00:00So we heard from US chip giants like AMD's Lisa Su and NVIDIA's Jensen-Huans,
00:05a head of Computex. Could you share the key takeaways from Intel CEO Libbuton's keynote speech today?
00:10That Intel is coming back, especially because of the 18A process node,
00:15which is kind of comparable of the TSMC N2P and N3P process node.
00:23As you remember, Intel was always lagging behind and had huge challenges, you know, with different CEOs.
00:28And now Libu is in charge. And he has a very engineer mind.
00:34And I have to admit that the way how Libu is structuring Intel, like the new Intel,
00:40with this engineering focus, and also not only for the lithography,
00:45so when you talk about the process nodes, also in advanced packaging.
00:51Advanced packaging is one of the key technologies today if you look into computing and if you look in AI.
00:59Intel has there some really good products comparing, like you have Cobos with TSMC,
01:05and so Intel has their own solutions. There is also a good fit into the ecosystem.
01:11So my feeling is they are coming back. So they are coming back into the game,
01:16and they are trying to basically also revive their legacy into the market because the AI market is so big
01:23today.
01:24So NVIDIA has announced a new AI chip designed for Windows laptops.
01:28How can companies like Intel and AMD compete? And is it a zero-sum game moving forward into the AI
01:35PC era?
01:35So now with NVIDIA entering the game, they come from a different angle.
01:40They come from the GPU angle into compute.
01:42And we need to see at which price point these notebooks or workstations will come out.
01:49Because we have to understand this is not for entry-level, not even for mid-level.
01:54So I normally would assume that this is more like on the higher-end level.
01:58So for people who really need the power to have kind of an offline AI and also for more privacy.
02:05There will always be customers or consumers. There are some discussions in the industry.
02:10It should be between two and a half thousand to three thousand dollars.
02:13But that is a lot of money for just a normal notebook.
02:16So I think we want to see how this filter evolves.
02:19Lately, we have a shortage on CPUs.
02:23And the shortage on CPUs comes when you look at the AI.
02:26You have large language models. That is basically when you train.
02:29But later when you use the models, you do inference.
02:33So you do queries and you are asking.
02:35And there you need more CPU power, traditional CPU power.
02:40And that is why the CPU makers like a traditional one like Intel or AMD
02:45also see a much bigger demand now than maybe in the last two years.
02:49So this market is really exploding so wide that we are not at the moment in a situation
02:55that you are biting market share away.
02:58So we see many tech giants coming to Taiwan during Computex.
03:01Many looking to secure the supply chains.
03:04But how can Taiwan manage the demand in the sector?
03:07Taiwan, you continuously build new factories, right?
03:10So there is an expansion.
03:12I mean, if you look at Narnia building a new memory fab, right?
03:15Winbon building a new fab in Kaohsiung.
03:17And then TSMC building several fabs here the next two or three years.
03:21But people have to understand that also all the data from all the other factories,
03:25they come together here in Taiwan.
03:27Because this is the core value of TSMC to understand yield and manufacturing process.
03:35So I think for Taiwan, Taiwan is right now at the epicenter of AI hardware.
03:43The hardware, definitely we are here, I would say, kind of the leader in the world.
03:48Because if you think about Quanta, Compa, Foxconn, Vistron,
03:53these are all next level infrastructure companies who make the racks.
03:58Of course they have factories around the world in different areas,
04:02but they are all Taiwanese companies.
04:04So I think Taiwan has always a foot in the door and you cannot push them out.
04:10They are a vital part. Without Taiwan there is no AI.
Comments