the company is doing and how it's trying to

[00:00:54] [SPEAKER_03]: offset regulatory risks.

[00:01:04] [SPEAKER_03]: But first, in the summer months, food scraps in your garbage might really smell. You

[00:01:10] [SPEAKER_03]: could try cutting down on your food waste or like our personal technology bureau cheap Wilson

[00:01:16] [SPEAKER_03]: Rothman, you could try a new device called the Mill Food Recycler. For $1,000 or a $30

[00:01:24] [SPEAKER_03]: month rental plan with a buyout option, this trash can can fight the garbage stink and

[00:01:30] [SPEAKER_03]: produce what makes it to landfills. So is it worth it? Wilson is here with us, so Wilson

[00:01:36] [SPEAKER_03]: tell us about the Mill Food Recycler and how does it work?

[00:01:41] [SPEAKER_01]: It looks like a trash can and you throw your food waste in there, the scraps from your

[00:01:45] [SPEAKER_01]: kitchen are the leftovers from your plate and it heats them up, churns them around for

[00:01:51] [SPEAKER_01]: three to four hours and outcome something that looks like coffee grounds but smells more

[00:01:56] [SPEAKER_03]: like dog treats. Okay, so you said you can put your food waste in it, can you tell us a little

[00:02:01] [SPEAKER_01]: bit more about what exactly can and can't go in? Obviously any fruit peels or like peach pits

[00:02:08] [SPEAKER_01]: and things like that that come when you're just snacking or whatever, those can all go in.

[00:02:12] [SPEAKER_01]: Anything you're chopping on your cutting board usually is going to be able to go in,

[00:02:16] [SPEAKER_01]: pizza crust that the kids don't want to eat, the stuff like that. You get into a little

[00:02:20] [SPEAKER_01]: more trouble when you get into bones, like you can put chicken bones in but you can't put

[00:02:24] [SPEAKER_01]: like big beef bones like that or like a leg of lamb, oyster shells are not cool to put in there.

[00:02:30] [SPEAKER_01]: You don't want to dump a lot of liquid or a lot of oil in there or you don't want to put an

[00:02:35] [SPEAKER_01]: overload of sugary stuff like cakes or cookies but the app is really super useful. It's got

[00:02:42] [SPEAKER_01]: a really fun little library where you can look up any food in the world and see whether

[00:02:46] [SPEAKER_03]: it can go in or not. So describe what this looks like a little bit for us is it just really a

[00:02:52] [SPEAKER_01]: smart trash can sitting in your kitchen corner. It just looks like a trash can it's like a nice

[00:02:57] [SPEAKER_01]: little white thing that you step on with your foot and the lid pops open and you dump your stuff

[00:03:02] [SPEAKER_01]: in. When you look inside you know something's up because it doesn't go all the way down.

[00:03:07] [SPEAKER_01]: It ends about a third of the way down and you see these two kind of plastic little propellers in

[00:03:13] [SPEAKER_01]: there that kind of churn the food around when it's doing it cycle. So you definitely know once you

[00:03:19] [SPEAKER_01]: are at your average trash can. And you mentioned the app what does that do? So you know like other

[00:03:26] [SPEAKER_01]: smart devices of our times you can use the app to turn it on and stop it and do other things like

[00:03:31] [SPEAKER_01]: it weighs all the scraps that I add essentially. So what I know is after two months I have turned

[00:03:37] [SPEAKER_01]: 74 pounds of my family's food waste into 14 pounds of these dried up grounds.

[00:03:45] [SPEAKER_03]: Okay so what happens with all those grounds you've created? The company will take them off your hands

[00:03:52] [SPEAKER_01]: they actually will provide for a $10 a month subscription they will provide boxes with little

[00:03:58] [SPEAKER_01]: bags inside where you can just dump all your grounds in the bags if the bag close the box and mail it back.

