you may be watchin: gpu ai base also. you may be watching this _ gpu ai base also. this from _ gpu ai base also. you may be watching this from home - gpu ai base also. you may be| watching this from home whilst type at work and remote work increasingly common. lots of businesses, office buildings are empty. now, vacancy rates are empty. now, vacancy rates are at a 30 year high in many american cities. new york's vacancy rate was at an eye watering 22% in the first three months of this year. that faces questions. what should be done with all that empty space. from new york, michelle sent this report. the big apple, defined in part by the the towers are sinking under the weight of skyscrapers made useless by remote work. high interest rates are making a bad situation worse. many are financed with short—term loans. the is $1 trillion of debt you in the next 18 months and so that'll be hard to be finance these loans and a lot of landlords decide that it is not worth it and they will decide to give back the keys and said an default on those loans. some real estate _ an d

this gpu ai race also.se— themselves into this gpu ai race also. well, you may be watching — race also. well, you may be watching this _ race also. well, you may be watching this from - race also. well, you may be watching this from home . race also. well, you may be i watching this from home with hybrid and remote work increasingly common. lots of office buildings are now empty. vacancy rates are at a 30 year high in many american cities and new york's vacancy rate was at high water and 22% of the first two months of the year. this raises a question, what should be done with all that empty space? we have this report. the big apple, defined in part by a slimming office towers, is sinking under the weight of skyscrapers made useless by remote work. i interest rates are making a bad situation worst. many are financed by short—term loans. same out there is $1 trillion due the next 18 months of that will be hard to be finance these loans lot of landlords will decide that it is not worth it and they will decide

they may do specific tasks very, very well, even better than the gpu. talking about the whole package, not necessarily the case >> hold on, jenny. i'm sorry. the tale of two chip makers if we're talking about nvidia and intel, because i want your comment on this, too, the gross margins over the last ten quarters -- i don't mean this to you, jenny, in any way -- it just shows the incredible divergence, gross margins of nvidia, red. intel, blue. look at the difference which underscores the story -- it's not like intel is a terrible stock -- hold on, josh. it's up 24%. nvidia is up 200% this year. they have that incredible ability to expand their gross margins to a degree others just can't compete with >> and their cash flow at $6 billion, hence the reason they had the buybacks the discrepancy there, we're talking about 71% gross margin to 40, not even, for intel, and probably will continue like that because they don't have the same margins, as you pointed out, like nvidia. in terms of competition, i still stand by the statement based off the research that

that is one of the reasons why developers want to develop on nvidia gpus. they're great and they know how to develop on them how sustainable is this? as far as what the disruption is, we think there could be some disruption if we think three, five years out, deep water is invested in rain.ai, a handful of companies that are building next generation chips that help to create what we think is 100 to 1,000x cost savings of energy with this. will have to make that shift to more efficient shifts because the current pricing environment isn't sustainable down the road. >> steve, you were nodding the competitive advantage which is their own programming knowledge. >> to further gene's point, that software isn't transferrable we know amd, they've talked about it, they'll have their own rival chips. better potentially the language everyone is using to code everything can't transfer over to amd if you want to go out a couple more years -- >> it's a wall garden. >> just like apple what other competitors can come in apple, microsoft is working on their own chips but it w

i said, listen, am i going to buy -- am i going to, you know, get a gpu in this? he said, you know, gaming, yes, if you want to game on a pc, absolutely but your normal pc doesn't really matter what's in there. it could be anybody. what matters is where it's connected to cloud >> as for the street's reaction today, we finally have no more sells as morning star goes to hold new street high, again, from rosenblatt as they go to $1,100, which is 40 times high 20s, which people would argue the stock got cheaper on a forward basis yesterday. >> one of the things i've always said about nvidia is if you go back and look, like in 2016, if you go back that far and you look at the analyst estimates, they were insane and then nvidia delivers a number, and it turns out the insane estimates were cheap. >> low >> i watched this stock go from a 44 multiple to a 9 >> and that's what i have been hearing this morning, even if there are those who argue about maybe it's pull forward demand or you're never going to see comps of 100% again. the counterargument is a 35 multiple is sort

, it's 2.75 times greater in terms of performance than existing gpus that mix alone is powerful if it's down, i don't think it will stay down i'm choosing to buy other names off the beaten path. >> this is why we are talking so heavily about what the importance of this stock is to the overall market if steph's right, then maybe a miss is not such a big deal because people step in and buy do you believe people would buy the weakness or do they sell tech if this misses? >> i think i said before, i think if it sells off people want to own it i would -- steph, i think if they miss, the stock is going lower. if they beat by less than the bulls are hoping and it sells off, then people say the fundamentals are there, it's not bad. i think if they miss tonight, it would be very surprising i think that surprise would mean the stock is down 10%, 15% >> maybe the more important thing is by what degree do they beat >> yeah. >> considering what they did last time -- >> i think that's where we are if it's by what degree do they beat, i think when they guided they knew they had a lot of that in the ba

