저자: Alex · 외국인
요약 보기
엔비디아 2분기 예비실적은 매출 67억달러로 가이던스 81억달러에 크게 미달했고, 게이밍 매출이 전분기 대비 44%, 전년 대비 33% 감소했다. 이는 팬데믹 특수와 암호화폐 채굴 수요 소멸 때문이다. 그러나 데이터센터 매출은 전년 대비 83% 성장했고 다음 분기도 60% 이상 성장이 예상되며, 발화자는 엔비디아를 산업·클라우드 컴퓨팅 기업으로 재평가될 것으로 본다. 자율주행 Hyperion, Omniverse, 기후 예측 AI 등 사업 확장을 소개하며 엔비디아를 '진정한 파괴자' 관심종목에 추가했다.
다듬은 글
nvidia just released a preliminary earnings report and it wasn't pretty they announced a huge decline in revenue about 6.7 billion dollars versus their initial outlook of 8.1 billion that's about 17 below expectations nvidia stated that the shortfall is due to big macroeconomic headwinds and weaker than expected gaming demand this drove their gaming revenues down about 44 quarter over quarter and 33 year over year here's the thing nvidia is viewed by the market as a gaming company because they make high performance graphics cards video game demand always comes down as the weather gets better but it's also way down compared to last year's numbers when everyone was stuck at home but another thing that was driving demand for graphics cards last year was crypto mining since the prices of bitcoin and ethereum were making new all-time highs now the prices of bitcoin and ethereum are roughly half of what they were a year ago so demand for graphics cards to mine them has dropped significantly so a year ago we had the pandemic keeping everyone at home and high crypto prices driving demand while the chip shortage was limiting supplies these are the kinds of numbers that nvidia is comparing against today so is that it is nvidia doomed well not exactly i first heard this news on cnbc's market alert with jim cramer but before you roll your eyes he actually made a really interesting point i would say this is a clearing event for a video nvidia is now going to be viewed as an industrial company right and that's what invidious industrial and and obviously data center i think jim cramer is actually spot on here nvidia may be still seen as a gaming company today but they actually had more revenue come from data centers last quarter then from gaming and their data center revenue is still quickly growing nvidia makes a lot of money from hyperscale and cloud computing and the demand for these services is only going to keep increasing as machine learning models get bigger and businesses process more and more data last quarter nvidia's data center revenues grew by 83 percent year over year this quarter those data center revenues are still expected to be up by over 60 percent year over year well after the most recent drop in their stock price simply wall street shows nvidia as trading near its fair value analysts still expect nvidia's earnings to grow by over 15 percent per year and that's on top of the amazing earnings growth they saw last year but it does look like there's been significant insider selling over the last three months simply wall street steps investors through an entire checklist of important factors like past performance current financial health and future cash flows all in a way that's easy to understand simply wall street is free to use forever but if you want to evaluate more companies each month or access their powerful stock screener you can use my link to get an extended free trial and 30 off your subscription when you sign up for now i'm adding nvidia to my true disruptors watch list which i'll be building as i keep covering the companies and technologies transforming our lives so consider subscribing if you want to follow along as i keep building out this list so let me show you some of the ways that nvidia is transforming our lives beyond gaming these are areas of their business that are poised to grow for the years ahead even if gaming ends up moving away from high-end pcs altogether let's start with automotive since that's one of their fastest growing platforms today nvidia has something called the drive hyperion system which is trying to provide automakers with self-driving as a service think of this as nvidia's version of tesla's cameras and full self-driving chip but with a few key differences nvidia recently announced hyperion 9 which supports 14 cameras 9 radars 3 lidars 20 ultrasonic sensors and doubles their processing speed of their current system the hyperion 8.