[00:04:04] [SPEAKER_01]: That is convenient although I don't like the price especially because

[00:04:09] [SPEAKER_01]: I myself my family has not produced enough to use it monthly. So what the company says you can do

[00:04:17] [SPEAKER_01]: is if you are a gardener you can use it as fertilizer you can use it to fertilize your lawn

[00:04:23] [SPEAKER_01]: it is not compost but you can also dump it into a compost pile and make it compost. But when it

[00:04:31] [SPEAKER_01]: comes out of there it's really something you don't want to use as kind of soil. I mean the point

[00:04:35] [SPEAKER_03]: of the mill and I guess even composting at home is to reduce how much food waste goes into the

[00:04:41] [SPEAKER_01]: trash. How big of a problem is this in the US? It's huge it makes up about a quarter of what goes

[00:04:49] [SPEAKER_01]: into a landfill and that is the single largest component of what goes into landfills and the bigger

[00:04:55] [SPEAKER_01]: issue is we don't really have anything better to do with it than put it in a landfill. So 60%

[00:05:02] [SPEAKER_01]: of our food waste goes to the landfill right now. All right that was Wilson Rothman the Wall Street

[00:05:09] [SPEAKER_03]: Journal's personal technology bureau chief coming up will tell you what Nvidia is doing to become a

[00:05:15] [SPEAKER_03]: one-stop shop for the key elements in the data centers where AI is produced. That's after the break.

[00:05:25] [SPEAKER_00]: First the bad news. SAP Business AI won't help you generate cubist versions of your family's

[00:05:31] [SPEAKER_00]: holiday photos. But it will help you understand which supplier is best to help you roll out your

[00:05:36] [SPEAKER_00]: plant-based packaging in Southeast Asia or identify the training your junior project manager needs

[00:05:40] [SPEAKER_00]: to rise up the ranks and automate repetitive tasks while you focus on big innovations.

[00:05:46] [SPEAKER_00]: So you can be ready for the next opportunity. Revolutionary technology real world results.

[00:05:52] [SPEAKER_03]: That's SAP Business AI. Nvidia's chips are essential to powering tools like OpenAI's

[00:06:08] [SPEAKER_03]: chatGPT. The company has an estimated 80% of the AI chip market share. But its CEO,

[00:06:16] [SPEAKER_03]: Jensen Huang is broadening Nvidia's focus. He's talked about becoming a one-stop shop for what

[00:06:22] [SPEAKER_03]: he calls AI factories. The data centers where these powerful tools are created and deployed.

[00:06:29] [SPEAKER_03]: Here to tell us more about these efforts is our reporter Ace of Fish. So Aceza how is Nvidia

[00:06:34] [SPEAKER_02]: going about this expansion? The company has built up its portfolio of services and products,

[00:06:42] [SPEAKER_02]: unlike hardware chips in the data centers for a number of years. They made some acquisitions.

[00:06:47] [SPEAKER_02]: Like they bought a company called Melonox for about $7 billion a few years ago

[00:06:51] [SPEAKER_02]: that is centered on networking with sort of wires and other infrastructure that connects these things

[00:06:56] [SPEAKER_02]: together very important for doing things quickly in AI. They've done other things like they've built

[00:07:01] [SPEAKER_02]: up their capabilities in designing the infrastructure itself. The cabinets, the servers, the racks,

[00:07:07] [SPEAKER_02]: the things inside of data centers that people put the computing power into.

[00:07:13] [SPEAKER_03]: And how do customers feel about Nvidia trying to take this approach?

[00:07:17] [SPEAKER_02]: I mean a lot of customers like it because it means that if one company in video can do more

[00:07:22] [SPEAKER_02]: within the data center essentially they can basically supply more of a turnkey solution.

[00:07:27] [SPEAKER_02]: And a lot of people like that in the context of the current AI boom where speed really matters.

[00:07:32] [SPEAKER_02]: Getting something in there quickly is paramount and Nvidia because it offers so many things

[00:07:37] [SPEAKER_02]: within the data center can typically do things very quickly or you can do more faster if you

[00:07:44] [SPEAKER_02]: choose Nvidia than perhaps one of its competitors. That's contrary to how things might work in a

[00:07:49] [SPEAKER_02]: more mature market where people are more concerned about what cost less, what way can I do things

[00:07:54] [SPEAKER_02]: are much more tailored to my own needs and don't necessarily use everything from one supplier.