most often the hardware in gpus. believes we could be sitting on a 6 to 12 month gpu shortage, which is substantial. that means any company like oracle in this case with a stockpile is already ahead of the game or what they call a, quote, speed to deployment edge. especially since so many companies already store their data with oracle software. another reason why it is a strength. they like oracle with the 47% pop on the year to date stock price and suggest companies are only just starting to ramp up their a.i. cloud infrastructure and that leaves room for growth with oracle cloud products maybe the midas touch. >> they were smart enough to stockpile a bunch of nvidia chips. >> smart. also they have the relationship and able to get the chips in advance compared to some other companies that have been struggling to do so. >> you treat your best customer best, right? >> precisely. it is about connects and money. >> and who you know. >> and money. >> and money. >> always money. >> always money. >> money connections. >> m

they are direct competitors for gpu's and nvidia's off session highs.y morgan stanley's estimations, nvidia is only meeting 50% of the demand out there so when they talk about improving supply, it gives confidence. caroline: the analysts reaction, the price targets being raised on note single cell on this stock. ed: the values tripled beyond $1 trillion of market cap in the final analyst threw in the towel . they gave us a note of caution and set on china that even if u.s. technology export restrictions are increase in the near-term, there wouldn't be an immediate effect because demand around the world for the server design is so great but long-term, they called it the removal of opportunity and what is the world's key market for data center and electronics. china is a long-term question. caroline: we are approaching that particular ipo and the exposure to them at demand there. ed: we got the numbers out of the way we have to look forward. let's bring in franklin equity group. you guys hold nvidia across multiple fronts. what was your take from this nvi

and companies willing to invest in artificial intelligence and the backbone of that is nvidia's gpushich go into the data centers to train the a.i. models which compares to chatgpt. chatgpt brought this technology to the floor and shows what the technology is all about for the companies. you have the supply tightness which is supporting the prices the makers are charging. they are in a perfect spot of huge demand. not a shortage, but not enough supply of the chips which is helping them look at the margins. 7 71%. mammoth because of the way they are ramping up the supply with the demand much of that is driven by the data center business and not to dismiss the gaming business. this used to be a gaming company. chips going into gaming monitors it is eclipsed by the broader data centers. >> this is an extraordinary story when the stock took off in the spring a arjun, stick around. we have the strategist from hsbc frank, arjun gave us the highlights how do you expect consensus earnings to evolve now that we had this earnings beat and this strong guidance from the company? >> sure, yeah. i

so, each generation of nvidia's gpu server, we have the product.ontribute a lot back to nvidia. >> you mentioned this partnership is going way back for the ai boom. did you see this explosive growth would happen or did it exceed your expectations? >> we've been waiting, waiting and waiting. in this industry, checking all of it for 5, 10 years, that's pretty normal. i think our investment into the ai -- from my perspective, it actually starts from the consultancy. i think less than 10 years, so i feel good. also, it's the adoptive by microsoft, currently light of facebook. google, amazon, they are all adopting that. and it's all just within one year everything comes out. i do believe this is really not -- in the flesh. it's going to maintain and you see he will keep getting stronger and stronger for the upcoming years. yvonne: i want to talk to about what happened recently. nvidia's founder came to taiwan and he personally came to your booth and it was a great moment where you got to wear his leather jacket. first of all, how did that moment happen?

each generation of nvidia's gpu server.he most important thing is we also work to contribute a lot back to nvidia. >> you mentioned the partnership goes way back before the ai boom. did you proceed -- foresee explosive growth was going to happen or did it exceed your expectations? what surprised you the most? mike: we have actually been waiting, waiting, and waiting. in this industry i think investing in certain technology for 5, 10 years, i think that is pretty normal. pretty normal. so i think our investment into ai, the real, from my perspective, the real investment , that actually starts from 2015. i think less than 10 years. so i feel good. also, the point is, the generative ai is actually not adopted by microsoft, but also, light, currently, by facebook. google. amazon. you know, they are actually there. it is around us within one year. everything coming out. so, i believe this is really not a spotlight. it will maintain. you will see it will keep getting stronger and stronger for the foreseeable upcoming years. yvonn

on gpus, which could include 5,000 nvidia chips. bulls are calling for data center revenue to jump to $15 billion in future quarters, versus the $8 billion estimate for q2. a contrarian view, nvidia's ability to guide materially above the current level is, quote, highly unlikely. if supply is tight, how are they going to keep guiding that much higher right now? the few other concerns, this current gpu shortage and gold rush mentality are almost certainly leading to double ordering right now, which might be especially true in china due to the fear of getting cut off because of export restrictions. so could that be a hurdle later on for nvidia? and if we're going to talk about the stock moving today, nvidia's contractor, taiwan semi is up almost 2%. today it reiterated its prior outlook in light of prior reports suggesting conditions have weakened. one name i wanted to bring up is analogue devices, higher ahead of earnings about 1% higher. there are some concerns, though, about analogue demand, and weaker demand coming specifically fr

