NVIDIA to $500? Jim Cramer Drops a Shocking Clue

요약 보기

CES 앞두고 엔비디아 상승 기대. H200 중국 수요 급증, 소프트뱅크 1000억달러 오픈AI 투자 완료가 촉매로 제시. JPM은 엔비디아 목표가 상향(대만 35%)과 낙관적 실적 전망. 베라 루빈 플랫폼은 2026년 하반기 출시 예정. AMD는 엔비디아 추격에도 생태계 격차로 AI 주도권 유지 가능성 낮다고 평가. 자율주행·로봇택시 확대로 메모리 수요 증가 전망.

다듬은 글

※ 자동 자막 원문(다듬기 전)

01. Hey computing chips by Bitcoin. Jim Kramer scare Nvidia give Nvidia a hint by showing how they see the next industrial revolution using accelerated computers and what I call the stack. You will hear a lot more about the stack. The stack tells you that we are not pieces. We've moved on from oil from Venezuela to the trade and artificial intelligence.

02. Let's do some that's fun because today we're seeing a rise. No big risk on maybe bitcoins. No, it's because Inennia bought a lot of high band with memory. This AI rally is different in some ways. They have a lot of market which is why Micron keeps going up. We're not building it fast enough. I'm going to go back to Venezuela.

03. If you want to talk about Halatin, I don't. I like Hallebertton as it is. No, no, no, no. Please stay on this because it's the market today, the second business day of the year and there does seem to be an AI related rally. We have a CES address this evening or rather this afternoon. Our own John Ford will be talking about it and I think it will be a very big event which is why I think Nvidia is already up.

04. AMD is going to talk tonight. I agree with Jensen Huang that that is something that should be front and center at CES and in the data center. Truth be told, I think the last few months have been hard on Nvidia. And maybe the doubters are done now that the H200 is in such high demand in China. A huge portion of the 50 billion bike dances that were bought.

05. All the lies that said China wasn't interested. Those people will get away with it. Remember that with my new plan, I'm not going to let anyone off the hook. People who said China wasn't interested. It's possible that the PRC wasn't trying to stop them, but China demand H200 and the big lie would take it all.

06. Yes, the big names at Amelia got better today, which raised their goal for Taiwan by 35%. Okay, another story from Foxcon gives an unbelievably high number, which again shows the Nvidia complex. And as I already said, the most important thing was that if the high bandwidth memory that you see in stock twice, that was what everyone would need.

07. Nvidia knew this would happen. David, where are you in the OpenAI IPO right now? Memory loss. No one paid for it. I don't think I'm anywhere in the IPO, though. I'm not sure if something should come this year or next. As much as I'm trying to focus on a number 100 billion raise that Open AI is currently going through at a possible value of $750 billion, I knew this was where I could find out more about Jim and Sheriff.

08. It would be good to talk to Sheriff. Sure, let's do it. You should know that Sam was in St. Barts. A few people talked to Sam Alman along with everyone else and he seemed to be in a great mood. It's possible that they felt like they made a lot of work on the fund raise, but I can't give you any specifics.

09. Also, I don't understand why 100 billion is such a huge amount of money. Would it have Nvidia's new money in it? Let's go back to the hundred billion idea of Soft in 2025. Yes, if you remember, we told you that Soft Bank finished its full fundraiser about a week ago. This is likely the most important thing that will happen in the first quarter because it means you can buy Oracle.

10. It means you can get Verdive. You can buy Nvidia with the $100 billion OpenAI catalyst. Yes, that Nvidia and Alphabet work together is interesting. Google is our rival. Google has two competitors because it spent a lot of money and had to deal with a lot of high-speed memory. But I think Nvidia is the one to keep an eye on because it hasn't done well in the last month and is going to start off quickly.

11. And I believe Jensen's groundbreaking why Nvidia could move fast out the gate speech tonight will make it clear why fast computing is often left out of the picture. Yes, today JPM's TMT desk made some picks about Nvidia for the coming year. Easy to switch to Vera Rubin. China wants more H200s than anyone thought they would.

12. If CPX demand goes over 10 billion dollars and margins stay at 70% that's good but it doesn't show that I think it will have a great first quarter. The price range multiple is going to shrink when you look at the numbers at 41 now analyst profit margin and price expectations this week. Nvidia showed off a brand new Ruby platform that includes a CPU, GPU or other networking IC's.

13. We thought it was interesting to see that Nvidia can see that AI model sizes are growing about 10 times a year while token generation is growing about five times a year. Nvidia's brand new Rubing platform will handle this kind of AI computation going forward and they still plan to launch in the second half of this year.

14. I believe it's pretty verubin platform changes. It's great to see that Nvidia is on track to release a new system. This is also good news for everyone in the tech supply chain. Masahiro, what does this mean for the market for self-driving cars? If Nvidia works with Mercedes, how do they stack up against competitors?

15. Yes, I guess. When it comes to autonomous driving, I think it's really interesting because the robot taxi market will grow in the future. In fact, Bloomberg Intelligence had a really interesting story about the robot taxi market last year. You know, autonomous driving will kind of speed up the robot taxi market all over the world and Mercedes might be able to set his autonomous driving technology apart from other companies.

16. But you know, if we look at other things like autonomous driving, the AIN infras market is also growing, which will make a lot of people want memory chips or other networking chips. So, in that way, that should be pretty interesting for the whole tech supply chain from self-driving cars to robots and maybe even how that will affect the need for memory in the future.

17. Why the difference between AMD and Nvidia still exists? Okay, we also heard from Lisa Sue, CEO of AMD at CES. What do you think? Is it possible that their new goods and chips will make Nvidia lose some market share? Yes, AMD is working hard to catch up to Nvidia and I don't think AMD will be able to break the current competitive field.

18. I mean that Nvidia has built a strong ecosystem over the last 10 or even 15 years. Because of this, I think that Nvidia will continue to lead the AI business. AI needs info right away. Each GPU gets 1.6 terabytes per second of scaleout capacity from Connect X9. The Bluefield 4 DPU takes care of storage and security, so the compute can work on AI alone.

19. The Vera Rubin computer case fully. There are no longer any wires, hoses, or fans on the new Nvidia ecosystem advantage. It has eight connect x9, two Vera CPUs for Reuben GPUs and a blue field 4 DPU. It is the part of the Vera Rubin AI supercomput that does computing. Moving on to the sixth version of the MVL link switch.

20. It connects 18 computer nodes and moves more data than the world internet. Last word is $500 Nvidia. Really? It can handle up to 72 Reuben GPUs working together as one. Spectrex Ethernet photonix is the first Ethernet switch in the world with 512 lanes and co-ackaged optics that can handle 200bit. It can be used to turn thousands of racks into an AI factory.