with all of these sensors the hyperion platform gives automakers many different tools for many different kinds of driving scenarios but not every automaker wants to offer full self-driving some just want premium features like drive assist self-parking and smart summon nvidia's hyperion architecture is modular which means automakers can use only what they need nvidia also has drive map which is a crowd-sourced real-time digital twin of every road they get footage for by 2024 they expect to map every major highway in north america western europe and asia that's hundreds of thousands of miles of digital roads that they'll have physical twins for in the real world if nvidia knows where a car is in real life they know where it is inside their digital twin which means they know the environment around the car where all the road lines are and traffic lights and curbs and so on then hyperion sensors in the car can double check what they're seeing against that digital twin likewise nvidia's drive sim takes in recorded drive data and turns everything in the scene into an interactive object with this tool you can remove add or change anything in the recording to run a bunch of simulations in that environment or change the environment entirely with different lighting effects and weather conditions and road hazards and so on kind of like what nvidia already does in the gaming market today drive map and drive sim are connected to nvidia's omniverse so let's talk about that next the goal behind the omniverse is to create the most accurate digital twin of the world and everything in it extended to vast scales a digital twin is a virtual world that's connected to the physical world and in the context of the internet it is the next evolution and that's what nvidia omniverse is about digital twins virtual worlds and the next evolution of the internet the idea for omniverse actually came from pixar who needed a way to set up complicated 3d scenes so that hundreds of artists could collaborate on them at the same time that's a pretty tall order and to solve that problem pixar developed the universal scene description or usd usd is a system for representing all the elements physics and interactions that make up a 3d scene that way one team can work on character models while another team works on environments and so on the universal scene description is where you describe how everything works together well pixar open sourced this usd standard and it's been widely adopted by huge companies around the world for example autodesk which makes 3d modeling programs used in engineering architecture and entertainment uses usd integrations so does epic games the maker of fortnite and the unreal engine and apple uses usd to represent all of their augmented and virtual reality scenes so this is already a well adopted standard but nvidia's omniverse is the first application purpose built entirely around this universal scene description what pixar's shared workflow did for large movie scenes omniverse does for every other industry from architecture and engineering to art and game design that means that one part of a factory engineering team can be working on the digital twins of different robots while a separate team works on the conveyor belt system and a third team architects the building itself all in different industry specific programs then they can plug in all of their models into a shared scene hosted on the omniverse if someone makes an update to their designs an omniverse connector would push it to every shared scene the model is in as well what this unlocks for businesses is the ability for multiple teams to collaborate on shared 3d scenes of any kind at any scale connected cars could upload traffic data for logistics and delivery companies to plan and adjust their routes weather services could upload weather data and models for everyone to use in fact nvidia made a machine learning model that emulates the movement and flow of global weather patterns to help others predict extreme weather events like atmospheric rivers this is an ai-enabled digital twin called forecast net and it's trained on 10 terabytes of weather system data combined with omniverse forecastnet can predict the path of these atmospheric rivers up to a full week in advance and it only takes a quarter of a second on a single nvidia graphics cards that's a hundred thousand times faster than today's standard weather forecasting models and it's not just atmospheric rivers severe droughts are happening around the world thanks to higher evaporation associated with rising temperatures which also causes more wildfires while we often talk about predicting weather and climates up to a week out nvidia's new technologies could one day help us predict months years or even decades out long-term climate prediction involves modeling the earth's atmosphere terrain oceans human activity and how they all interact because these predictions can happen so fast nvidia can generate many thousands of simulations explore the many different potential outcomes and quantify the risks associated with the most extreme weather events risks that may have been missed because today's forecasting models take so much longer to run here's the thing to get a model with meter level accuracy you need to be able to account for things like small low-lying clouds and how they reflect the sun's radiation back into space if you don't account for all these little details the small errors will begin piling up which causes big errors over longer periods of time this kind of modeling is hundreds of thousands of times higher resolution than what we've been able to do today and it's not because of supercomputers it's because of advances in physics-based machine learning like the kinds enabled by nvidia's modulus supercomputing is just the icing on the cake super computers like meta platforms research super cluster or rsc which is expected to be the largest customer installation of nvidia's hardware ever this supercomputer will connect over 6000 nvidia gpus to do things like translate the voices of large groups of people each speaking a different language supercomputers like eos which nvidia is building in-house to provide ai training as a service eos is expected to be the most powerful super computer in the world about four times more powerful than meta platform's research super cluster it's also expected to be the fastest with around 70 terabytes per second of bandwidth eos can transfer netflix's entire movie library in a single second hopefully this quick episode helped you understand just some of the things that nvidia works on outside of gaming because of efforts like these i really think that nvidia will be known for their industrial and cloud computing solutions in just a few short years but there's another computing company clawing their way to the top by any means necessary and they're already worth more than intel so if you want a high-tech story full of lying stealing and corporate espionage the company that controls all of our chips even nvidia's this episode is for you yo
Comments 0