look, nvidia has the most powerful gpu, right? you can run lots of data processes in parallel rather than focus on one big task but it's best used for large language models that have billions of parameters. we are talking what anthropic is doing, but the future of ai is more democratized. people will use large language models that don't have billions of parameters. they have hundreds of millions and of those don't require the same computing power. nvidia's response is that means there's a much bigger market. there's more upside. it will keep going for them. but i think intel and amd would have something to say about that. they would say, we also offer relevant gpus and cpu's that can power lower power llm's. we agree -- the message is it's the beginning of a phase of accelerated computing and the computing needed for generative ai. guy: you have more to talk about later on. there was a lot of gpu -- whatever that was. there things going on there. caroline will explain. this is bloomberg. ♪ if you're trying to get a view of the who

super micro is down the lane when it comes to getting gpus. the gpus from nvidia are given to the hyperscalers first like google, meta, but there is overall concern about that supply level if that's the case, there's only so high nvidia can climb because they can't increase that utilization level. they can't keep producing the gpus if they don't get supply. the story is pc recovery we saw it in the latest earning report, return to profit for intel after two quarters this could be another reason this stock is trending upward and a reason why memory prices have also increased. there's a lot in there but the fundamental argument for wolf. >> you made it, thank you. kristina partsinevelos >>> next up, we'll turn to retail credit suisse raising their price target on walmart to $180, expecting momentum we'll also hear from competitors in retail earnings season next week senior retail reporter courtney reagan joins us with more on what we can expect it is interesting to see the wall street analysts continue to love this stock, walmart, even as it's m

and that's what arm wrestle himself into the gpu, ai race right now.l be watching this from home... with hybrid and remote work increasingly common, lots of office buildings are empty. vacancy rates are at 30—year highs in many american cities — new york's vacancy rate was an eyewatering 22% in the first three months of this year. it raises the question: what should be done with all of that empty space? from new york, michelle fleury sent us this report. the big apple, defined in part by its looming office towers, is sinking under the weight of skyscrapers made useless by remote work. high interest rates are making a bad situation worse. many buildings are financed with short term loans. there's about $1,000,000,000,000 in commercial real estate debt coming due in the next 18 months. and so that will be hard to refinance these loans. and a lot of landlords will decide that it's not worth it, and they will decide to give back the keys instead and default on those loans. some real estate companies will not survive the crisis. others are pivoting. so this

it led to special gpu stacks. >> is there a way around this?have to go to nvidia and pay up? we had analysts say nvidia has 90% of the chip market. >> they had a strong lead. they were early and decided to reinvest in the gpus. a.r.m. is doing work there. the ipo is coming up. they have the market share in the smartphones and they have chips which is use for a.i. and data center. for the next couple years, nvidia will really enjoy the lead they have. >> you have to pay to play. you innvest in companies. valuation is the theme of the year. i want to focus on the hyper scalers with direct business with the portfolios. amazon, microsoft and alphabet. 62 times and microsoft is 42 and alphabet is 23. you know how important these services are for tech companies. they seem more appropriate? >> i think the big cloud providers are looking at $30 billion to $70 billion a year. they are moving stinfrastructur to the cloud. those will represent the full market cap with the companies at $1 trillion to $2 trillion market cap. you get the free traditional bus

due to the shortage of gpus, because tsmc's capacity constraint and.s constraint. so think we have a very much limited supply. but next year i think nvidia will solve everything. if you are talking about the only just the gpu, i think there will be a couple of times of. rose. how many times can you put a number on it? that's very at least twice or triple the demand. yeah. wow. are you getting more visibility on supply now? this is actually very difficult to predict, but basically what i can see is that the demand is quite strong and the supply chain is actually not only just for the gpus, but also, frankly speaking, like a power supply pcb, you know, and also like the thermal module, you know, chases, you know, actually all kinds of supply chain. i think there eagerly to increase the capacity and the production. so i believe everybody actually work on the same direction. and hopefully next year will be better. geopolitics is a big part of that as well. how do you see that? you know, the tensions between us and china impacting how you manages your suppl

nvidia is deploying the second-generation gpu's already but amd is late to this game. the story for amd is ai accelerators. what is an ai accelerator? kunjan: i think most people are aware of the cpu, the general-purpose compute, designed to process a lot of different tasks and do it well but an accelerator is a chip specifically designed to vastly process one specific type of task, hence the name accelerator because it accelerates the processing of one specific type of task. a gpu is a type of accelerator, other types are chips like ones from broadcom, the tpu google uses is one. when there is a chip designed to only compute for specific task, or do it fast, that is what we call an accelerator. ed: away from the excitement of ai, this earnings season has been about the calling -- about the core businesses. what did you learn about amd's health in pcn data center? kunjan: similar to intel, the results around pc was really good. it shows the pc market bottom is turning and on its way to recovery and the second half will be a tailwind. and it comes to data center, we see

