- 18 hours ago
Category
🗞
NewsTranscript
00:00First of all, just your thoughts on the latest news of the Kimi K3 model.
00:04How is this changing the playing field, including for you?
00:07It is an excellent model. We've tested it. It's quite good.
00:11And we'll be using it in our products.
00:14I think if we just look at the trajectory, the models keep getting better and smarter.
00:20And there's no end in sight.
00:22And that's something I think that underscores the whole improvement.
00:26We also see open source catching up gradually.
00:30With the closed source narrowing the gap.
00:32So it looks like that's what it signifies.
00:35And having an open source model actually is great for enterprises.
00:39Because they want on-premise deployment.
00:42They don't want to send their data to the cloud.
00:44So I think that's probably the bigger news.
00:46There will continue to be good models from the U.S. and China.
00:50But I think the important thing is it's getting so smart.
00:53We've got to learn to use it for the most important tasks.
00:56Your company also has your own foundational model, the E-Series.
01:01So how do you see this relationship with Kimmy?
01:03Is it a competitive or a collaborative relationship?
01:06We think we want to focus our energies on building applications.
01:11And we think with these models getting so good, we're better off building applications on the very, very best model.
01:17And using their intelligence to change the financial statement of companies.
01:22Because the models, you know, three years ago were like an average human.
01:27But now they're smarter than all humans.
01:29So it's such a pity if these models are only applied to simple tasks.
01:34So we are building in my company a product called Boss AI.
01:38It is a product designed to give the CEO the God's eye view over all the data to make all
01:44the right decisions that will move the dial in the financial statement of the company.
01:49So that's where our company has pivoted to since about a year ago.
01:53So that means, am I right to say you're no longer going to invest on fine-tuning and expanding your
01:58financial foundational model?
02:00You're working with other companies and building applications on top of it.
02:03Well, ecosystem is emerging.
02:05And we're very relieved and happy with this choice because our expenses have dropped tremendously over the last year so
02:13that we can get to profitability fairly soon.
02:15I remember you mentioned that you want to be the first highly profitable AI 2.0 company.
02:21Where are you on that milestone?
02:25Our bookings have increased, you know, three to four times from last year.
02:29And we're busily hiring, fulfilling.
02:33And because the Boss AI product is so understandable, right?
02:37Many CEOs think about, I want to do AI transformation.
02:41What should I do?
02:42I want to ask my CIO, ask the IT department, install customer service, do legal, do finance.
02:47But at the end of the day, all they have is a bunch of sugar-coated demonstrations, but no substantial
02:53impact.
02:54So we feel that our Boss AI is the first product of its kind that superpowers the boss.
03:01And it really gives him or her the chance to gain insight, make the right decision, see the risks, and
03:08help the company become more profitable.
03:10I use it every day.
03:12So to the extent our business has developed well, it's thanks to Boss AI that's helped us, taught us how
03:20to sell the product to the customers.
03:22Okay, so some personal experience with your own product.
03:25So what's your timeline to profitability?
03:27It should be soon, sometime next year.
03:30Okay, amazing.
03:30And tell me about your client uptake rate.
03:33What are you foreseeing in terms of the demand pipeline going forward?
03:37Well, we're getting more demand than we can handle because in order for us to do the business, we do
03:44global business.
03:45So we service not just Chinese enterprises, but also global enterprises in Asia, South America, Middle East, and so on.
03:53So our challenge, frankly, is finding English-speaking, Arabic-speaking, and other language-speaking people who are qualified to be
04:02either consultant or FTEs who can fly to those countries and work with their teams to AI-infuse.
04:09So I think the limitation of our business, if there is one, is our ability to hire.
04:16Now, as we get into businesses, we'll see the uptake grow, and also there will be licensed revenue, subscription growth.
04:25So we can definitely foresee a single-digit, sorry, hundreds of percent of growth for the coming years.
04:33But we're not a token factory, so you're not going to see 50 times growth.
04:37Okay, so triple-digit growth still.
04:40We have been hearing that the government might be mulling restrictions on foreign access to advanced AI models in China.
04:48Is this going to impact your business in any way, especially your overseas portion of the business?
04:54I know nothing about the rumor you talk about, but we sell any model legally in any country.
05:02So if there is a Spanish buyer that wants to use a legal model to operate, and they want to
05:12use an American model or a French model, we'll do that.
05:14So we're agnostic, and we support every model from every country, but we operate it legally in the country in
05:21which our product is sold.
05:22Let's talk about some of the export controls, right, with restricted access to the advanced GPUs from NVIDIA's.
05:29How serious is the shortage of GPUs domestically?
05:33How is this, is this constraining the pace of your innovation or development, or is it raising the cost of
05:38your business?
05:38Well, I think globally there's not enough GPUs because the entire market is blossoming.
05:44I think export control was an imaginative but unsuccessful effort to contain China.
05:51It has failed.
05:52So to whatever extent China faces a shortage, it's not from the export control.
05:58It's from either businesses being frugal, or it's about the supply not being sufficient, or it's about businesses choosing not
06:08to buy.
06:09So the export control just simply has not worked.
