Connect with us

Published

on

Jensen Huang, co-founder and chief executive officer of Nvidia Corp., during a news conference in Taipei, Taiwan, on Tuesday, June 4, 2024. Nvidia is still working on the certification process for Samsung Electronics Co.’s high-bandwidth memory chips, a final required step before the Korean company can begin supplying a component essential to training AI platforms. 

Annabelle Chih | Bloomberg | Getty Images

Nvidia called DeepSeek’s R1 model “an excellent AI advancement,” despite the Chinese startup’s emergence causing the chip maker’s stock price to plunge 17% on Monday.

“DeepSeek is an excellent AI advancement and a perfect example of Test Time Scaling,”  an Nvidia spokesperson told CNBC on Monday. “DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely-available models and compute that is fully export control compliant.”

The comments come after DeepSeek last week released R1, which is an open-source reasoning model that reportedly outperformed the best models from U.S. companies such as OpenAI. R1’s self-reported training cost was less than $6 million, which is a fraction of the billions that Silicon Valley companies are spending to build their artificial-intelligence models. 

Nvidia’s statement indicates that it sees DeepSeek’s breakthrough as creating more work for the American chip maker’s graphics processing units, or GPUs. 

Read more DeepSeek coverage

“Inference requires significant numbers of NVIDIA GPUs and high-performance networking,” the spokesperson added. “We now have three scaling laws: pre-training and post-training, which continue, and new test-time scaling.”

Nvidia also said that the GPUs that DeepSeek used were fully export compliant. That counters Scale AI CEO Alexandr Wang’s comments on CNBC last week that he believed DeepSeek used Nvidia GPUs models which are banned in mainland China. DeepSeek says it used special versions of Nvidia’s GPUs intended for the Chinese market.

Analysts are now asking if multi-billion dollar capital investments from companies like Microsoft, Google and Meta for Nvidia-based AI infrastructure are being wasted when the same results can be achieved more cheaply. 

Earlier this month, Microsoft said it is spending $80 billion on AI infrastructure in 2025 alone while Meta CEO Mark Zuckerberg last week said the social media company planned to invest between $60 to $65 billion in capital expenditures in 2025 as part of its AI strategy. 

“If model training costs prove to be significantly lower, we would expect a near-term cost benefit for advertising, travel, and other consumer app companies that use cloud AI services, while long-term hyperscaler AI-related revenues and costs would likely be lower,” wrote BofA Securities analyst Justin Post in a note on Monday.

Nvidia’s comment also reflects a new theme that Nvidia CEO Jensen Huang, OpenAI CEO Sam Altman and Microsoft CEO Satya Nadella have discussed in recent months.

Much of the AI boom and the demand for Nvidia GPUs was driven by the “scaling law,” a concept in AI development proposed by OpenAI researchers in 2020. That concept suggested that better AI systems could be developed by greatly expanding the amount of computation and data that went into building a new model, requiring more and more chips.

Since November, Huang and Altman have been focusing on a new wrinkle to the scaling law, which Huang calls “test-time scaling.” 

This concept says that if a fully trained AI model spends more time using extra computer power when making predictions or generating text or images to allow it to “reason,” it will provided better answers than it would have if it ran for less time. 

Forms of the test-time scaling law are used in some of OpenAI’s models such as o1 as well as DeepSeek’s breakthrough R1 model.

WATCH: DeepSeek challenging sense of U.S. exceptionalism priced into markets, fund manager says

Continue Reading

Technology

Taiwan accuses China’s biggest chipmaker SMIC of ‘illegally’ poaching tech talent

Published

on

By

Taiwan accuses China's biggest chipmaker SMIC of 'illegally' poaching tech talent

A logo hangs on the building of the Beijing branch of Semiconductor Manufacturing International Corporation (SMIC) on December 4, 2020 in Beijing, China.

Vcg | Visual China Group | Getty Images

Taiwan investigators on Friday alleged that Chinese chipmaker Semiconductor Manufacturing International Co. (SMIC) illegally recruited high-technology talent.

Taiwan’s Ministry of Justice Investigation Bureau (MIJB) said in a statement that SMIC had used a Samoa-based entity as cover to set up a subsidiary on the island “under the guise of foreign investment” and has been “actively recruiting” talent from Taiwan.

CNBC was unable to independently verify the claims and SMIC was not immediately available for comment.