what we've learned about is that's not true that yes a gpu specifically the h 100 to a 100s they make of large language models. but very big ones. if you speak to intel or and, other chipmakers they will say hold on we offer gpu's or cpus that are needed. thinking about micron, makers of dram and memory chips. data still need storage so were learning from the chip perspective there are other investments you can make an this school of thought is changing. as ai becomes more democratized and accessible we see some of those benefits. guy: ed ludlow with the ringside seat on what's happening. i know ed and the team will be talking about all of this in around 90 minutes time. ♪ ♪ avalarahhh ahhh every business that's why comcast business de is launching theal. mobile made free event. with our business internet, new and existing customers can get one year of unlimited mobile for free. it's our best internet. powered by the next generation 10g network and with 99.9% reliability. plus one line of free mobile for an entire year. it's the mobile made free event-happening now. get started for ju

everyone all over the world is trying to make sure their data center is stocked with as many of the gpus as possible and this entire architecture is built on top of their proprietary software platform. that didn't change during the course of this quarter maybe a reiteration of that demand going out years, quite frankly. when you look at the report from amazon, alphabet, within the cloud more and more upgrades to gpu architecture from cpu are inevitable i'm in it for better or worse. probably the best prediction of the future i can give you. >> you mentioned the visibility, the general sit for the morgan stanley call that initiated this bounceback in the stock. back to the issue of where we are, whether the economy, steph mentioned atlanta gdp now, 5%, retail sales today beats expectations another issue is whether it's just too good, whether data is too good it engages the fed perhaps more than people want to realize. no landing, no good. steve liesman joins us now as neel kashkari is on the tape saying he's not ready to say he's not done raising rates yet. what do you think here >> neel

the point is there is no imminent amount you can because there's a physical constraint to how many gpu's can produce. as it emerged, the center has been pharaoh will but i agree expectations are high so they are going to have to hold some outstanding numbers to beat them. matt: if you are a hedge fund manager and have a couple hundred million dollars to play with today, do you go short nvidia just as a tactical move because the chances of them beating expectations are lower? ivana: we run a strategy that is a little more long-term focused and nvidia is a top holding in our etf. we would not go short here because you are fighting against a powerful, long-term trend. if you look at our ideas, nvidia currently has the best risk will reward. -- risk reward. it is not have the high upside because we have seen the easy money made of the lows but it does have good downside protection because there are a lot of investors that missed on the trade and are going to look to enter if the stock is lower. with hedge funds, it is about risk/reward. it is not just about how much can you make or miss, you

they do not have the gpu.ould catch up but on this x86 system, this goes to show there is a risk they will continue to lose share. i think there is a shared learning aspect to what nvidia is doing. matt: just to be clear, because my mom is watching, that is compound annual growth rate. we have to explain this. great having you. thank you for joining us. mandeep singh, bloomberg intelligence. covers companies like nvidia which is expected to drive markets. how key has nvidia been to the rally we have seen? gina: it has been a huge part. in the spring months, the peak concerns around concentration rest occurred. a lot of this was a reflection of nvidia's recent announcement that they would crush earnings expectations which re-floated the space. there -- the bigger trend for 2023 is the magnificent seven driving everything. matt: we will see this play out in this morning's trend. kara murphy of kestra joins us. this is bloomberg. ♪ the first time you made a sale online with godaddy was also the first time you hea

and it is giving rise to gpu as a service, all the second-tier gpu as a service firms, as well. so, there's a lot of spending and we are yet to see whether the apps all come and the enterprise adoption comes, but the spending cycle is going to last well into next year, probably through next year >> so, the bears will say there could be overordering. there's a lot of pull forward to this, and you actually make the point that they're trying to build a more durable stream of revenue, that they are looking -- >> yeah. >> to create this sort of eco system can you explain that i feel like that is, you know, the notion that it's a one-off, it is sell its chip and done is the pillar of the bear case. >> i'm really glad you asked this question. every once in awhile, a special company comes along that's a plot platform this is one of them. there was apple, microsoft, but every once in awhile, one of these comes along, it's rare, and nvidia is probably the only company that has the ability to do that. you generate an outsized proportion of the profits for that industry, and it's more durab

we think we'll get there and the exceptional h-100 demand gpu is harder to get than a toilet paper inmic is how one person put it this is completely hard to get. nobody can get it. it is booked out for next year. the question is how high does nvidia go on the data center growth and i think we'll see a beat for sure. charles: how about historically when something has been hot, right? you know there is inability to meet that demand we've seen historically, no matter what the product is that demand start to wane or alternatives are ought out? at some time will they be able to deliver the particular server or chip? >> that's a great question. it comes down to supply and nvidia has a lot of supply to serve the demand but ultimately there is really no other a.i. accelerator on the market like nvidia and that is the key piece that the bears are not understanding is when you have only a near monopoly, 95 to 99% monopoly in the data center gpu market that is the key piece here where it is a little bit of a mix much supply and demand where nvidia has all of that supply. charles: beth, let me br

