Beijing Weighs Curbs on the AI Models Americans Love
WSJ What’s News
July 9, 2026
A.M. Edition for July 9. Maine’s populist Senate candidate Graham Platner drops out of the race under heavy pressure from Democrats. Plus, WSJ tech reporter Sam Schechner explains why China is considering restrictions on the homegrown AI apps U.S. companies are addicted to.
Speakers Daniel Bach, Sam Schechner, Graham Platner, Aaron Zittner, Brian Gormley, Stephanie Davis
TopicsDaily NewsNews
SPEAKER_1 (0:00)
This podcast is brought to you by ReliaQuest. Cyber criminals are constantly attacking. They want your data, they want your identity, they want your innovation. ReliaQuest fortifies your business with agentic defense, AI that detects, contains, and eliminates cyber threats in minutes. It helps your security team move faster at the work that matters most to protect the business now and delivers insights to help them predict what's next. ReliaQuest, agentic defense for the enterprise. Learn more at reliaquest.com.
That's reliaquest.com.
Daniel Bach (0:33)
Maine's populist Senate candidate, Graham Platner, drops out of the race under heavy pressure from Democrats. Plus, China considers restrictions on the homegrown AI apps US companies are addicted to.
Sam Schechner (0:44)
American businesses have been adopting them because they cost a lot less to run. But now, Beijing's concerned that sharing some of this technology could help adversaries or malicious actors against China.
Daniel Bach (0:58)
And fears of an AI wealth boom in San Francisco fires up the housing market.
It's Thursday, July 9th. I'm Daniel Bach for The Wall Street Journal, filling in for Luke Vargas. And here is the AM edition of What's News, the top headlines and business stories moving your world today.
Graham Platner (1:17)
We are suspending campaign operations.
Daniel Bach (1:20)
Maine Democratic Senate candidate, Graham Platner, has ended his campaign following a sexual assault allegation which he fiercely denies.
Graham Platner (1:27)
This is incredibly difficult because I know that some will think it's an admission of guilt and it most certainly is not. We're not doing it because of the allegations. We're doing it because of the structures that are being taken away from us by those in power.
Daniel Bach (1:41)
In an 11-minute video, Platner blamed the media and party establishment for forcing him out of the race.
His exit leaves state Democrats scrambling to find a replacement to challenge Republican incumbent Susan Collins on the November ballot. Journal editor Aaron Zittner says that losing this pivotal seat would complicate the Democrats' path to winning back control of the Senate.
Aaron Zittner (2:01)
The collapse of the Platner campaign is a big problem for the Democratic Party. Really any path to taking a majority in the Senate next year runs through Maine. That was supposed to be the easiest of the four seats they need to gain in order to win a Senate majority. Under state law, the process for picking a new nominee falls to the Democratic state committee, but we don't know yet what process they will use to select a nominee. We only know that they need to do it quickly. Under state law, they have until July 27 to name a new person and put that person's name on the ballot.
Daniel Bach (2:33)
If the Democrats lose Maine but hope to win back majority in the Senate, they'd have to win several states such as Texas, Ohio, and Iowa, normally considered solidly Republican. And we'll hear from Aaron later on in the show about how political polarization has had an impact on America's faith in democracy.
If there's one common thread of the AI boom across corporate America right now, it's an addiction for cheaper Chinese models. In many cases, particularly in Silicon Valley, companies are saving money by using DeepSeek or Moonshot AI instead of models made by OpenAI and Anthropic.
Sam Schechner (3:10)
Until recently, China and the officials there wanted to encourage the rapid spread of Chinese AI models, sort of saw it as a form of soft power. And that's been working. Some of these models perform close to the levels of proprietary American ones. And American businesses have been adopting them because they cost a lot less to run. But now, Beijing's concerned that sharing some of this technology could help adversaries or malicious actors against China.
Daniel Bach (3:40)
That's Journal Tech reporter Sam Schechner. He says Beijing is now considering tightening its grip on homegrown technology.
Sam Schechner (3:46)
What we've seen is that there's a brewing price war around AI models. Tokens are expensive. And using the top tier models from Anthropic or OpenAI is costing companies lots of money. So what many of them are doing is using switchers or different technologies that kind of root different queries that they may be using internally in business processes where the complex ones, the higher end orchestrating queries might go to an American model that's expensive and more routine ones or submodels running kind of parts of tasks will be rooted to cheaper Chinese models.
Daniel Bach (4:23)
At the same time, China said it has found what it called security backdoor vulnerabilities in several versions of Anthropic's popular cloud code tool released between April and June. A government-run cybersecurity platform said a built-in monitoring mechanism could pose a serious threat and advise users to update or uninstall the software.
9 more minutes of transcript below
Thousands of transcripts fetched by people building searchable podcast archives
Fetch the whole transcript
The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000776103965