Nvidia, Sam Altman and the trillion-dollar AI dream

the biggest figures in AI they have just thrown out the bees and started talking in teas to AI Titans they're now prophesying trillions of dollars in Opportunity stratospheric predictions from Nvidia Chief Jensen hang and over the course of the next four or five years we'll have $2 trillion worth of data centers and then there's Sam alman's ambitious new Venture $7 trillion 7 trillion 7 trillion he likes he likes to big numbers that's more than the entire us fed budget two times the UK's annual GDP more than the value of Microsoft and Apple combined it's not even conceivable nobody raises trillions of dollars first of all alman's response to the critics you can grind to help secure our collect a future or you can write substacks about why we are going to fail substacks or in our case this week's Tech check ai's trillion dollar opportunity nvidia's latest earnings report it crushed analysts already Skyhigh expectations for the AI darling we start with that massive move in Nvidia you're going to hear that name a few times here bet the over on this one $22.1 billion in sales that's 10% higher than the street even expected forecasting $24 billion for the next quarter beating estimates by another 10% a lot of it though is driven by data center Revenue it's bread and butter which grew 38% quarter over quarter to 18.4 billion there had been signs of worry in the market that nvidia's Stellar run was unsustainable shares fell nearly 10% in the week leading up to earnings you have to wonder what point does something trigger this thing I mean the story can't I don't think last forever and double ordering triple ordering margins compression is more people get into the space more competition I don't know we're getting close concerned that demand might Wayne more competition and that invidious chips wouldn't pack the same punch as the industry shifts from something called training to to inference here's what that means the first step in developing an AI model is training it feeding it tons of data images video text that requires heaps of compute power with those highly sought after Nvidia h100 chips after the models are trained though the next step is inference the model is able to start generating its own text and pictures based on that training data one analyst wrote demand growth for NVIDIA chips will eventually moderate AI chip demand will eventually normalize once the initial training build has been completed the inference phase of AI is going to require less computing power than the training phase but Nvidia says it's cornered this Market too the amount of inference that we do is just off to charts now almost every single time you interact with chat gbt you know that we're inferencing every time you use mid Journey we're inferencing the inference part of our business has grown tremendously we estimate about 40% 40% of data center Revenue from inference that is two times higher than Morgan Stanley estimated and it crushes the Bear's warnings inference will scale like the more queries I have the more use cases I have the more compute I will I will need installed and and it's not all happening like at the edge a lot of that's going to be happening in the cloud and a lot of that will be on gpus too plus it's not like the training phase is over open ai's next GPT model is still in development training can actually get very big too because the model sizes you know G gpt3 had 175 billion parameters gbd4 they they haven't released the specs but it's rumored to have over a trillion parameters and and and those are going up and that could help explain why Nvidia doesn't see demand going anywhere anytime soon we expect the demand will continue to be stronger than our supply provides and through the year and we'll do our best the uh cycle times are improving another concern among investors the infamous cyclical nature of the semi industry it's known for being prone to downturns the idea that there's a Year's long but constant rotation between scarcity and over supply high demand for specialized chips that leads to a shortage and price hikes that leads to more investment and eventually more competition more Supply and then leading to a correction that could still happen and according to databrick CEO Ali godsy it could happen as soon as this year I actually think there's going to be a GPU glut this coming next year when there is in Market a scarcity the market rushes to prod so I think we're going to see lots more gpus next year what is that going to do to prices I think the prices of gpus of course comes down but nvidia's management they don't see that happening anytime soon we expect our next Generation products to be Supply constrained as demand far exceed Supply and the US is only the start my expectation is that what is being experienced here in the United States in the west will surely be replicated around the world and these AI generation factories are are going to be in every industry every company every region so the company has begun laying out their opportunity in trillions instead of billions all of those older uh data centers are going to be modernized so that'll make them more efficient over the next 5 years that's going to create a$1 trillion doll opportunity the second trend is going to be General uh Enterprises that are spending on creating large language models and so they're purposly buying new GPU data center uh units to build and so that's another trillion opportunity within the next 5 years after the company already added a trillion dollars to its market cap over the past year but Jensen Hong is not the only Top Dog in AI minding his T's and B's Sam Alman wants to transform the entire Global chip industry the picks and shovels of the AI Revolution and he reportedly may need up to seven trillion dollars to do it his chip dream is about developing AGI or artificial general intelligence because we're not going to have enough chips or enough electricity to power the chips we need to to get to overall artificial general intelligence open AI defines AGI as systems that are broadly smarter than humans and has the ability to teach itself thereby creating new even potentially smarter agis the implications for Humanity they're huge AGI could find new ways to cure diseases uncover new insights about the universe lead to breakthroughs and Technology but but there are also risks AGI could become so powerful that it might stop listening to humans and do things to harm us so the development of responsible AGI that is the true Holy Grail for AI researchers and chips or semiconductors critical technology to get there Sam malman is on the other demand side of nvidia's Supply shortage the argument across the board is simple Nvidia is the only