[00:08:01] [SPEAKER_02]: So we're at a moment right now where Nvidia has this opportunity to play more of a central role

[00:08:05] [SPEAKER_02]: in all of the things that go into the data center beyond just its chips.

[00:08:10] [SPEAKER_03]: And how are rivals responding to Nvidia's plans for expansion? A lot of rivals are responding

[00:08:15] [SPEAKER_02]: by essentially trying to do the same thing offering more services and products and things like that

[00:08:21] [SPEAKER_02]: within the data center so you have companies like AMD which recently bought a company called

[00:08:27] [SPEAKER_02]: ZT Systems or agreed to buy that company for almost $5 billion and they're going to strip

[00:08:33] [SPEAKER_02]: out the data center design portion of that business. They basically help people figure out how

[00:08:39] [SPEAKER_02]: to set up their data center infrastructure, how to build it essentially.

[00:08:44] [SPEAKER_03]: So what are some of the challenges to this approach?

[00:08:47] [SPEAKER_02]: Well the risk for Nvidia really is that the market does mature and there is more competition

[00:08:53] [SPEAKER_02]: in these individual areas whether it be networking, data center design, any other area where

[00:08:58] [SPEAKER_02]: Nvidia's tried to make a broader push. If those areas become individually competitive

[00:09:03] [SPEAKER_02]: and people aren't moving quite as fast they don't have this need to get their computing for AI

[00:09:08] [SPEAKER_02]: setup yesterday Nvidia is in less of a powerful position. There's no indication that's happening yet

[00:09:15] [SPEAKER_02]: analysts I've talked to you say that's a risk looking forward for Nvidia and pursuing this kind

[00:09:20] [SPEAKER_02]: of vertical strategy. And what about regulatory concerns? Those concerns are in their

[00:09:26] [SPEAKER_02]: infancy but they could play a big role in the way that Nvidia's business shapes up looking for

[00:09:32] [SPEAKER_02]: regulators have started to look at the company carefully. They've rated its offices in France

[00:09:36] [SPEAKER_02]: last year should be said that there have been no charges brought against the company but

[00:09:41] [SPEAKER_02]: Nvidia has a dominant market share in the provision of AI chips and there is growing concern among

[00:09:48] [SPEAKER_02]: regulators around the world that it is getting to a point where it should be investigated at least

[00:09:55] [SPEAKER_02]: on antitrust grounds. Nobody knows the stage where that's going to go if it's going to go anywhere

[00:09:59] [SPEAKER_02]: but if Nvidia's out here building up a bigger portfolio of businesses and types of products

[00:10:05] [SPEAKER_02]: is using to help people create these AI factories if you will that could potentially make it more

[00:10:11] [SPEAKER_02]: vulnerable to those kinds of concerns or make those concerns more amplified. And what is Nvidia

[00:10:17] [SPEAKER_02]: doing to offset these concerns? You know the company has been very clear that it's not going

[00:10:23] [SPEAKER_02]: to do everything and it doesn't really make sense financially for the company to do everything because

[00:10:26] [SPEAKER_02]: they're parts of this business that are quite low margin. So one area where they're not doing

[00:10:31] [SPEAKER_02]: everything is in actually building up the servers that these AI chips go into. So they give people

[00:10:38] [SPEAKER_02]: reference designs but they're not actually producing those things for people so that's one example

[00:10:43] [SPEAKER_02]: of something that Nvidia has stopped short of fully producing and supplying it. That was a reporter

[00:10:49] [SPEAKER_03]: Ace of Fish and that's it for Tech News Briefing. Today's show was produced by Zoey Colkan with

[00:10:55] [SPEAKER_03]: supervising producer Katherine Melsop. On Zoey Thomas for the Wall Street Journal we'll be back

[00:11:00] [SPEAKER_03]: this afternoon with TNB Tech Minute. Thanks for listening. First the bad news. SAP Business AI

[00:11:09] [SPEAKER_00]: won't generate amusing holiday cards but a will person allows career paths for your people

[00:11:14] [SPEAKER_00]: and let you know which suppliers are best so you can be ready for the next opportunity.

[00:11:18] [SPEAKER_00]: Revolutionary Technology Real World Results. That's SAP Business AI.