00:00We just heard from Anthony talking about how Meta's Muse, for example, has led to that huge rally in the
00:05AI trade overnight.
00:07We're seeing everything in today's session rallying around technology.
00:11When it comes to the rise of AI agents, those can infer, that can execute, that can continuously surge.
00:19How does that change the demand picture and your business at VeriSilicon?
00:29Yes, the first half revenue almost doubled compared with the same period last year.
00:36Right now, we have 1.8 booking on the first half.
00:44The newly booking was 2.2 billion.
00:50So, of course, they're mostly from the training, cloud, but as you mentioned, it's quite important to consider age and
01:00device.
01:01Because those are the place to make money.
01:03Usually, I say, the training data center, like a trunk, we need a trunk, like a tree, but we also
01:11need a branch and a leaf.
01:13So, fine-tuning, like for the specific domain, it's like a branch.
01:20They're growing leaf, like an inference.
01:23So, as you correctly pointed out, those are the place to close the loop, such as when Internet days, we
01:30need a cell phone, right?
01:32So, this is, you correctly mentioned, the key issues.
01:35These days, people talk more about the inference, rather than a few years ago, most people talk about training.
01:47Yeah, that's really interesting.
01:49You have ARPC.
01:51Yeah, yeah.
01:53Go on, go on, go on.
01:57Yes, I think that the existing market, like a cell phone, PC, if it's not AR cell phone, people may
02:05not change the phone or change the laptop.
02:08But I think most importantly, emerging market, AR glass, even AR toys, and AR rings.
02:16You see, I have this ring, no display, no camera, right?
02:23But it costs $400, $500, and they went to IPO.
02:28So, because health, right?
02:30The amount of health continues very conveniently than the watch.
02:33And the reason that we develop algorithms can measure blood pressure, kind of calibrate, actually, unless they tell you to
02:41measure, you don't need to measure the blood pressure.
02:44Because they monitor.
02:45So, wait.
02:45So, those, the age device is really important.
02:49Yeah.
02:50I think we're having a little bit of difficulty because of the noise over where you are and the technical
02:55lag.
02:55But let's keep this conversation a little bit shorter and talk a little bit about your business itself.
03:00Because you've talked about listing in Hong Kong that will give you more global exposure.
03:04Given right now the political situation, how feasible do you expect to be the reality out there of trying to
03:12get Chinese chips into the world?
03:16I think the export Chinese chip, actually, you can tell it's growing rapidly, right?
03:23Most problem age device.
03:26I think very silicon, 30% revenue from outside China, most from U.S.
03:34Although 98% are in China, right?
03:40So, this is, we always keep the things more international, more global.
03:48We have offices worldwide, we have Germany, France, right?
03:53We have offices in Tokyo, Korea, whatever.
03:56So, in addition to U.S.
03:58So, I think this company almost IPO in Nasdaq 2015, later when China opens a new market and will come
04:07back.
04:08But I think the majority, right now, significant customers still, even given the current geopolitical situation, still outside China, 30%.
04:20So, Wayne, I mean, China is also leading to this new form of AI economics where you have cheaper, open
04:28-weighted models as well.
04:29What is the strength of very silicon in this ecosystem?
04:34And do you expect the Chinese sort of AI setup to be actually available overseas to be exported worldwide,
04:44especially when it comes to smaller emerging market economies that have to choose, perhaps, a way to go when it
04:51comes to artificial intelligence development, West versus China?
04:57Yes, we used to be called the leading IP provider, worldwide leading.
05:05But now, we call it AI-AC, AI-AC.
05:10So, you know, very silicon, no product, no branding, of course, no inventory, no boundaries, right?
05:18So, we are custom design for our customer.
05:21It could be chip company or could be non-chip company.
05:25So, we just prep platforms, right?
05:29So, I always say the product is the first layer.
05:33The higher layer actually is platform or service.
05:37Then, another higher, it will be ecosystem.
05:40Highest will be the standards, right?
05:43Not everyone needs a product.
05:45So, very silicon, perfect example, grow so fast without any product.
05:49So, that's the enabling technology.
05:52We have the six major processes, including GPU, AI, whatever.
05:58But also, in addition, we have 1,700 mixed signals, I, interconnect, all those IPs.
06:10So, but those IPs used by other, our customers, the chip is belong to our customers.
06:18But we have to provide turnkey service, means we have the chip, manufactured, tested, right?
06:26Manufactured, tested, right?
06:27Manufactured, tested the chip to our customer, one by one.
06:29So, this is called turnkey service.
06:32dort.
06:32So.
06:32Not giving off questions.
06:32We are still waiting for a viewer.
06:32We are totally annoyed, keeping it up.
06:32We try deeply and eagerly to come out.
06:32You sit against them.
Comments