AI chip game in town the strong demand for AI infrastructure will continue to outpace supply which is already constrain and everyone seems to be just clamoring to get these GPU chips gpus are those specialized Cutting Edge chips that Nvidia dominates the market for they're needed to power the most advanced AI models according to Altman there scarcity is the only thing holding back him open Ai and a breakthrough in AGI in Sam's view we need you know just a Quantum Leap in the number of chips we have it's hard to wrap your mind around what that leap would look like it's sort of interesting though in terms of what it supplies Sam Altman believes we need in terms of computing power though like is are they really like looking at things where you could need hundreds of millions of GPS open AI is already awing users just recently releasing another feature called sora that can generate videos just from text this is an AI generated video of two pirate ships inside a cup of coffee and this is a video showing logos Nigeria in the year 2056 and Wildlife on a river in Malaysia all hypothetical videos created purely with AI this technology open AI was able to develop with a GPU shortage so imagine what Alman might be able to accomplish with $7 trillion worth of gpus you talk about Investments and things usually don't talk about7 trillion that that is deep I got a lot of power a lot of computing power that's human level sentience that that is coming soon but there's a shortage for reason solving it is way easier said than done especially for Sam Alman who has spent his life in software and the Silicon Valley bubble chip making is entirely physical like cisos routers and switches enabled the rise of the internet in the 9s high-end chips are the hardware of the AI era I always say like these are the most complicated things that Humanity has ever made so there are only a handful of companies that are able to make those chips right now including Intel Taiwan semi and Samsung and it took them Decades of R&D to build up the capacity to manufacture them it makes more sense for him to partner with various foundaries or help them to get funding he could buy Intel for a lot less than that and fix Intel's problems for a lot less money than $7 trillion designing chips has proved just as difficult deep pocketed Mega caps like Google Amazon Microsoft meta they're all trying to make custom in-house chips but they all still rely on nvidias it's orders of magnitude larger than anything that like anybody's ever done the semispace is already constrained SMC and Intel right now are building you know they're having trouble finding enough construction workers each to each build like one Factory Alman may be looking to build hundreds if you could build them like where where could you get the tools again we're talking a number that is like multiples of the the amount that the entire industry has ever spent the lead time alone for getting chipm tools was as much as 12 months as of early last year how could you actually ramp up production to to to get the you couldn't and finally where do you even find $7 trillion during the pandemic the global chip shortage underscored just how Reliant America is on just a few companies the world's largest semi manufacturer tsmc produces 90% of the world's chips including all of nvidia's and its headquarters is in Taiwan a region with complex geopolitical Dynamics raising National Security concerns for the United States the fact that we are so overly dependent to a couple of countries in Asia to act semiconductor chips that we need for life-saving medical equipment cars uh every piece of technology showed us uh we got to get to work we need to get back to work making more chips in America it's why President Biden's $53 billion chips act aims to fund domestic chip manufacturing giving America more control over a critical industry but Altman's ambition to remake the industry at large it is a far bigger project one that would create a complex partnership between open AI chip makers investors power providers even governments it would require capital from even bigger sources which is why Altman is reportedly talking to Sovereign wealth funds in the Middle East with trillions of dollars in assets he sees the Middle East as a source of funds for those foundaries and he's looking for really towards the future where chips are the new oil and look guess who's going to be owning all these uh foundaries you know being the big investors the place that controls the all now it's kind of ironic which raises yet another question if semiconductors are an issue of National Security do you let Abu Dhabi get a foothold I would try to find a way to create more of a regulatory sandbox where people could experiment with this technology and what sort of auditing what sort of safety measures do we want in place before you can deploy like a super intelligence or you know however you want to call an AGI and I think for a bunch of reasons the UAE would be so well set up to be a leader in the discussions around that Alman reportedly met with the United Arab Emirates top National Security official to discuss the effort in recent weeks that official also chairs the Abu Dhabi based AI firm g42 which openingi has partnered with here he is with the g42 CEO but g42 is also under scrutiny for its ties to China the CIA and other agencies have issued warnings about its work with large ch companies in response g42 said that it would try to pair back its presence in China it's going to be in the trillions in the Middle East in the middle e in the Middle East that's going to go over well seven I mean Global I mean good for him Sam alman's chip Ambitions they are so big that some are comparing him to Elon Musk legendary tech genius troll maybe both on social media Altman is leaning into a provocative Persona after the journal article about his 7 trillion doll chip Ambitions after that was published Alman posted to X why not8 followed by quote our comms and legal teams love me so much a few weeks later he would then downplay the development in a conversation with Intel CEO Pat Ginger and as Nvidia surpasses $2 trillion dollars in market cap some are drawing comparisons to Apple nvidia's large profit margins and Cuda software ekko's Apple's proprietary offering that has underpinned its decade plus long dominance a new era of tech Titans trillion dollar dreams the AI transformation may just be getting [Music] underway

Channel: CNBC Television

Views: 281,416

Keywords: app, business news, cnbc, digital, disrupt, funding, innovation, investors, nasdaq, nyse, online, silicon valley, startup, stock market, tech, techcheck, technology, venture, wall street

Id: JrOgRQFqhbM

Length: 13min 57sec (837 seconds)

Published: Fri Feb 23 2024