we're seeing a shift back towards nvidia, and a new kind of data center in play, gpu only data center. so there's these private companies. i think you guys hosted the ceo. that company is go inc. to go from $30 million in revenues last year to $500 million in revenues this year to $2 billion next year. and this is nvidia pure play from a data center angle. that's where a lot of your demand is coming from. >> there is this thing call the law of large numbers. i'm not sure what the law of large numbers says, by the way, but it seems to me that what it fundamentally says, when you grow at this rate and you get to this scale, compounding growth at the current rates become more and more difficult as you move forward. so there is an argument, i suppose, that this is as good as it might get for nvidia. talk to me about that and tell me i'm full of water. >> i don't think you are. i think you are spot on. and something very important from a standpoint of how this is. i think we are at peak growth rate. 78% sequential growth rate followed by 20%, something sequential growth rate. these are as

global race to accelerate ai capables put a major strain on the gpu chip supply chain and could limitr-term upside for nvidia can't get the parts, can't sell the chips. supply will be the primary determinant for its data center revenue at least through 2024. but, wall street, of course, continues to pump up its name. today wells fargo, ubs all raised their price target. you can see on your screen, well above 500 bucks ahead of nvidia earnings out next wednesday. keep in mind, there has been recent stock fluctuations with nvidia down 5% on nont, so it shows investors are nervous this ai phenomenon will cool. it's still too early to get off the train. consider it nvidia has that first mover advantage, for now at least. >> they are the name in the category so far as i can tell. christina, stick around far couple minutes as we are joined by our next guest, runs a company caught the middle of the race to secure ai chip may be a bit of an under the radar name shares up around 4% right now today. and more than 150% this year a lot of ceos would like that kind of performance. the company stil

. >>> carl, dominance may be one estimate, ai-related gpus, most people know this it's not just about the hardware, the gpus there is an entire ecosystem that includes a key software layer, specialized computers, that makes nvidia more of a one-stop shop for ai developers. some of that, like the software layer, is designed to keep customers within its system. for example, it can reduce flexibility for the biggest cloud provider aws, working with a lot of ai companies. so competition, while it is still far from challenging that dominance, it is beginning to emerge, and we dived into it on the software level a number of startups are working on tools that they say reduces the cost of training and running machine learning models compared with nvidia products others are targeting alternative software that would work entirely outside of the nvidia system, thus freeing up developers to use different chips. now one such startup is modular, and it just received $100 million in new funding from general catalyst, google ventures, among others i spoke to the ceo yesterday who said there are thous

under the marches , honoring the advanced brigs took place at the beginning of 1933, on behalf of the gpus first photographer was secretly sent to the white sea canal he made more than 2.000 tips of alexander rotchin on the channel with him along with the management. the channel will be very easy to meet and the steamboat karl marx on which to mark the end of construction will be written or headed by maxim it was decided to perpetuate the construction of the channel and glorify it on august 5 , 1933, a decree of the central executive committee appears that it is necessary to publish a monograph on the channel under the leadership of the ogpu for this purpose the best soviet writers go to the channel maxim gorky alexei tolstoy mikhail zoshchenko vera inver. leonid leonov and this collective monograph. now this is a bibliographic rarity, the circulation was almost completely withdrawn in 1938. after heinrich yagoda, the head of the canal construction, was repressed, the publication contains queues of german construction managers with a lot of data on technological methods on may 18, 1933 du

and very often that sort of gpu only comes from american tech companies. and case in point, in many cases it was nvidia. so nvidia was providing a ton of the chips that were processing that were used in the applications to process large amounts of video coming in to recognize these faces. so a lot of the face and image recognition was based on these chips and it wasn't just the high end stuff, like low end stuff, like hard disk drives, because you need huge amounts of storage space to store all that video and often the chinese police would call for video to be stored for 365 days. so a lot of these hard disk drives came from companies like seagate and western digital. so all in all, i mean, what we found was in the not we got less of supply chain, commercial or financial partnership. there was a ton of links between silicon valley and the chinese surveillance state. yeah. so what are you describing here is and i'm taking this some of some of the readers questions which i'm also seeing right now here. and so i'm trying to shape them onto a follow up for wha

gpu's with a key component that enabled the process. without that, you never would have been able to scale out the way that you can today. sorry for getting a little bit too technical. to take it back up a level in how you apply a and how it will enable the industry today, it will help everyone get their work done, as long as you have the appropriate structure in place. folks that are learning how to program can go and use the copilots or large linkage model generators that can help to write code or proof it for you. it is important industry realizes you can put structures in place to make sure sources are cited. it really is being seen as a productivity enhancement tool. i see ai as being infused every day, aside from generated ai, but also in terms of helping large organizations be more efficient. there is lots of work being done in areas of document processing. i worked on a pilot to show how you could extract handwriting recognition from forms, then extracted to be used as data that can be integrated into systems. in the federal gov

