- 6 hours ago
Category
🗞
NewsTranscript
00:00Tech stocks tumbled again this week over concerns about AI spending and semiconductor supply,
00:04but it was the surprise announcement from a Chinese startup drawing parallels to last year's
00:09Deep Seek moment. That's right. Joining us more is Mike Shepard, Bloomberg's senior editor for
00:13technology and strategic industries, excuse me, and Bloomberg's China correspondent Min Min Lo.
00:18Thank you both so much for joining us. Min Min, I want to start with you. What is this announcement?
00:23What was the reaction to this moonshot AI? And you were over there with all these tech people.
00:29What was the energy like?
00:33The energy was really insane. It was so crowded today. And mind you, it was a very, very hot day,
00:40but it's drawn a huge amount of crowd at this World AI Conference in Shanghai. I went to Kimi's booth.
00:47I mean, Moonshot AI's booth, that's the Chinese startup that was behind the new Kimi K3 model.
00:52They just launched that model. It is the largest ever in China, 2.8 trillion parameters. And the
00:59company says that it outperforms all rivals except for Anthropik's latest Fable 5, as well as OpenAI's
01:07ChatGPT 5.6. So it's really, I think, taken the world by surprise because of how fast
01:15China's AI model is closing the gap with the US, despite not being able to access some of these
01:21advanced chips from NVIDIA. And that's why we saw a huge market reaction. The sell-off that we saw
01:27across Asia, the cost fee is down more than 6% because investors are now really reassessing
01:33whether or not all that CapEx by Samsung and SK Hynix, $600 over billion into new factories. Is that
01:40really justified, given that China is doing it better and cheaper?
01:45Mike Shepard, I learned a new word this week. It's distillation. I heard it a lot at that Aspen
01:49Security Forum. And I guess it's kind of code for stealing or using models that are in existence,
01:54kind of replicating them, building on them. And there are a lot of accusations, I guess,
01:57that that's what's happening with a lot of these Chinese AI companies. Talk a bit about that and kind
02:01of give us your sense. Put this into context, if you would, sort of where this new model stands in
02:05contrast with the ones that we talk about so much here that are made in the US.
02:08Well, you've brought up an issue, David, that is increasingly a sore spot between Silicon Valley
02:14and China. And the leading AI companies here, OpenAI and Anthropic, have increasingly beaten the drum
02:22for policymakers to hear that they think that some of the results that their models are generating
02:28are being extracted en masse in an unfair and improper fashion, a way that violates their terms of
02:35usage by Chinese startups, and then being used to develop a rival generation of chatbots. They
02:43haven't pointed the finger specifically at Kimi K3, but they have recently, and we reported just a few
02:50weeks ago, that Anthropic itself was pointing the finger at Alibaba, accusing it of a large-scale
02:56campaign of distillation. Now, to be clear, this is a technique that is commonly accepted when it comes
03:03to AI when you have a larger model used to build a smaller model, a larger frontier model whose data
03:11are extracted and results are extracted to produce something that's a little smaller, more focused,
03:17and less advanced. But the claim from Silicon Valley is that, look, what's happening is that all of the
03:22physical investments that we're making and spending all these billions of dollars on are being instead
03:30used by Chinese competitors to build this rival generation. Now, earlier today, we did hear from
03:36the Chinese deputy foreign minister rejecting those accusations. It was really the first and the most
03:42prominent public statement from China's government with respect to that. But we can expect to see this
03:48come up between Presidents Donald Trump and Xi Jinping if their planned meeting in September goes ahead.
03:54Yeah, the timing of this, Mike, is really interesting. I mean, I'm assuming this is all related. This is
03:59why we saw this huge dip in chip stocks when this announcement came out. Is that just about this
04:03model, or does that kind of compound on some of the concerns about this cap-ex investment? Especially,
04:09it seems to me, when you look at what China can bring to the table if they have the technology
04:12side,
04:13here in the U.S., we focus so much on the energy aspect, on building these data centers, on whether
04:17or not
04:17they can support them. It seems a lot easier to do coming out of a top-down system like Beijing.
04:22Is that part of
04:23this concern as well? It certainly is. And what you're seeing really is a return of that crisis
04:29of confidence in whether the AI trade will really sustain itself. And can all of the investments that
04:36these companies are making in what could be trillions of dollars in infrastructure for data
04:42centers and chips and other materials, will they keep going? And can they justify it? And it's harder
04:50to justify if there are cheaper and just as good models out there being offered, and also models
04:56that can be downloaded. One of the differences that you see with OpenAI and Anthropic is their models
05:03are generally accessed through a website, through an API, at an enterprise level. With the Chinese
05:11approach, you've seen these startups use what's called open weight. That means you can just download it,
05:16modify it, adjust the parameters, and really tailor it to your needs. It's a lot harder to do that with
05:23some of the frontier models being developed by the U.S. AI labs right now. And that is, again,
05:29something that people are also questioning, the interoperability and functionality of what the U.S.
05:35labs are producing. And then there's the cost. And Min Min pointed to the fact that many of these
05:40Chinese companies are now offering their services at a much lower rate. And the companies here in the U.S.
05:46that are trying to buy these services are starting to look at the bottom line and thinking, hey, all
05:51this AI that we're telling our employees to use, it's starting to cost us. So that is, again, another
05:58factor in this. And it is causing this big downdraft across the markets here.