06:13And given the sort of two-track tech ecosystem, in the U.S., it's been the leaps in the hardware.
06:19In China, it's been mainly the innovation on the software side.
06:23Where do you think China and U.S. stand in this race, maybe in three to five years?
06:28Well, I think in models, U.S. will make a lot more money, but China will have a larger market
06:33share
06:33because of open source being a powerful distribution and on-premise deployment being such an important thing to enterprises.
06:41On applications, I think U.S. will have more profitable enterprise applications.
06:46But when it comes to sort of the heavy lifting, like what we do and what companies like Palantir does,
06:53we actually think the Chinese flexibility, willing to work hard, willing to fly anywhere to help our customers be successful,
07:01is going to be a specialty that makes us stand out in this category of enterprise AI transformation.
07:08In consumer software, both sides will come up with ideas, but the Chinese large companies, like Tencent, ByteDance,
07:15will clearly beat the American companies in coming up with clever AI applications.
07:21And then in hardware, China is ahead, as we know.
07:24So you talked about how the U.S. might have more profits and China might have more market share.
07:30Then therein lies the question around monetization.
07:32Given that many of these Chinese AI models are priced at a fraction of U.S. cost, is this sustainable?
07:38Well, history has told us if you can win the hearts of users, even if it's a free product,
07:44there will be ways to figure out how to monetize later.
07:47We've seen that with internet applications, with websites early on, with social media, with products like TikTok.
07:56So there will be ways.
07:59If the companies continue to care about users, build good products, and win the hearts and minds of the users
08:07and developers,
08:08I think money will be made in the future.
08:11But in the short to medium term, the enterprise money favors the American approach.
08:16Going back to your company, I know you said you are reaching profitability very soon.
08:21What are your fundraising plans, your timeline to an IPO?
08:25We hope very soon also.
08:27But a lot of more work has to be done to be eligible for filing for IPO.
08:33But I think our financial numbers are ready right now.
08:37But it's a matter of going through the process.
08:40Is it going to happen in 2027?
08:42I hope so.
08:44All right.
08:45Tell me as well, when you talk about selling your boss AI product,
08:51how have you been able to compete with the suite of enterprise solutions out there?
08:57You have benchmarked yourself against Palantir in the past.
09:00What is the unique selling point or advantage that you have against the competitive lead scheme here?
09:05Okay, so we sell globally.
09:08We don't sell much to the U.S. because the market is not so receptive.
09:12But other than U.S., we sell globally.
09:15I think Palantir and we have the only two products that are good enough to do AI transformation
09:21and do it from the top and show results within two to three months.
09:25So we're very proud to be in that company.
09:28We are very respectful of Palantir, who is a pioneer in the space.
09:32And we're very thankful to them because they don't come to the countries where we have the most business.
09:38They choose to focus on U.S. and American allies.
09:41And those are exactly the countries we have trouble selling to.
09:45So we have a two-player horse race, a two-horse race,
09:49where the other horse chooses not to run in the tracks that we're entering.
09:54So the big question then, if these U.S. players start to look overseas to those markets,
10:00and you're, of course, constrained in the U.S. market because of geopolitical headwinds,
10:04I mean, how do you see things playing out?
10:07Well, actually, I think the approach is, while the products are similar, the approach is quite different.
10:13Palantir chooses to do a lot of defense military.
10:16We don't touch that.
10:18And also, landing in Asia, for example, is going to be a difficult proposition for Palantir, even if they choose
10:27to.
10:28And also, their guiding principles chooses not to come to these countries.
10:32So all these reasons combined, I think it will be unlikely.
10:36Also, they're doing extremely well, you know, if I were Alex Karp, for the same, you know, 100-person investment,
10:45he'll make so much more money going to U.S., Europe, and Japan.
10:48I can't imagine why he would consider coming to our countries.
10:53But I also think there is a first-mover advantage,
10:56and we're already gaining that in many of China and the Belt and Road countries.
11:00So when that happens, I think it will become a very defensible mode.
11:05I want to ask your take on where do you think we are in the AI development phase,
11:10especially in China, because there are now concerns about whether or not we are in a boom,
11:15you know, whether we are in a very unhealthy period of price competition.
11:19Where do you think things stand?
11:21Well, I think what's really going well for China is government is hugely supportive,
11:26and we're going to see companies that are worthy of going public go public.
11:30and raise the money they need to deliver their business goals.
11:36And we're also seeing a phenomenal consumer and enterprise enthusiasm about AI.
11:42I mean, you see the show here, everyone's so excited.
11:46And also the poll in China shows 84% of the people think AI is a great thing,
11:52whereas in the U.S. it's under 50%.
11:54So I think, you know, customer-consumer optimism is important.
11:59Having the technology is important.
12:02Having a supportive government policy is important.
12:04We do have these.
12:06That said, there are some issues with enterprises' willingness to pay substantial fees for AI.
12:13That might be a challenge.
12:15And other than that, I think everything is going to go extremely well for China
12:20with an unassailable position of at least number two in AI
12:26and possibly number one in many of the key areas.
12:30Thank you for listening.
12:30Thank you so much.
Comments