Skip to playerSkip to main content
  • 5 days ago
Transcript
00:00I think our group feel excited today but it's also just a normal day.
00:04We will invest all the resources to achieve our mission.
00:08We set up from day one which is intelligence with everyone.
00:12Also it's our understanding about AGI. So what's intelligence?
00:15We want to build, to bring the world-class technology foundation models to all the world.
00:21So in the past almost four years, every single year we make like technology breakthroughs.
00:27We deliver world-class models in each modality to the world and what's with intelligence.
00:34So we hope all the world users, the enterprise could use our models.
00:39So currently we are serving around 200 million users across all the world
00:44and more than 100k enterprise customers from more than 100 countries are using our model.
00:50So how do you scale your multi-modal capabilities without being profitable
00:56and without having a large capital base unlike an open AI or anthropic and others?
01:02You're a open source lower capital model.
01:06How do you scale those multi-modal capabilities not only in China but abroad as well?
01:12I think I have been working in AI for around like 10 years.
01:15I think the beauty of foundation model is like a general purpose use model.
01:20So you make one model and different kinds of use cases, no matter if it uses our enterprise customers,
01:26you are using your models.
01:27So we put most of the resources on the models.
01:30If the users, they think your model has the best performance in their use cases,
01:34they just directly use it.
01:36As you said, so in the past around like four years, we really focused on the capital efficiency,
01:42cost efficiency. So we had only spent around 500 million USD in total.
01:49Probably only one or two percent of the big gens spending.
01:52So we did all the optimization, made creativity, made innovations on it.
01:58How do you compete though if you're going to go global?
02:00How do you compete with an open AI or others that are throwing lots of capital at their business?
02:07Is it viable? And what is the technology gap looking like in the next few years?
02:12Actually, I don't think it's a competition. It's more about all the top talent across all the world.
02:19They bring the technology breakthrough to the society, to the end users.
02:23So you will see each model has pros and cons.
02:26So we focus more resources on building the models, building the best product experience
02:32to get more and more users from all the world to use it.
02:35Now your proceeds have been stated they're going to be used for research and development for other
02:40models over the next five years or so. What are you going to be your chip needs?
02:45We know that there's a bifurcation in the chip world. We're hearing as well the Chinese government
02:50might approve NVIDIA's H200 for China in this quarter perhaps. Is that something you need to look at,
02:59importing more advanced accelerators for your training and the like?
03:03I think in every industry with a really high growth, when the new technology comes here,
03:08so you will see every startup and all the companies face resources constraints. Probably cheap
03:14constraints, computing constraints is only one of the challenges. So in the past you will see our
03:20model performers and our product performers. So our resources are stable, we can get access to all of
03:26them. So we focus more on our R&D and foundation model. So we just use whatever chips that makes
03:33sense for us at that specific stage. Is your goal to stay indigenously with your chips, locally made
03:40chips, or would you buy NVIDIA chips as well? Actually, we have a really high capital efficiency and we
03:49really focus on cost efficiency. So it doesn't matter what kinds of chips it is. It's more about
03:56which chips can give us best ROI, which can help us to achieve our mission to make the best technology
04:03to accessible to all of the users across the world. So when does the emphasis turn from efficiency
04:10to profitability and also getting that steady revenue growth that will then justify a higher
04:17stock price valuation? So we put lots of R&D resources and innovations on the efficiency part.
04:24You will see our API, we have the enterprise business, you will see the gross profit margin of
04:30our API is more than 65 percent, which is probably already one of the highest globally. So we focus more
04:37on make the best technology and the best user experience that people will love to pay for the best
04:43performers models for their use cases. Now, speaking of the different models and you have multi-model
04:50capabilities, you've also come out of a price war, essentially, a race to the bottom, some have
04:57called, because you're trying to keep costs down and you want to survive as well. There was a battle
05:03of survival and you have been one of the survivors. But at the same time, the startups are not the only
05:09ones, the big players, Alibaba, ByteDance, Tencent and others. How do you compete with them at a low-cost basis?
05:17Actually, it's not, I don't think that's a press word. It's more about the performers competition.
05:25So you need to make your model best and the people, I mean, end user or enterprise customers,
05:30they don't choose your model because of it's cheap or expensive. It's more about its performers.
05:36So that's the key. As you said, I think it's not about the competition between like all these big
05:42gens, but it's more about some of them, they are like our cloud providers and some of their products,
05:47they are using other models. So it's more about the collaborations. All this company brings the
05:53foundation models to all the world. So where are going to be your more feasible areas to expand abroad?
06:01Because again, you know, going overseas is going to be a big challenge, isn't it?
06:07Actually, we set our strategy going global, grow up global from day one. I studied in the U.S. I had
06:16been working there. I'm working here right now. So I think the beauty of foundation model is just to
06:21build one model. All the world in different languages, they just use this model. Yeah,
06:26the same model. So focus on your model, focus on your product. It's the key. Yeah,
06:31there is no boundaries among like different regions.
Be the first to comment
Add your comment

Recommended