06:03We compare these models. We compare the kind of technological advancements. It's also interesting to
06:07look at the contrast between the U.S.'s approach and China's approach. And that's something I talked about
06:11with Elizabeth Economy, who's a fellow at the Hoover Institution at Stanford University a couple of days ago.
06:15We talked about just sort of the Chinese strategy compared to the U.S. strategy. Let's take a
06:18listen. Xi Jinping has a grand vision for how they're going to deploy AI in every sector of
06:24economy, in the educational system, in the logistics system, certainly in the military. I don't feel as
06:28though we have that vision of taking all of our incredible innovative work, you know, the most
06:35advanced models in the world. But how are we going to make them practicable for our economy? How are they
06:42going to be used? And again, in that transition, how do we make it not destabilizing?
06:47Min Min Lo, let me turn to you on this first, if I could. I'm curious sort of if you
06:50could expand on
06:51that a little bit. You heard from President Xi yesterday at this conference. Talk a bit about
06:56kind of the way that the Chinese government is embracing advancements in AI and thinking about
07:00it at kind of a larger strategic level. Yeah, it's definitely an important priority for the
07:06Chinese government. In fact, they are grooming this entire AI sector as a next big export engine
07:12for China as the country transitions away from older industries like property, which is really
07:19in a big slump for several years now. So they are pouring a lot of investments to provide the patient
07:26capital to groom all these startups in the sector. And it's all part of the five-year plan that China
07:31had laid out early this year, right, in March at the two sessions. They have a AI Plus plan where
07:36they want to encourage adoption of AI across all sectors of the economy over the next decade or so.
07:42And just to give you a sense of how important AI is, this World AI Conference in Shanghai is in
07:48the
07:48ninth edition. And I remember in the early years, not many journalists would even bother attending,
07:54but now it's a completely different story. Even President Xi himself came to town to grace the
07:59event personally. And this really represents an upgrade to the status of this event, which in
08:04previous, in the last two years, it was attended by Premier Li, Li Qiang, who is the number two leader
08:10in
08:10China. And President Xi gave this big speech this year. A lot of it was, of course, also a veiled
08:15message
08:15to the US as well, because he talked about how when it comes to AI governance, the world needs to
08:21take a
08:21multilateral approach. The world needs to learn from one another. And AI development shouldn't be a solo
08:27development of one country. And all of the Chinese state media, you look at the Communist Party
08:32mouthpiece People's Daily, they have been publishing editorials about how tech and AI should not become
08:37a monopoly. It should really be treated as a global public good that is shared with the rest of the
08:42world. So again, President Xi positioning himself as this leader of the global south, advocating for
08:48bridging the digital divide. So he announced a few key initiatives, including offering at least
08:535000 training opportunities in AI for people in the developed countries and allowing them some 30
09:00countries to tap into this AI powered meteorological system to safeguard their homes, for example. So
09:05really, and also, yes, a couple of days ago, there were some 29 countries who signed up to be part
09:11of
09:11this China led group called the World AI Cooperation Organization. And there's just yet another platform
09:17that could give China more sway in shaping global AI norms. Mike, how does this impact American
09:23companies? I mean, obviously, they're already down the road with a lot of these investments.
09:27Is there they can't claw them back at this point, but they have to be looking at this and being
09:31very
09:31concerned? Is there a way for them to pivot? And then you also have the Trump administration starting
09:35to talk about having a new regulatory agency specifically for AI. Does that compound the difficulties
09:41American AI companies could now face? Well, take the second part first there, Christina, because it's
09:48the latest and there was a scoop from last night from our team here in DC. And the companies actually
09:55would welcome some sort of a coherent approach to AI safety reviews. And at issue is, you know,
10:02they did not like the way that Anthropix, Fable 5 and Mythos 5 were essentially ordered off the market
10:10via export controls over concerns that the guardrails on them weren't adequate. And we saw
10:16the same sort of pressure put on open AI. But the companies had felt that it was overboard and also
10:22kind of ad hoc in the approach. And they would like to see a more coherent framework. And what's taking
10:27shape is, you know, a FINRA-like model, one that would resemble what the broker-dealer industry on Wall
10:35Street faces, that it's an industry funded and not industry run. But industry certainly has a lot of
10:43input to how the rules of the road are set. And in this case, it would be the safety testing.
10:49And that
10:50would give them a little bit more of, I guess, a clear path forward when it comes to dealing with
10:56the government and how far they can take their models as they develop them. More broadly, though,
11:02the question of how companies are reading this moment, look, many of them are playing the long
11:07game. And all this week we heard from TSMC, they were reporting, look, we are seeing demand for AI
11:15chips and AR hardware going through 2030. We are, it is going to be hard to meet. And that is
11:21a message
11:21that was echoed by SK Hynix during Ed Ludlow's interview with the CEO just last, just a little more
11:30than a week ago, as they were making their public market debut here in the U.S. And it's a
11:36message
11:36that's been reinforced by NVIDIA Chief Executive Officer Jensen Wang. They just see the demand from
11:42hyperscalers here in the U.S. and even elsewhere as being off the charts. And NVIDIA also caught a bit
11:48of a break, and some of the other American companies caught a break, too, with the declassification of
11:56the United Arab Emirates, which is trying to establish itself as a real center of development
12:01for AI. And they see that as a new market that they can break into and try to export the
12:06American
12:07tech stack. But really, you do see this contrast of visions of Xi Jinping laying out something that is
12:13very multilateral, and President Donald Trump arguing one that is very much America first. And it really
12:19doesn't have all the pieces together in the same way that you do see China putting on the table now.
Comments