in theinterim we are going to have a major shortage of these gpus which is going to cause to a deflation. >> you just referred to it as a bubble, the a.i. bubble. do you think this whole thing is a bubble, that it's all overstated? i was talking to a ceo last week who said we've been playing with open a.i. and new microsoft products and a bunch of things out there. day-to-day use case, still limited. >> yeah, i mean, i definitely think it's a bubble. two things can be true. there can be a tremendous tail wind ahead of us with an.i., which i do strongly believe and the value capture can be quite limited. if you liook historically, ther have been probably, andrew a million or so a.i. products that have launched in these past few quarters and still today, august of 2023, the only a.i. products are oki, and others and i think we are at a stage where the hype is way ahead of the sort of revenue capture, revenue generation and i do continue to believe that the majority of the value here is going to be captured here by the big tech. you have the top tene enterpris and the top ten with $50 billi

gpus allow ai to happen. everyone is ordering them. the earnings come out next week.he earnings projection, the price/earnings down the road is astronomical. >> we are looking at $2 trillion ai market. they will play a big role. stuart: adobe is doing well. >> the thing people don't realize, google has better chips than the ai race and the processor unit, that you have right now in nvidia, those willpower the next generation of ai. you might see that at the google next event. stuart: all three of those stocks look solid do you? >> they look solid. they will be on the base for a long time. stuart: the best performers of related stocks? >> microsoft, can't discount that. apple, stock is high at this point. stuart: stay there. i have more for you in a little while. they are about to open the markets. right hand side of the screen, the nasdaq. when someone reaches across, the market is almost open. five seconds, the bell rings 10 seconds before the opening. it is 9:30 and the market is open as we open, the dow was up 150 points. not a huge loss. . 42%. there you have it,

spending on cpus to the graphic processing units which nvidia is making amd is ramping up to mass produce gpund intel is getting into the game as well. competition is ramping up. nvidia won't be the only game in town that is part of the equation as investors are looking at nvidia to see how much it maintains going into the rest of the year and into next year as demand for the a.i. chips are likely to continue. >> we had call tqualcomm and we antiand adve -- we had anticipated apple and then nintendo is up for the year with revenue coming in at 460 billion yen. "super mario" is a positive for the stock. >> "super mario" quarter it reported a decline in revenue two quarters and decline in profit they were concerned. has nintendo switch peaked has nintendo peaked in the current cycle? then the june quarter and revenue up 52% it was on two factors. the "super mario" movie has had a positive effect across the products it has driven people to other mario games on the switch. it has boosted revenue in the mobile division and the company is seeing the positive effects in the portfolio the second big

what about the next gen gpu with greater capacity? how does that drive the story for nvidia? options market pricing a 10% swing, almost 4% of the s&p 500 and more than 30% of the nasdaq's rally year to date. we cannot wait for these earnings after the bell. the opposite story is peloton. the stock hit a record low after its fiscal fourth-quarter earnings. revenue of 600 million is way below expectations, even at the high-end of its given range. the story in the quarter gone? the seat post issue, a recall. 750,000 people say i do want a new seat post. subscribers hit pause. they lost 29,000 subscribers. peloton going from a hardware maker to a subscription and software provider is unraveling. let's get more of the peloton earnings. what worried you most about what you heard from peloton? >> thanks for having me, ed. there are a few things. peloton, it is clear that they are having some potentially structural issues in terms of their ability to sell more hardware. the fundamental question for peloton remains whether there are already enough household owners for fitness equipmen

ubs is confident that oracle is carving out an edge in the gpu and cloud infrastructure.hat should be enough to attract new customers. shares are up 2.5% in pre-market. >>> and morgan stanley downgrading farfetch to equal. it says high volatility and limited path for a catalyst makes it difficult to be overweight on the stock in the medium term. shares down 2%. >>> time for the global briefing. shares of apple are moving higher in the pre-market as the company prepares to challenge a $2 pa$2 billion lawsuit. apple has hidden defective batteries in software updates. >>> and toyota halts operations at 12 plants due to a computer error. it doesn't believe it was caused by a cyber attack. >>> and shares of byd are surging. profits jumped 200% compared to a year ago. byd is the biggest china auto brand and counts warren buffett as the biggest investor. >>> coming up on "worldwide exchange," the one word that every investor needs to know today. plus, the key piece of economic data that jenny harrington is watching today and what it means rer the equity outlook. mo "worldwide ex

