저자: ChamathPalihapitiya · 외국인
요약 보기
엔비디아는 엣지 추론(로컬 GPU PC·로봇·산업 현장 실시간 AI) 수요가 커질 것으로 봄. 로봇 테스트벤치와 자율주행 분야에서 테슬라가 시장을 주도할 것이며 완성차 업계 재편이 올 것이라고 봄. 발표자들은 엔비디아가 장기적으로 밸류에이션을 정당화할 것이라는 입장이며, 브로드컴은 커스텀 AI칩과 하이퍼스케일러 수주로 강세를 보이고 있다고 전함. 소프트웨어 스타트업 피벗 관련 논의도 포함.
다듬은 글
thoughts shath on nvidia's dipping their toe into maybe taking the bottom 30% of the stack of self-driving I don't have much of an opinion on that to be honest I think that sort of along the lines of what I said on the prediction show I think that way and Tesla are going to run away with this market and I think it's going to force a bunch of consolidation in the traditional Auto oems I think the interesting thing is that they really doubled down and created a pretty decent test bench for robot I thought that was pretty interesting so I think that reinforces what a lot of smart people including you know what freeberg and Gavin also spoke about just in terms of the long-term future for robots I think that that was cool I was I was a little confused by the lowend PC I don't understand what the point of that is maybe it like creates some crazy deepend Market where you can buy GPU and then contribute it to some distributed Network and allow some distributed workload to run on that I guess I don't know I think it's a toy a hobbyist kind of device that becomes like a bridge and we see this often in technology where somebody creates like even the original PCS let's face it they were kind of like toys and hobbyist devices arduinos and the original drones were kind of hobbyist yeah I guess the point is a toy to do what because if you're trying to do inference like everything is telling us that we are reaching the limits of training and that's an L though right the point is it's not yeah let me get to yeah yeah so so in this world of AI that we know it today there's training and there's inference and right now we think that there's training that's at a limit and so now the market shifts to inference so if you're going to buy this jacked up personal computer what are you going to use it for my suspicion is some sort of test time compute use case which is an inference use case but it's not clear to me why that's a better solution than all of the AI accelerators plus tensors that are now just prolifically being exposed to the market whether it's Amazon exposing what they've done whether it's Google exposing what they've done a whole Litany of startups exposing what they've done so I was just confused I don't really know what the whole point is what do you think about this the robotics thing was interesting if the market develops in the way that they think so we're talking about maybe two or three different pieces here freeberg which one do you think is super interesting than the this $3,000 sort of GPU for your desktop that you attach to your computer you get to play with things locally do you think that's promising where would that go if you have to guess so I think the bet he's making is it's not just llms which is predicting text but you know we've talked a lot about Machine Vision models graph neural Nets that that are being used for weather forecasting there's now these kind of genome language models that are trying to predict genomic output for Bio uh Tech applications there's also going to be kind of Real Time Machine Vision and robotic response like we're working on this at ohal and we're trying to figure out what's the right kind of runtime environment for these sorts of systems that are going to be using a machine vision and a robotic kind of response type system and there's a lot of these industrial applications that are emerging let's say you're running a robot in a warehouse do you really want that robot in the warehouse to be sending data to the cloud and waiting for a model to run in the cloud and getting a response the probability is you want to have that at the edge of the network you want to have something local and I don't think he necessarily has a strong point of view on what the types of models and Industrial applications will be but the bet he's making is that the models are good enough and now the chips are good enough that they can actually Realize Real Time responses is using Machine Vision using realtime input and then respond quickly with a local model running whatever that model is to drive some output in the industrial setting and that there'll be a lot of these sorts of applications whether that's making predictions for biotech research or whether that's for running robots in warehouses or building new research models or maybe you could strap this PC on the back of something like a car a tractor a lawn mower a humanoid robot or any other set of applications last year's stock market saw Tech Giants leading the charge with AI stocks pulling indices like the s&p500 and NASDAQ to remarkable Heights at the Forefront of this revolution companies like Nvidia and broadcom became household names capturing the imaginations and Investments of many but what's next in this thrilling AI Saga are these Tech Titans poised for even greater growth or is there a new Contender lurking in the shadows chamat piaia thinks Nvidia is securing the top spot here once again let's analyze if he is indeed right about that welcome back to invest ocracy so cyan what do you think of these announcements and some of the other ones he made I know you were excited to talk about this yeah I'm really excited to talk about it because I think I've been trying to figure out how they justify their valuation over the long run and you know I'm not a public market person but I am fascinated with Nvidia and uh you know their Cloud GPU business is definitely Ma maity of uh the revenue so I think a lot a lot of what we're seeing is them trying to grow into that and trying to expand in case you know the music stopped now I don't actually think the music's going to stop you know it's insane to me we haven't even barely touched what AI is going to do and change and all of the various things that are going to come from it and you know the early adopters cannot use Claud without getting shut down because of scaling issues and I don't think those are artificially created based on the type of investing I'm doing and um so I'm very bullish on Nvidia it is interesting it's an interesting thing to go consumer and you know the thing that really hit me was the fact that uh he kind of declared Tesla one of the most valuable companies in the world in the long run it's interesting that he got behind Toyota but at the same time you know I don't think of a there's one single car company out there that has the kind of data full self drive has and and Tesla has um so if they enter the robo taxi Market I actually think they should buy Uber then it's kind of you should think Tesla should buy Uber oh yeah I think they should buy Uber well that would be about 10% of Tesla's market cap at this point if they paid a premium it might be 15% so it would be very similar kind of hard