Nvidia Praises Chinese AI Models After Resuming AI Chip Sales In The Region - 1

Photo by BoliviaInteligente on Unsplash

Nvidia Praises Chinese AI Models After Resuming AI Chip Sales In The Region

  • Written by Andrea Miliani Former Tech News Expert
  • Fact-Checked by Sarah Frazier Former Content Manager

Nvidia announced on Monday that it will resume selling its H20 AI chips to the Chinese market. The company’s CEO, Jensen Huang, also described Chinese AI technology developed by companies such as DeepSeek as “world-class” and “revolutionising.”

In a rush? Here are the quick facts:

  • Nvidia will resume selling its AI chips to China, as the U.S. government has granted the company its license again.
  • CEO Huang Jensen described Chinese AI technology as “world-class” and “revolutionising.”
  • Huang said he expects to bring more advanced GPUs to the Chinese market.

According to Reuters , the tech giant said it would soon begin selling China its graphics processing unit for AI systems—the flagship AI chip model specially designed for the Chinese market —after the U.S. government agreed to reverse the export restriction imposed in April over national security concerns.

Despite the sales restrictions, Nvidia became the first company in the world to reach a $4 trillion market valuation last week. Now that the U.S. government has granted the company its license again, Nvidia expects to sell more AI chips in China and continue its expansion.

“H20 was released from its ban, the memory bandwidth is extremely good, for LLMs and other new models it will be excellent,” said Huang during an event in Beijing on Wednesday, as reported by Reuters . “I hope to get more advanced chips into China. Today, H20 is still incredibly good, but in the coming years, whatever we are allowed to sell to China, we will do so.”

Experts have considered the decision of the U.S. government “rare” as it has been criticized by American legislators who initially backed the restriction to prevent China from advancing its AI capabilities.

“The H20 is a powerful chip that, according to our bipartisan investigation, played a significant role in the rise of PRC AI companies like DeepSeek,” said Republican John Moolenaar, chair of the House of Representatives Select Committee on China.

At the Beijing expo, Huang praised Chinese AI models, including those developed by DeepSeek, Alibaba, and Tencent. He described the technology as “world-class” and said their AI was “revolutionising” supply chains.

Hackers Can Remotely Trigger Brakes on U.S. Trains - 2

Image by Alexander Shustov, from Unsplash

Hackers Can Remotely Trigger Brakes on U.S. Trains

  • Written by Kiara Fabbri Former Tech News Writer
  • Fact-Checked by Sarah Frazier Former Content Manager

Many trains in the U.S. continue to face a security risk that allows hackers to activate their brakes from a distance.

In a rush? Here are the quick facts:

  • Vulnerability discovered in 2012 by researcher Neil Smith, still not fixed.
  • Railroad industry ignored warnings, demanding real-life exploit proof first.
  • CISA calls exploit complex, but researcher says it’s “low attack complexity.”

404Media reports that this serious security flaw has been known for over a decade, but is still not fully fixed. The flaw was first reported by the U.S. Cybersecurity and Infrastructure Security Agency (CISA) and independent researcher Neil Smith.

Smith first discovered the vulnerability in 2012. He explained that the vulnerability can be exploited through radio frequencies used between the front and back sections of a train. “All of the knowledge to generate the exploit already exists on the internet. AI could even build it for you,” Smith told 404 Media.

Smith explained how hackers can activate the exploit using a FlipperZero device within a 200-foot range, or an aircraft transmitter at 30,000 feet, to achieve coverage up to 150 miles.

The flaw stems from a safety system introduced in the 1980s called the End-of-Train and Head-of-Train Remote Linking Protocol (EOT/HOT), designed to improve communication on trains. Smith said, “The radio link is a commonly found [frequency-shift keying] data modem that was easy to identify,” but figuring out the meaning of the data was harder, as reported by 404Media.

When Smith alerted the railroad industry in 2012, he says they ignored him. “The Association of American Railroads (AAR) […] would not acknowledge the vulnerability as real unless someone could demonstrate it to them in real life,” he said. AAR declined to comment on the issue.

CISA official Chris Butera revealed that rail sector stakeholders have monitored this vulnerability for more than a decade, but state that its exploitation requires both physical access, advanced technical skills, and specialized equipment.

But Smith disagrees, calling CISA’s statement “overly complicated,” noting the exploit is actually “low attack complexity,” as noted by 404Media

The vulnerability remains unfixed, with industry efforts to update the system ongoing but slow. Smith criticized the railway industry’s response, saying it follows the insurance industry’s “delay, deny, defend” approach, as noted by 404Media.