China’s New AI Model, Tap-to-Pay Troubles, and Underdog Plays the Field 7/17/26

**SPEAKER_1** (0:00)
Did you know that NeuroSymbolic AI can help your business find new ways to make money? EY Parthenon is the only one offering this groundbreaking growth platform. NeuroSymbolic AI analyzes hundreds of millions of data points to reveal new growth strategies. EY Parthenon teams deploy this innovation so you can uncover hidden value and scale beyond existing boundaries. Want to know what NeuroSymbolic AI can do for your business? Contact EY Parthenon today. Solutions that work in practice, not just on paper.

**SPEAKER_2** (0:30)
This episode is brought to you by Schwab Market Update, an original podcast from Charles Schwab. Join host Keith Lansford for this information-packed daily market preview delivered in 10 minutes or less, including projected stock updates, monetary policy decisions, and key results and statistics that may impact your trading. Download the latest episode and subscribe at schwab.com/marketupdatepodcast, or find Schwab Market Update wherever you get your podcasts.

**Gabriella Von Rouge** (0:58)
You're listening to The Exchange. Here's today's show.

**Contessa Brewer** (1:04)
Dom, thank you and thank you for joining us here on The Exchange. I'm Contessa Brewery and for Kelly Evans today. Stocks are off the lows this hour, but still we're looking sharply down on these new AI jitters. This is courtesy of China and that ongoing weakness in the Semi's trade. The SOC Semi's ETF was down as much as 3% earlier today, but now just lower slightly. It's still down 8% for the week and on track for the worst week since last April's tariff downswing. And look at the Mag 7 now, also broadly lower. You've got Meta as our biggest decliner now down more than 2%.
There are new reports out today, it may lease computing power to Anthropic in a potential $10 billion deal. And take a look at SpaceX. We're seeing it back now down to 126 That's way below its IPO price of 135 It follows a second scrapped launch of the Starship rocket and continued growth in short interest. Let's begin here though with the latest news adding pressure to the AI trade, which is a tiny startup releasing a new model to challenge Anthropic and OpenAI. Kate Rooney is looking at the impact here and Eunice Yoon is live in Shanghai with the very latest. Eunice, we begin with you.

**Eunice Yoon** (2:20)
Thanks so much, Contessa. Well, that startup is called Moonshot and it released the world's largest open weight AI model. So that means that anyone can download it and customize it. And in terms of the size, it's 2.8 trillion parameters. So that means that it's basically trained on a lot of input data and in theory, it means that the model is better. Another big feature of the company says it has 1 million token context windows. So that just means that when you use it, you can put a lot of text and other stuff into the window in order to ask it complicated questions. So because of that, there are third-party benchmarks who have rated it very highly. Arena AI ranks it first for building web interfaces. Moonshot itself says that this AI model called KimiK3 is comparable to models from Anthropic, OpenAI.
It says that it outperforms its Chinese rivals like ZAI on coding tasks, and that affected the stock prices of some of those Chinese companies in Hong Kong today. The release also coincided with a big AI conference here. President Xi was here for a keynote speech. He was delivering a message about how the open weight model and that open source in general is a much better way to go as compared to the US. In fact, he took a veiled swipe at the US model, saying that the proprietary AI models are really not necessarily something to go by. He said that AI development should not be a solar performance. And he also praised what's seen here as China's version of the US Pax Silica Initiative. So this is called the WAICO. It's a group launched last year but didn't see much membership until this week when 29 countries including Russia, Brazil, Serbia, from Europe joined. US did not.
And also it's all about sharing technology as well as expertise in AI and ultimately with the goal to be able to promote global AI rules and standards. So this is the first time that President Xi had attended this conference, and it was also the first time when he addressed AI safety concerns, saying that the AI must be secure and controllable. So a really very interesting contest. I think though that an important point to make is that when China says openness, it's not always very clear by what they mean because at this conference, we did not see Anthropic, we did not see OpenAI or any of the major US labs. And also those same labs generally are not allowed to operate here.