less the hyper scalars and if those companies can't monetize these gpu's than it could be a bit of ather words they won't be the magic pill there seems to be right now by the industry. haidi: when you take a look at some of the alternatives to china, we saw earlier in the year a lot of close going into japan. india now seems to be the darling. mark: we like both. in japan the story is basically buybacks. they really are happening and they try to get them happening 10 years ago and it never really took off. i think because his popularity waned after he change the constitution in 2014. he was not able to implement the corporate forms he wanted to. it is happening now and then the yen is probably the most competitive currency in the world. it is bizarrely undervalued, it cannot stay that way forever but as long as it does it makes japan credibly attractive. in india i think the story in a nutshell is i think the middle class will get better -- bigger and as it does they will invest in stocks. they have something similar to 401(k)s called systemic investment plans. i think there is over

using nvidia to train, you will then highly likely use it it has a pole position and you can't get the gpus until next year some time i think lisa is incredible and i'll see what happens here nvidia remains the most direct way in the public market to play earnings revenue growth and the shift out of ai. the word of earnings was optimization they are spending on ai and that road leads to nvidia >> joe, amd before we move on. >> amd was added to the portfolio along with sky works solution we own 13 semiconductors it's white hot, intense. 10% of the etf is in semiconductors in april, i don't know if i didn't have a rules based strategy if i would have said, i want to buy nvidia that's what the strategy had the courage to do. >> one more chips stock will be qualcomm, a big one, it reports after the closing bell kristina, what are we looking for? >> qualcomm gave us two warnings, hand set weakness. they said to expect a high growth but overall that's one warning sign and then the inventory drawdown just across the board they expect the inventor to continue drawing down they said that last quarte

relationship and the fact, one person pointed out himself in a call that they have already throttled back the gpuay given to china and if -- they have given to china and if the regulation gets narrower, they will have to lose out more revenue. stacy: that will happen anyway whether or not they never regulation. they put restrictions on what can be shifted to china and nvidia has designed product and there are thresholds against bandwidth that prevent you from shipping. the speculation is they will lower the thresholds lower. if they don't, they will never raise the thresholds so other -- overtime, the gap between what is legal to sell in china versus the rest of the world will get bigger and bigger so that china long-term opportunity will probably be under pressure regardless. if they do narrow the sanctions or make them tougher, you will get to that point quicker but my personal opinion, it will happen anywhere -- anyways. they have plenty of other problems -- over the long-term, this is something that needs to be thought of. matt: what is your view on the valuation? what are the most important

people talk about the ai gpu crunch.were trying to do is focus on the set of challenges that ai gave companies and bring people together who want to work at the vanguard and make it possible for companies to experiment. we are giving people enough capital to get going, 150,000 dollars on friendly terms and partnering with microsoft, openai and others to start to make it possible for people to experiment from the hardware side. it's in part because we think this is a special moment in time and we want to bring together the vanguard community and we think we can bring our network together and people will solve problems together. today's the last day of that location. ed: what is it you are looking for? if you are going to get a new wave of applicants right now, how did they set themselves apart? what are you seeking to unearth? >> i think the fundamental thing we are looking for with any entrepreneur besides the entrepreneurial traits of grit, the enthusiasm and deep understanding of a particular problem space, is actually

nvidia has circumvented the other restrictions by producing a low-power gpu.hat is your perspective on how u.s. technology names who want to do business in china can get around some of these executive orders. joanne: for a while i think companies like nvidia and amd will continue to sell to china. as nvidia has done they will have to sell less capable chips than elsewhere in the world. as you saw not just what tokyo electron said about the continued demand for equipment into china, china will want to build up its own semiconductor manufacturing capabilities, encouraged by all of these restrictions, not just from the u.s. but from europe. in addition we saw that according to the financial times nvidia is receiving $5 billion worth of orders from chinese companies because they are afraid the restrictions will only get more severe. china remains an opportunity even if they are constrained. they will want to get as high level chips and as high-level equipment in the semiconductor industry as they possibly can. it is pretty clear china has been making its own inves

people put money where their ai mouth is, is all of the hype around ai translating to big growth for gpu's the data centers? we will keep tracking it. this is bloomberg. ♪ ed: after nearly 20 years of investing in europe, silicon valley silicon valley firm ivp takes the leap and is opening an office in the u.k., relocating to london in 24 hours time, joins us now, i want to start with a big question which as you have chosen london, you are investing in europe, why london and not the capital of continental europe? eric liaw: london is closer in time zone. it is easier for my family, my wife and children are joining the deck pick. i have discovered in my decade plus investing in london that i am lucky that most people in europe are doing business in english, there are hours outside of the day that it is easier for us. ed: many who say london is just fintech and nothing else. your response? eric liaw: we have seen a lot of fintech investing there, we have made our first london-based investment in fintech companies recently called volt.io. forcing ai companies attract talent and being built th

