Connect with us

Published

on

A smartphone with a displayed AMD logo is placed on a computer motherboard in this illustration taken March 6, 2023. 

Dado Ruvic | Reuters

American semiconductor company Advanced Micro Devices has failed in getting a made-for-China AI chip past U.S. regulators and will need to apply for an export license, Bloomberg reported Tuesday.

The report said AMD designed the chip to have lower performance than its premium products in order to comply with U.S. export restrictions. But Bloomberg reported the Commerce Department did not clear the chip for sale in China because it was still too advanced.

AMD will now have to obtain a license from the department’s Bureau of Industry and Security, the report said.

It’s not clear if the company will apply for the license. AMD and the Commerce’s Bureau of Industry and Security did not immediately respond to CNBC’s requests for comment.

While the U.S. has restricted sales of products containing the nation’s most advanced semiconductor technologies to China, citing national security concerns, American companies have continued to sell mature or less advanced technologies to the massive market without licenses.

AMD’s products include chips that can be used to develop and train AI models – something U.S. officials have warned that Beijing could use to gain military advantages. 

AMD CEO Lisa Su: AI is the most important technology that has come in the last 50 years

In 2022, U.S. President Joe Biden’s administration unveiled an initial set of export controls to curb China’s access to advanced semiconductor technologies. Leading AI chip company Nvidia subsequently said it would sell slowed-down versions of their premium AI chips that comply with U.S. restrictions. 

However, those chips were also banned in October, when the U.S. expanded restrictions to include more technology and target chips that were seen as circumventing controls.

Nvidia has since redesigned products to be less powerful for the Chinese market to align with the 2023 restrictions. 

In the lead-up to the October restrictions, Nvidia had warned that further U.S. export curbs on its chips to China would risk a “permanent loss” for American semiconductor firms to lead in one of the world’s largest markets. 

In Nvidia’s November earnings call, Chief Financial Officer Colette Kress said China and other regions targeted by U.S. export controls had consistently contributed approximately 20% to 25% of data center revenue over the past few quarters. While Nvidia reported blockbuster fourth-quarter results, Kress noted during the February earnings call that data center revenue from China declined significantly following the U.S. export curbs.

Compared with Nvidia, AMD had a smaller foothold in the Chinese AI chip market prior to the trade restrictions. But the company has begun targeting the AI chip market more aggressively, launching a new MI300 product line that is seen as a challenge to GPU products from Nvidia.

It’s not clear which Chinese customers AMD designed the chips for. Some leading Chinese tech giants, such as Tencent have reportedly stockpiled enough advanced chips from Nvidia to train their AI chatbots’ capabilities for “at least a couple more generations.”

Meanwhile, U.S.-sanctioned Huawei is reportedly developing its own chips and chipmaking tools, along with other domestic firms, as Chinese companies attempt to fill the gap created by U.S. restrictions.

Despite restricted sales to China, shares of both Nvidia and AMD have soared amongst an AI frenzy. Nvidia is up more than 250% in the past year while AMD surged over 150%.

Read the full report from Bloomberg.

Continue Reading

Technology

Google announces new health-care AI updates for Search

Published

on

By

Google announces new health-care AI updates for Search

A Google corporate logo hangs above the entrance to their office at St. John’s Terminal on March 11, 2025, in New York City.

Gary Hershorn | Corbis News | Getty Images

Google on Tuesday announced health-care updates to Search, including a way for people with specific health conditions to compare their experiences with others.

The company unveiled a new feature called “What People Suggest,” which uses AI to pull together online commentary from patients with similar diagnoses. A patient with arthritis would be able to look up how other people with the condition approach exercise, for instance. The feature is available on mobile devices in the U.S., Google said.

Google said it has also expanded its knowledge panels, or the information boxes that appear to the right of search results, to cover “thousands” more health topics. The panels are coming to new countries and languages, including Spanish, Japanese and Portuguese, starting on mobile devices.

The tech giant has launched several health-care projects and features over the years, but it has struggled to outline a consistent business strategy within the sector. The company built out a formal Google Health unit starting around 2018, which swelled to more than 500 employees, but it was dissolved in 2021.

Karen DeSalvo, Google’s chief health officer, told CNBC months later that the company was “still all-in on health.”

In recent years, many of Google’s health-care initiatives have centered around AI.

Google introduced artificial intelligence summaries called AI Overviews last year, and the feature shows a quick summary of answers to search questions at the very top of Search. The rollout was rocky, as users were quick to share examples AI tool giving incorrect and controversial responses, like encouraging users to add glue to pizza.

AI Overviews appear for some health-related queries, like “How do I know if I have the flu?” But some experts have encouraged users to use caution with these answers, according to a December report from The Senior List. Out of more than 200 health searches, a panel of medical experts said 70% Google’s AI Overviews were considered risky.

Google said Tuesday that recent health-focused advancements with its Gemini models have allowed the company to improve AI Overviews for health topics.

In late 2023, Google announced MedLM, a suite of AI models designed specifically for health-care, to help clinicians and researchers carry out complex studies, summarize doctor-patient interactions and complete other tasks.

The company also unveiled Vertex AI Search for Healthcare that year, which is a generative AI tool that clinicians can use to search for information across disparate medical records.

Watch: Google to acquire cloud security startup Wiz for $32 billion.

Google to acquire cloud security startup Wiz for $32 billion

Continue Reading

Technology

Google to acquire cloud security startup Wiz for $32 billion

Published

on

By

Google to acquire cloud security startup Wiz for  billion

The Wiz website on a smartphone arranged in New York, US, on Tuesday, July 16, 2024. 

Gabby Jones | Bloomberg | Getty Images

Google on Tuesday signed a “definitive agreement” to acquire Wiz, a New York-based cloud security startup, for $32 billion in an all-cash deal.

