Transcrição completa de Instant Reaction: Nvidia CEO Jensen Huang Ousted by Nvidia Board

all the relationships in terms of the ecosystem that Jensen had created. And I think that's where anyone stepping in new may have their own way of going about things on the product side as well as the financing side that you alluded to. 00:03:54 Speaker 5: Tell me with a GPU, folks, Mandeep Singh, we underplay it. We ask him dumb questions. He is incredibly good at the electrical engineering and computer science on this. What is the skill set? to go from GPU chip whatever to the next GPU chip. Is that a Mandeep Singh engineering decision? Is it a guesstimate about the depreciation. 00:04:22 Speaker 1: Of the older chip? How does that work? 00:04:26 Speaker 4: Yeah, look, I mean, what NVIDIA has done is they have launched, you know, new versions of their chips almost at a six-month cadence. And we are talking about, you know, their latest Vero Rubin architecture, which is, you know, five times better in terms of throughput than their Blackwell architecture. So... to be able to iterate on these things and stay ahead of everyone else, which is why they had such a big share in the GPU market, is very commendable in terms of execution. But that's where, if it's somewhat technical, it will make a difference in terms of how they approach the product side of things. 00:05:07 Speaker 5: Have you met Harvey Jones, the new chair? I mean, the dominant feature there is he's one of, I believe, off the Bloomberg Bandip for original directors. I mean, do you have any perspective of Mr. Jones' relationship with Mr. Wong and with the rest of the team? 00:05:25 Speaker 4: Yeah, I mean, look, I think when it comes to the board and, you know, the new chair, obviously, you know, everyone is very close to Jensen, given he's the co-founder and, you know, has been with the company for over 30 plus years. So from that perspective, I think I don't know Harvey in particular, but at the same time, I think it would be somebody who Jensen knows quite well. 00:05:53 Speaker 1: Well, we're going to come back. I want to talk to you about Ian Buck and all that he's done for the company. 00:05:57 Speaker 5: This is a headline moment for technology across America. A few more minutes with Mandeep Singh here, Global Head of Technology, Media and Telecom for Bloomberg. 00:06:08 Speaker 1: I look, Mandeep, at the product cycle here. 00:06:12 Speaker 5: Ian Buck seems to be incredibly important here. He is the Stanford technologist behind the moment here. How important is position now? And what does Ian Buck need to do in the coming days? Yes. 00:06:28 Speaker 4: So he's definitely the front runner when it comes to internal candidates. And clearly he is the guy with the CUDA aspect of the Bloomberg product lineup, which is why, sorry, the NVIDIA lineup, which is why NVIDIA has done so well. And from that perspective, you know, having someone technical will be quite beneficial at this point of time to calm down all the anxiety that investors have. 00:06:59 Speaker 5: And Mandeep underplays the ginormous technology. 00:07:03 Speaker 1: Feat of Ian Buck. 00:07:05 Speaker 5: This was Brook, B-R-O-O-K at Stanford. then went over to CUDA, C-U-D-A, which Mandeep will explain to me over a beverage of my choice at some point. And then you bring that over to the miracle that is NVIDIA. Mandeep, what is your experience of beast technology people, quite frankly, like you with great respect, running as a CEO company? Can that actually work? 00:07:36 Speaker 4: Well, what Jensen was really good at was he was both very good on the technology side and also great in terms of, you know, that ecosystem that he had created and the supply chain aspect of it. And that's a hard combination to find in terms of, you know, at the CEO level. So clearly, Ian has all the credentials when it comes to the technology side of the NVIDIA products. But I think to have those kind of relationships with TSMC and all the suppliers, which is why NVIDIA got to this position in terms of ramping up capacity so fast, I think that will be a true test of somebody who has gotten, you know, all the different aspects of what is needed for this company. 00:08:20 Speaker 1: Mandeep Singh, thank you so much. Greatly appreciate it this morning. 