to but it's true like yeah it is true yeah and then and then you launch that Robo taxi service and you know maybe there's some sort of secondary aftermarket solution kind of like comma AI or something like that that you can do for people cars where you can actually get anybody's car into the fleet and start self-driving but it is true this is going to be the largest breakout in robotics we've ever seen if wh's any indicator and you know I I read somewhere I think that Amazon or some somebody was looking at I don't know what was going on with whmo but um oh lify Amazon was gonna buy lift yeah that makes no sense right like to me Dand would the point be I I do think you're correct I think maybe delivery or something like that I can't figure out what their play is there well it's not and it's also not Global but you know looking at the Amazon and weo Tesla and Uber I think weo plus Uber Amazon Plus Uber or Tesla plus Uber defines who number one is right because you'd have a global footprint and for the five 10 years maybe 10 years it takes to roll out taxis globally you could have people D I mean it's a really interesting thought process you have your say imagine if there was an interm step where they sold less Teslas this year slightly than last year you could just keep producing lots of model wise and give them to the Uber drivers keep reinforcement loaning going while the taxes and regulations get set and then you would be able to put another instead of selling 1.8 million Teslas you could sell three million Teslas four million Teslas to Uber drivers get all that data and have the safety driver in while each region decides if they want Robo taxis where how Etc Nvidia often celebrated for its cuttingedge gpus has dominated the AI landscape with Stellar performance its Revenue figures painting a picture of success recently their latest Venture the Blackwell architecture set the market a buzz with excitement promising unparalleled capabilities with demand exceeding Supply Nvidia seems to have a winning formula on the flip sideline stands broadcom equally impressive in its domain the sheer reach of broadcom's Technology impacting over 99% of global internet traffic is nothing short of phenomenal driven by demand for AI accelerators and networking enhancements broadcom's recent Financial achievements have set new records their strategic acquisition of VMware only amplifies their growth rent what's going on why are we seeing such weakness in Nvidia and broadcom really gaining so much strength pack this and start off with a couple of stats in the latest quarter broom's Revenue was just over $4 billion up 51% year on-ear Nvidia meanwhile posted just over $35 billion in Revenue up 94% year on year broadcom trades at a higher forp than Nvidia yet broadcom is getting all of this backing from investors right now there's a sense I think that investors are looking for the next Nvidia the company that's going to give them the outside returns riding this AI wave so you're seeing some of that rotation happen there but it's important to know also that broom and Nvidia are not necessarily an Apples to Apples comparison Nvidia develops gpus these Graphics process tessing units that have been used to train huge AI models like those made by open AI uh there is still no real Challenger here in this space meanwhile broon Focus really has been on designing custom AI chips for hyperscalers the big cloud players like Google Amazon or Microsoft and why these hyperscalers still need Nvidia chips they're increasingly looking towards some of this custom silicon that broadc con's designing that can be used for certain tasks or processes these hyperscalers then sell those Services through their Cloud Division and one reason the market is excited is because of recent comments by the broadcom CEO Hawk tan who said uh three hyperscale customers have developed their own AI chips on broadcom's own platform that will be deployed over the next 3 years and the company has received also a round of price Target increases from Wall Street all of that right now helping broadcom's recent rally to over a trillion dollars Frank with AI initiatives and VMware integration propelling forward momentum broadcom is a giant in its own right according to tweets from industry ins ERS broadcom's consistency makes it a formidable player in the AI Wars the conversation inevitably circles back to a common question if you had to bet on one would it be the Innovation Powerhouse Nvidia or the steadfast Juggernaut broadcom for those contemplating this decision their respective stock performances and Market positions offer significant Insight we were picking famously Uber you had to pick between scar lift and Uber there were three people doing it and it was pretty clear who was the most qualified you know amongst those three you now to your point if you're want to be involved in Tax Plus AI or legal plus AI you might be looking at 50 companies a 100 and it was tradition in Silicon Valley to not bet on competitors there were some notable exceptions when you run an accelerator like I do Tech Stars or Y combinator you aren't Bound by that because 50% of the companies pivot almost by Design so I think you just have to I think at precede because people pivot you just have to tell people like listen we have a lot of pivoting going on people are going to run into each other I can't just bet on one thing right in in a space but I think that's a reasonable compromise if if all the founders are going to keep pivoting to each other's businesses how can the investors even keep track of that it's like being aircraft traffic control of 10 airports at once it's just not feasible uh let's go ahead you wouldn't think it but there's still a lot of spreadsheet companies out there you know you think you'd run out of them but they're still out there you know you look at and I think this is where AI is really going to make a difference like RFP proposals for governments you know something that takes like 30 days and it's manual and you have to submit these horrible documents you know you can ingest your entire Corpus of you know all of your previous bids and submit them at a break neck speed now and win more contracts that becomes like a National Defense company at that point and and so I think we're going to see a lot of really interesting things where a lot of Cru is going to disappear and and that'll be a really interesting wave that I'm I'm looking forward to yeah in fact jamaath has made a big bet there with his time uh with his uh software startup that he's created as the debate continues between Nvidia and broadcom it's clear that AI driven growth presents numerous opportunities for aspiring investors nvidia's Relentless pursuit of innovation ensures its premium valuation remains Justified while broadcom's solid Market position offers an attractive bargain so dear investor as you consider your options where do you see yourself placing your bets and why
Comments 0