The deal, which will be Google’s largest-ever acquisition, will improve its cloud security offering in a world of advancing artificial intelligence and cybersecurity threats. Wiz will become a part of the company’s cloud business. Google said it expects to close the deal in 2026.

“Google Cloud is a leader in cloud infrastructure, with deep AI expertise and a track record of industry-leading security innovation,” Google said in a release. “Bringing all this to Wiz will help make their solutions even better and more scalable, benefiting customers and partners across all major clouds.”

The acquisition comes after CNBC reported in July that Wiz had walked away from a potential $23 billion acquisition by Google and announced to employees that it would pursue an initial public offering instead.

“Saying no to such humbling offers is tough,” Wiz co-founder Assaf Rappaport wrote to employees in a July memo obtained by CNBC. At the time, a source familiar with the matter told CNBC that Wiz walked away from the deal in part due to antitrust and investor concerns.

Before talks with Google were reported, Wiz had set its sights on two goals: an IPO and $1 billion in annual recurring revenue. In the memo at the time, Rappaport wrote that the company would pursue those milestones.

Wiz was founded in 2020 and has grown rapidly under Rappaport, with the company hitting $100 million in annual recurring revenue after just 18 months. The company’s cloud security products include prevention, active detection and response, a portfolio that’s appealed to large firms and would have helped Google compete with Microsoft, which also sells security software.

“Becoming part of Google Cloud is effectively strapping a rocket to our backs: it will accelerate our rate of innovation faster than what we could achieve as a standalone company,” Rappaport said in a blog post Tuesday.

Google has a long history in dealmaking and snatching up smaller companies to broaden its offerings to customers. Its largest deal before Wiz was the $12.5 billion acquisition of hardware marker Motorola in 2012. Two years later, the company sold some assets to Lenovo for $2.9 billion. Google has also made cybersecurity acquisitions in the past, paying $5.4 billion for Mandiant in 2022.

Wiz’s products will still work on competitor platforms including Amazon Web Services, Microsoft Azure and Oracle Cloud, the companies said. The Wall Street Journal first reported Monday that the companies were in advanced discussions.

While the agreement may still draw government scrutiny, many on Wall Street have been hopeful that President Donald Trump’s new White House administration will be more amenable to tech industry deals. Alphabet is currently battling an antitrust suit over its online search dominance.

— CNBC’s Jennifer Elias, Jordan Novet and Rohan Goswami contributed to this report.

Continue Reading

Technology

Baidu, once China’s generative AI leader, is battling to regain its position

Published

on

By

Baidu, once China’s generative AI leader, is battling to regain its position

Pictured here is the Ernie bot mobile interface, with the Baidu search engine home page in the background.

Future Publishing | Future Publishing | Getty Images

Chinese tech giant Baidu has released two new free-to-use artificial intelligence models as it vies to regain its leading position in the country’s fiercely competitive AI space. 

The Baidu models launched Sunday included the company’s first reasoning-focused model, and come ahead of plans to move toward an open-source strategy. 

However, experts told CNBC that while the release of the models is a positive development for Baidu, they also highlight how it is playing catch up as its Ernie bot — one of China’s earliest versions of a ChatGPT-like chatbot — struggles to gain widespread adoption. 

“The new models make Baidu more competitive since the company has been lagging behind in a reasoning model release,” Lian Jye Su, chief analyst at Omdia, told CNBC.

A reasoning model is a large language model that breaks down tasks into smaller pieces and considers multiple approaches before generating a response. It is designed to process complex problems in a similar way to humans.

Chinese startup DeepSeek upended the global AI race and transformed China’s ecosystem in January when it released its R1 reasoning model, which rivaled American competitors despite costing a fraction of the price.

Baidu has said its new ERNIE X1 reasoning model “delivers performance on par with DeepSeek R1 at only half the price,” and has “stronger understanding, planning, reflection, and evolution capabilities.” CNBC has not been able to independently verify this claim.

According to Wei Sun, principal analyst of artificial intelligence at Counterpoint Research, Baidu’s future competitiveness could hinge on whether its new models deliver on the promised performance and cost advantages. 

“Baidu is clearly in catch-up mode, largely due to its slow innovation pace and underestimating rapid shifts in market dynamics,” Sun said. 

What happened? 

Baidu rolled out its first generative AI platform to the public in 2023, giving China one of its first answers to OpenAI’s popular AI chatbot ChatGPT. 

However, despite initial momentum, Baidu’s Ernie product has since been eclipsed by competitors including startups as well as large-tech companies such as Alibaba and ByteDance.

Experts list a number of reasons for Baidu’s struggles and slow rate of innovation.

“Baidu fell behind when they tried to build proprietary models and compete for funding for AI,” Ray Wang, principal analyst and founder of Constellation Research, told CNBC. He added that the company has also suffered from recent government crackdowns and was distracted by “regulatory nonsense.” 

CFOTO | Future Publishing | Getty Images

Proprietary models keep their source code and underlying architecture confidential, in contrast to models from the likes of DeepSeek, whose source code is made freely available on the open web for possible modification and redistribution.

“Using a closed-source approach means that [Baidu] was training its model from scratch whereas the open-source models were able to leverage certain parts that were communal to developers,” said Kai Wang, a senior equity analyst for Morningstar. 

Baidu, however, said last month that it would make its next-generation AI model Ernie open-source from June 30, according to Reuters.

“Baidu has always been very supportive of its proprietary business model and was vocal against open source, but disruptors like DeepSeek have proven that open source models can be as competitive,” said Omdia’s Su. 

He added that Baidu is “merely following the footstep” of its biggest competitors in China, namely Alibaba, DeepSeek, and Tencent, which have all now released open-source models. 

Baidu’s advantages

Continue Reading

Trending