00:08:24 Speaker 5: I'm sure I have a long and hellacious day. It is widely understood at Bloomberg that they can't be on the same airplane. They can't even be in the same room. Anurag Rana in Chicago, as we have Mandeep Singh in New York, and we're thrilled to that Anurag could join us here for years at KeyBank and driving all of technology at Bloomberg. 00:08:44 Speaker 1: He is our definitive force on cloud. 00:08:48 Speaker 5: What is the challenge here, Anurag, in the relationships with all the people that buy these chips for data centers, etc.? Do you have to get out and sell it, as Mandeep was talking about, or do they even get an interregnum of 30 days? 00:09:06 Speaker 2: Yeah, I think the biggest thing is going to be, you know, NVIDIA has had a very massive, you could say, gap between the others when it comes to AI accelerators. But this is the chance for the others to catch up, you know, whether that's Amazon or Microsoft or Google can close the gap and say, well, why am I paying so much money to NVIDIA for those chips when I can have my own? So far, they are forced to buy those chips because That's what the customers demand. That is the best in the business. But if this gives them a little bit of a chance to catch up, I think this could be good for the hyperscale cloud providers. 00:09:42 Speaker 4: Anurag, you've covered Apple forever. 00:09:45 Speaker 3: As I think about the impossible task of filling the shoes of Jensen Wang, one of the names I think of is Tim Cook. 00:09:53 Speaker 1: Talk to us about that. Oh, you are such a gossip beast. Yeah, I thought about that as well. Oh, come on. 00:09:59 Speaker 5: It would be. 00:10:00 Speaker 1: play Tim Cook in the movie? No, I don't know. Richard Gere. That would be it. And Iraq continue on Tim Cook. 00:10:06 Speaker 2: Yeah, I think, you know, one of the biggest thing is they just did the transition. The CEO is very young. He, Tim said he's going to be chairman for a while. He's going to make sure that he navigates the political headwinds or the political currents both in China and the U.S. I'll be surprised if he takes off the responsibility of ensuring that Apple's in a great hand before leaving Apple. But to be honest, you don't know at this point. 00:10:35 Speaker 1: Can I impress Anurag with my vast knowledge? 00:10:37 Speaker 3: Okay. 00:10:37 Speaker 5: Anurag, I know you're big on Microsoft and them as a holistic Mag7 future. 00:10:43 Speaker 1: So they have Azure Maya, M-A-I-A, I'm probably pronouncing it wrong. 00:10:48 Speaker 5: And they're going to go from Maya 200 out to their newest Maya 300. NVIDIA has that same analog, right? Who's going to make the decision on the next Maya 300 at NVIDIA? 00:11:02 Speaker 2: Yeah, the thing is that I think is the biggest question because as Mandeep was saying, NVIDIA has done a phenomenal job of going from one generation to the other in a very short period of time. And the gap they have with, let's say, Maya. Now, Microsoft has had Maya for years. But to be honest, nobody's rushing out and buying that capacity on that chip. and Microsoft has to be forced to go out and buy Nvidia chips. If we see some shortening of that innovation curve, you could see down the road that people are using training or inference on Microsoft's via chips. We've already seen that with Amazon. Amazon chips are in demand by customers. And Google's obviously far ahead in that game. And they are trying to come up with their own systems by the end of next year. So they are already in the market. So I think the competition is going to get very tough over the. 00:11:53 Speaker 1: Next few years. You say who's ahead. Who's behind? 00:11:58 Speaker 2: I think if we were to do the math right now, NVIDIA being number one in terms of technology, then Google, then Amazon, and then Microsoft. And my math is only driven by what the customer demand is. If the customer is not demanding big Maya workloads all for their inference and training, then I think they are behind. But Microsoft has started to change their own internal first-party apps with Maya chips, which means if you are running Copilot, or for that matter, any other app that runs AI workloads, they are slowly moving it in the background before the customers get comfortable with it. 00:12:35 Speaker 1: Anurag Rana with us, folks. 00:12:36 Speaker 5: Across America and worldwide, the shock of a resignation at NVIDIA by Jensen Wang, an interim CEO announced, the chief financial officer, and a new chair of the board, Harvey C. 00:12:49 Speaker 1: Jones.