nearly a year ago, we started going deep for relevancy moving from cpus to gpus and getting to recommendedn what we were using the year before that resulted in a 10% lift in perceived relevancy by users, getting to 94% which in the world of search and discovery is just unheard of so we're seeing great progress there on the organic engagement side and now we're bringing that to our ads platform where our ads models are 100 times larger than they were a year ago and that's leading to a fantastic benefit. >> i usually don't like to be, i like to be empirical she's a parsons graduate she's genteel. doesn't want combat. letter x, formerly twitter, is about combat isn't that going to be to your advantage that people want a genteel, kinder place to go? >> well, i'd say there's a couple of things that really make pinterest the first is that most of social media has the user in entertainment mode on pinterest, more than half the users are there to shop. they're in a lean forward mode so that's different than much of the rest of social media a second thing you were touching on that has been a big fo

and hopper -- it's all nvidia, but grace is the cpu and hopper is a gpu.md makes cpus, intel makes cpus. they're on the x86 model. this is not that. this is a major break through by arm to be the partner. remember, nvidia tried to buy arm. the head of arm is from nvidia, and they're very close -- very close with jensen. the reason i'm very excited about this deal is because, if they're the cpu chosen to be in grace hopper, which is a gh 200, the fastest computer -- what they'll need for soup your computers, that's a huge leg up on amd which i like and intel which i don't care for. going away from the 286 toward arm. arm is the winner. they have the best technology. that's why i'm excited about it. >> i think a key part of the story that they'll be selling to wall street will be ai-related. of course, it's real and offers particularly significant growth off the numbers you just saw. >> spectacular. >> what's spectacular? >> what arm has. by the way, arm is so savvy, you could argue they are going to displace intel, but also amd. >> they don't make their own

spending more money on ai builds, and nvidia is the leader right now, by providing the ai chips, the gpus there are no competitors in sight. they will continue to gain that market share it is an argument for near-term success for the company, possibly in the next guidance, the quarter. the other successful factor, i guess, for nvidia is demand coming from china, front running a lot of orders because of export controls. we learned from the "ft" saudi arabia is purchasing nvidia chips. you have china, a bunch of companies, hyper scalers as well demand is clearly there. yesterday on your show josh brown made the argument how can nvidia continue to blow past analyst expectations it was 50% past estimates the last quarter talking about guidance can they keep that momentum going when so many companies, countries, et cetera, are front running their orders right now to get a leg up on the ai push >> also the issue, hey, make sure you ask stacy ras con on "closing bell" about supply constraints because of the quality, quote, unquote, of the chips nvidia produces and their dominant position in that

the gpu's people use in ai, i think this is a big problem. more than that, the market necessarily won't panic if they can't keep up with supply. it is, if there is no roadmap. the stock price reaction is interesting on that coming out. guy: do you agree? >> yeah, i think a big question for big tech this year is, we have seen a strong rally despite the higher yields. the reason for that is maybe people are more positive about the sentimental's given the ai boom. you need numbers to back that up. you need economic resilience to back that up. the question is, can big tech actually survive these higher real yields? guy: which is going to shake the market more. that is what i am trying to figure out. what has the potential to shake it's more? we are bought into the ai idea, into the idea the fed is ultimately going to cut rates next week. if either of those two ideas go into next week, next year -- if either of those two ideas proved to be wrong this week, how seismic will those events be? >> exactly. the real yields part may be the focus. at this

you have to buy their gpu processers in order for you to put the code on and all that stuff.is where we actually get to see whether or not ai is blake g adopted as we hope and honestly as how much is priced into the stock already. already. stuart: you wouldn't bite stock at the time of the earnings report coming out. you'd want to wait for the call. the call tells everything, that look to the future. >> you've got to hear the future guidance. last quarter in may, the future guidance, they raised revenue estimates by 50% above where street had them at. that nevers, especially for a company of that size. i want to hear obviously future guidance, where they're going to be heading. jenson wang happens to be optimistic most of the time, that's good. i hope there's results we can stack. stuart: to be clear, nvidia runs the ai chip industry. largely. >> i read a research report that says they have 90% share of that chip industry right now fai. stuart: and their chips are faster than anyone else's? >> they're afew years ahead of anyone else. stuart: keep it for years. >> if it sells

networks, and fundamentally, these gpus have to be connected and they're sharing information, and theyave to do it very quickly. we've talked about the fact that we think the networking opportunity underneath the gpus could be three to four times -- >> this is underneath nvidia they have to connect it to azure. they have to connect it to amazon web services. >> and the historical way of connecting that is through a technology called infini band. >> owned by? >> nvidia. however, the cloud providers are telling us they want to move to ethernet, which is what we do, and we actually are a founding member of something called the ultra-ethernet consortium, which is designed to actually tweak the standards a little bit it's a little geeky, but tweak the ethernet standards a bit to make it more adaptable for these super high-performance environments >> now, you have been a leader in two things. one is work from where you want, and the other is, helping the homeless in what was your headquarter area still is give us a sense of where people are working and also, frankly, what are you doing for t