00:00Thank you for taking the time to come in. What were the highlights for you?
00:03Yes, we delivered a very strong set of results late last night.
00:07As you mentioned, group revenue grew 43% year on year,
00:11and we also delivered a significant milestone of both net profit and adjusted net profit
00:18that have solidified our growth into the long-term structural trend.
00:24Particularly, I would like to highlight the growth in our API, which is our token factory.
00:28We saw our revenue grew more than eight times,
00:32and also token demand surged close to six times or more year on year compared to first half last year.
00:40Tell us a little bit more about me.
00:41How do you compete in this whole ever-changing AI race here right now?
00:46Obviously, you have the big traditional tech players that are also kind of developing their own models and the like.
00:51You have peer plays that are making really big waves in terms of user growth and whatnot.
00:57But how do you fit in this whole AI race right now?
01:01It is a very well-asked question.
01:04We've been asked these questions repeatedly.
01:06Since the founding of our time, we committed to being neutral, open-source platform.
01:12We work with all model companies, and we work with all compute companies.
01:16And hence, with that, now our virtual GPU is able to provide more than 200,000 models for our enterprise
01:25customers.
01:28So many of those, what you've mentioned about competitors in the marketplace,
01:32are actually very good, long-term trusted partners of ours that we also service as paying customers.
01:40Token consumption, you brought that up.
01:41Yeah.
01:43I mean, if you chart the number of your growth, 620%, that's year in year, that's almost going to look
01:50exponential.
01:51Is that the proper way of looking at your runway moving forward?
01:57Or at some point, does this plateau to more modest gains?
02:01If we look at structural trend, we have done quite a bit of analysis, dive through.
02:07We look at all the third-party data.
02:09We look at global numbers.
02:10Structurally, we see significant strong growth demand from the demand side of the equation.
02:16For the next few years, particularly going into 2030, we believe that the growth demand will likely to be multiple
02:23folds.
02:23And hence, we are seeing the second half of this year likely to deliver even higher growth versus first half.
02:29And that's likely to continue for the next three years, a few years, mostly driven by a global demand of
02:36AI development, AI innovation,
02:38and more importantly, is the AI deployment into scale.
02:42Okay.
02:45I mean, just tell us about, there's so much debate about how these Chinese AI platforms are making money, right?
02:51Because the fact is, it's low cost, it's open weight, it's very different from what the U.S. is doing,
02:58and the visibility around, really, this ROI is still very debatable right now.
03:02What is your answer to that question when people ask you about the prospects of ROI when it comes to
03:07AI spend in China?
03:09I think I'll answer this in a couple of ways.
03:12One, it's very well understood by the market.
03:16China has highly effective, cost-effective open source models.
03:20That's easy to deploy, that's flexible.
03:23And two, China has a very large amount of cheap, effective, and efficient power supplies,
03:30which is also another very important component for AI deployment.
03:36At least for fancy, if you look at our paying customers,
03:41the number of, amount of paying customers year on year compared to the same period last year has grown more
03:46than five folds.
03:46All those are from large enterprises.
03:50Those are eagerly looking for more effective, scalable AI solutions,
03:55and we are not yet seeing a slow demand in that.
03:59And I think, like mentioned earlier, structurally, the huge demand is likely perceived at least into 2030.
04:05Sure.
04:07Sure.
04:07And just give us an idea of price sensitivity.
04:10People talk about how the China model wins out simply because it's more economical.
04:15Like, your direct conversations with your clients,
04:18does it seem like there is a really, really price-sensitive customer base there?
04:23If I'm a large enterprise that's deploying AI solutions, it's not only about the price.
04:30The quality of AI solutions is also equally important.
04:35So when we look at this equation, we look at both cost, structural,
04:38but also the quality of the token output and also token services.
04:45I think this is where we are significantly well-positioned for it.
04:49A lot of people ask me about CapEx.
04:50So what do you think is going to be your CapEx planning when it comes to your outlook here in
04:56the near term?
04:56And tell us a bit more about just when do you think things are going to be,
05:01I guess, when investors are going to start looking at this a little bit more positively,
05:05the AI spending, the mounting costs it takes to really do this?
05:09As a listed company in particular, we are extremely disciplined
05:13about looking at a growth profile, growth prospect, as well as our CapEx expansion.
05:18For one, when it comes to CapEx demand, we will remain extremely disciplined
05:24and also we look for a variety of funding source channel
05:28that we need to consider about cost structure on our balance sheet
05:31and also the growth profile for our shareholders
05:33and also the growth profile of the quality.
Comments