The ministry said Taiwan began investigating the issue in December 2024. Eleven Chinese enterprises suspected of paoching talent were investigated, it said, with agents conducting searches at 34 locations and questioning 90 individuals.

SMIC is China’s biggest semiconductor manufacturing firm. It was thrust into the spotlight in 2023 when it was revealed to be the maker of the 7 nanometer chip in Huawei’s smartphone at the time. A few years prior, SMIC was put on a U.S. government export blacklist.

China has been trying to ramp up its chipmaking capabilities via SMIC, but the company remains behind competitors like TSMC in Taiwan. Chip export restrictions imposed by the U.S. also mean SMIC is unable to access the latest chipmaking tools from critical suppliers like ASML that could allow it to catch up.

Taiwan is a hotbed of talent in the semiconductor industry as it is home to TSMC, the world’s biggest and most advanced chipmaker. The U.S. has sought to tap into this talent, and bring more chipmaking capabilities to its shores, by convincing TSMC to build more manufacturing capacity in the country.

Taiwan’s MJIB said it set up a special task force at the end of 2020 to investigate allegations of “illegal poaching” of talent.

“Chinese enterprises often disguise their identities through various means, including setting up operations under the guise of Taiwanese, overseas Chinese, or foreign-invested companies, while in reality being backed by Chinese capital, establishing unauthorized business locations in Taiwan without government approval, and using employment agencies to falsely assign employees to Taiwanese firm,” the ministry said.

Continue Reading

Technology

Assassin’s Creed-maker Ubisoft’s shares surge 9% on deal to spin off top game franchises

Published

on

By

Assassin's Creed-maker Ubisoft's shares surge 9% on deal to spin off top game franchises

Artwork for Ubisoft’s upcoming “Assassin’s Creed Shadows” game.

John Keeble | Getty Images

Ubisoft shares surged on Friday after the French video game publisher announced a deal to spin out some of its best-selling franchises.

On Thursday, Ubisoft revealed plans to form a new gaming subsidiary, with Chinese technology giant Tencent investing 1.16 billion euros ($1.25 billion) into the unit.

The subsidiary will include Ubisoft’s best-known games brands, including Assassin’s Creed, Far Cry and Tom Clancy’s Rainbow Six, according to a press release.

Ubisoft shares jumped 11% Friday morning before paring gains later in the day to trade 9% higher.

The game maker said Thursday that its newly formed unit would “focus on building game ecosystems designed to become truly evergreen and multi-platform.”

Saudi Arabia has highest penetration rate of women professional gamers globally: Saudi prince

The investment from Tencent values the new subsidiary at 4 billion euros, Ubisoft said. That’s more than double Ubisoft’s current market capitalization.

Investors were hoping for a move from Ubisoft to clear up uncertainty surrounding its future after a series of challenges faced by the firm in recent years.

Ubisoft has been plagued by various issues, from financial struggles to delays to some of its key games — including its most recently released Assassin’s Creed Shadows title.

In February, the firm reported a 52% drop in fiscal third-quarter net bookings, beset by underperformance of some of its key games.

Ubisoft released its newest title Assassin’s Creed Shadows, the latest installment in Ubisoft’s top game franchise, last week.

The game secured generally positive reviews, getting an average reviewer score of 82 on review aggregation site Metacritic.

Continue Reading

Technology

U.S. tech giants are betting big on humanoid robots — but China’s already ahead, analysts say

Published

on

By

U.S. tech giants are betting big on humanoid robots — but China's already ahead, analysts say

Unitree’s G1 robot at the Mobile World Congress 2025 in Barcelona, Spain, on March 6, 2025.

Nurphoto | Nurphoto | Getty Images

American tech giants like Tesla and Nvidia are racing to develop humanoid robots, stressing their importance to the future economy. But analysts warn they are already at risk of losing out to China.

So-called humanoid robots — artificial intelligence-powered machines designed to resemble humans in appearance and movement — are expected to provide a range of use cases, such as filling industrial and service sector jobs.  

Investor excitement surrounding the robots has been mounting amid increased mentions from tech leaders like Nvidia’s Jensen Huang, who ushered in “the age of generalist robotics” earlier this month when announcing a new portfolio of technologies for humanoid robot development.

In the manufacturing of the robots themselves, Tesla’s humanoid robot project, Optimus, appears to be leading in the U.S., with CEO Elon Musk announcing plans to produce about 5,000 units this year.

