英伟达展望强劲推动科技股上涨，亚洲芯片股反应平淡

Bloomberg Audio Studios podcasts radio news. Welcome to the Daybreak Asia podcast. I'm Doug Krer. In the US equity market on Thursday, we had a rally among tech shares following that blockbuster outlook from Nvidia. It was after the bell on Wednesday. The chipmaker said that revenue in the next fiscal year will grow about 70%. So today, shares in Nvidia were up nearly 9% and that gain added some $442 billion to Nvidia's overall market cap.

For a closer look at trading in the Asia-Pacific, I'm joined by Sunonny Banga. Sunonny is in our studios in Sydney. Thank you so much for being here. What are you observing right now when it comes to chip stocks, particularly in South Korea? I know that the Nvidia news broke after the bell in the US on Wednesday. So, Asia had a chance to respond in the Thursday session, but I'm seeing kind of a a lackluster tone right now when it comes to the chip stocks in the APAC region.

That's that's correct. The Nvidia results were quite spectacular, particularly the forward guidance that the CFO gave on the earnings call. Uh unlike in it's very unlike Nvidia, but also just unlu unlike any semiconductor company to give that forward guidance. But the guidance was as much for investors as it was for the supply chain in Asia. It was a it was a discussion I guess to say well there's revenue being missed not just by Nvidia but also its Asian supply chain uh business partners such as SKH High, TSMC and Samsung electronics.

So whilst the stocks did respond well um yesterday uh they did fade into the afternoon and then the momentum hasn't really carried on today. I guess it's sort of opens up more questions. Um, will the Asian supply chain respond and increase capex because it's obviously lost revenue for for everyone in this space.

So, I was struck by news today that SKHENX is looking at ways to deepen its cooperation with the Japanese flash memory maker Kyoka Holdings. Now, I know that SKHix already has a significant indirect state in Kyoka. Can you give me a sense of what the strategy here is? what may be at play.

That's a very interesting observation because a lot of what we're heading down in terms of the road map for a lot of these memory companies is increasing amount of code design um with the partners. So you've seen a lot of co-design between Nvidia and its partners in terms of HBM and stacking and ensuring uh where memor is placed relative to chip but also it's important where NAND is placed relative to GPU and relative to DRAM and so the partnership probably indicates that there needs to be more of a code design roadmap that uh both companies have to embark on going forward. uh and obviously this is an important uh way to improve efficiency for the chip and in terms of improving efficiency for memory makers so that um memory cost can be kept under relative control but also it's it's important that the that the uh western players stay ahead of the game because in China you've got YMTC and you've got CXMT also looking to move into uh HBM memory as well.

So I'm glad that you mentioned a company like CXMT. To what extent will they be able to compete with a company like SKHix right now? Yeah, I mean this is the this is the big question for the memory market because it is uh increasingly a emerging player on the scene and um they've had quite significant success in mobile DRM um and really using now this current tight market in DRAM which is which is leading to great profits across the whole space but also leading to very high profits for CXMT and those profits, you know, we will find out on the earnings call this evening.

Are they now looking to expand capacity in HBM and uh it's it's inevitable that they try because this is obviously the big bottleneck for China getting HBM3 and moving up the the DRM stack. So, um I think you know the the the important thing to watch is that um there is a lot of effort by the Chinese government to support this sector. A lot of the big Chinese tech stocks have all invested in a mutual fund that will allow for uh investments in this space.

So I mean there is a big uh sovereign initiative to to push CXMT into into into HBM and and and compete against the western players.

So what is your sense of the degree to which western players will need to expand production capacity and if that is the case how is that going to be funded? Yeah, that's look I mean if we just take Nvidia's uh forecast uh of 70% and then you know look at what they you know are arguing they could do without the supply chain constraints which isn't you know which is 100% growth versus a 70% that sort of amounts to uh about $120 billion US of of loss revenue for Nvidia.

Um so when you sort of take that through um really what we're saying is there's not enough uh capacity in uh wafers a TSMC there's not enough capacity with with the memory makers the three large memory makers. So look the the the interesting thing is that the cap the capex to se to revenue at a lot of these memory makers and TSMC has actually been falling. So there is a lot there is a lot of scope to rise that these companies are producing records amount of free cash flow.

The memory makers cumulatively will nearly approach a trillion dollars of free cash flow. Now they are committing that to buybacks which which makes total sense. They need to support their share prices but also there's a lot of free cash flow there to redirect to capex. Um but these things take time and you know these uh you know clean rooms take one to two years to start. So I think it's you know TSMC and and the memory makers really are probably going to work closely with Nvidia and and supply chain and and you know in the results Nvidia's made very large pre-commitments to the supply chain um overing $169 billion just for the quarter.

Uh so that's that's an important signal being sent to the supply chain in Asia that Nvidia is opening up its wallet and using its balance sheet to to encourage these companies to expand capacity.

Sunonny, I'm curious about the adoption of AI across the Asia-Pacific. We know what's been happening in the states between Open AI and Anthropic, two of the leading names. We have seen the situation on the mainland in China. Deepseek obviously was the seinal moment. But I'm curious about how AI is being adopted in jurisdictions like Japan, South Korea, Taiwan. Can you shed some light?

Of course. I mean this is uh this part of the world has uh you know through legacy uh you know when we look at different technology cycles uh was a very fast adopter of the internet uh boom. Um in many of these markets uh you know payments is all digital uh they've already bypassing the rails. Um so this technology uh is essentially uh back you know getting backboneed by the existing infrastructure and so the adoption is definitely at a consumer level very high um and there's a lot of uh interest to test this out.

It's where it's probably lagging is at the enterprise level and and and I don't think that's going to be the case for uh forever. I I think just just naturally uh the adoption of claude co-work and um and codeex have been very US you know let's call it western hemisphere at the moment heavy um and and uh as the as those products have been very successful it's it's moved in through the enterprise enterprises of a lot of US companies and uh that is something I think that's going to happen in Asia as well enterprise adoption will will rise and that's a you Asian perspective exchina.

I think what's happening in China is is is quite different. Uh in China, there's obviously it's it's almost like a different universe. You've got this open-source uh ecosystem. Um Alibaba, uh Tencent are really pushing the adoption hard in time in terms of their user base, both consumer and enterprise. And losses are rising. when we saw in Alibaba's results last week, losses for Quinn have risen and that's probably because they're subsidizing adoption.

Um, so I think what we'll