While Musk’s ambitious plans could give it a leg up on U.S. competitors like Apptronik and Boston Dynamics that are yet to hit the mass market, he will face stiff competition from a familiar source: China. 

Jensen Huang, co-founder and chief executive officer of Nvidia, speaks about humanoids during the 2025 CES event in Las Vegas on Jan. 6, 2025.

Bridget Bennett | Bloomberg | Getty Images

Hangzhou-based Unitree Robotics last month briefly sold two humanoid robots to consumers on the e-commerce platform JD.com, as per local media. Meanwhile, Shanghai-based robotics startup Agibot, also known as Zhiyuan Robotics, has matched Optimus’s goal to produce 5,000 robots this year, according to the South China Morning Post

As Chinese electric vehicle companies like BYD begin outpacing Tesla’s growth and undercutting its prices, experts say a similar dynamic could play out in humanoid robotics. 

“China has the potential to replicate its disruptive impact from the EV industry in the humanoid space. However, this time the disruption could extend far beyond a single industry, potentially transforming the labor force itself,” said Reyk Knuhtsen, analyst at SemiAnalysis, an independent research and analysis company specializing in semiconductors and AI.

Dancing on the competition?

In a research note in February, Morgan Stanley estimated that current building costs of humanoid robots could range from $10,000 to $300,000 per unit, given different configurations and downstream application requirements.

However, Chinese companies are already undercutting U.S. competitors in terms of price thanks to superior economies of scale and manufacturing capabilities, according to Knuhtsen. 

For example, Unitree released its G1 humanoid robot for consumers in May with a starting price of $16,000. In comparison, Morgan Stanley estimates that the selling cost of Tesla’s Optimus Gen2 humanoid robot could be around $20,000, but only if the company is able to scale, shorten its research and development cycle, and use cost-effective components from China.

Unitree made a major splash in the robot’s space in January when 16 of its highest-performing H1 humanoid robots joined a group of human dancers to celebrate the Lunar New Year in a demonstration broadcast on national television.

But there are signs that China’s progress in robots go much further. Morgan Stanley’s February research note found that the country has led the world in patent filings mentioning “humanoid” over the past five years, with 5,688 patents compared with 1,483 from the United States.

Large players such as Xiaomi and EV makers, such as BYD, Chery, and Xpeng, are also involved in the humanoid robot space. 

“Our research suggests China continues to show the most impressive progress in humanoid robotics where startups are benefitting from established supply chains, local adoption opportunities, and strong degrees of national government support,” the note said.

Beijing has increasingly backed the space, with government departments promoting their development. In 2023, the Ministry of Industry and Information Technology issued guidelines for the space, calling for “production at scale” by 2025.

'Very fast' decline in component costs of humanoid robots in next few years, says analyst

According to Ming Hsun Lee, head of Greater China automotive and industrials research at BofA Global Research, China sees humanoid robots as an important industry because of their potential to mitigate looming labor shortages. 

“I think in the short-term, three to four years, we will see humanoid robots initially applied in production lines to compare some workers, and in the midterm, we will see them gradually spread into the service industry,” he said. 

Musk predicted that he’d have over 1,000, or a few thousand, Optimus robots working at Tesla in 2025. According to Chinese state media, EV makers like BYD and Geely have already deployed some of Unitree’s humanoid robots at their factories.  

Lee said that increased adoption will coincide with a “very fast” decline in component costs, also noting that China owns around 70% of the supply chain for these components. 

According to a report by SemiAnalysis earlier this month, the Unitree G1 — “the only viable humanoid robot on the market” — is entirely decoupled from American components.

The report warns that China is the only country positioned to reap the economic awards of intelligent robotics systems, including humanoid robots, which “poses an existential threat to the US as it is outcompeted in all capacities.” 

“To catch up, U.S. players must rapidly mobilize a strong manufacturing and industrial base, whether domestically or through allied nations … For Tesla and similar firms, it may be wise to begin reshoring or ‘friendshoring’ their component sourcing and manufacturing to reduce reliance on China,” said SemiAnalysis’ Knuhtsen. 

Bank of America analysts predicted in a research note this month that the deployment of humanoid robots will accelerate rapidly, aided by the development of AI, with global annual sales reaching 1 million units by 2030 and 3 billion humanoid robots in operation by 2060. 

Continue Reading

Trending