Nvidia's Singapore GPU sales are 28% of its revenue, but only 1% are delivered to the country: Report

Unlike DeepSeek, OpenAI’s o1 is proprietary, meaning consumers and businesses pay the company to use its model and services. While some companies prefer to use proprietary technologies—because they are vetted by their builders and come with built-in cybersecurity controls—others prefer open-source technologies because they are easier to customize and control.

However, numerous security concerns have surfaced about the company, prompting private and government organizations to ban the use of DeepSeek. Here's what you need to know.

Taiwan banned government agencies from using DeepSeek’s AI model last week, citing security concerns.

The chatbot was removed from app stores after its privacy policy was questioned in Italy. The Italian goverment previously temporarily blocked ChatGPT over privacy concerns in March 2023.

Regulators in South Korea, Ireland and France have all begun investigations into how DeepSeek handles user data, which it stores in servers in China.

While there has been no nationwide ban, Texas became the first US state to ban DeepSeek on government-issued devices.

The US Navy has reportedly banned its members from using DeepSeek - though it has not confirmed this to the BBC.

Australia bans DeepSeek on government devices over security risk

Western countries have a track record of being suspicious of Chinese tech - notably telecoms firm Huawei and the social media platform, TikTok - both of which have been restricted on national security grounds.

The initial reaction to DeepSeek - which quickly became the most downloaded free app in the UK and US - appeared to be different.

'Old Donald' described it as a "wake up call" for the US but said overall it could be a positive development, if it lowered AI costs.

An Australian science minister previously said in January that countries needed to be "very careful" about DeepSeek, citing "data and privacy" concerns.

Interestingly, Microsoft CEO Satya Nadella sees the emergence of DeepSeek's R1 AI model and other iterations as good for business amid Azure's slow growth rate. According to the executive:

AI will be much more ubiquitous. And so, therefore, for a hyperscaler like us and a PC platform provider like us, this is all good news as far as I'm concerned."

Microsoft integrates DeepSeek’s AI model into Azure and GitHub (Owned by Microsoft).

Nvidia -- along with analysts and investors -- has taken a more constructive view, arguing that DeepSeek's ability to design an AI chatbot comparable to OpenAi's ChatGPT should be lauded, and that its may represent an opportunity for the entire AI market. With that in mind, should investors buy the dip on Nvidia?

"I think people are moving from this craziness around the model, understanding that, thanks to open-source ... a lot of these models are going to be free and freely available,” Wolf told CNBC. Hugging Face is a popular code repository for open-source AI projects.

DeepSeek might not be as disruptive as claimed, firm reportedly has 50,000 Nvidia GPUs and spent $1.6 billion on buildouts

Does humanity know what it's doing?

Led by Ph.D. candidate Jiayi Pan, the team replicated DeepSeek R1-Zero’s reinforcement learning capabilities using a small language model with just 3 billion parameters.

It performs a number of tasks as well as its American competitors do, but it censors topics such as Tiananmen Square.

Misc. tests associated with Taiwanese language
1. Gemini | On tackling Taiwanese tone sandhi
2. ChatGPT | On tackling Taiwanese tone sandhi
3. DeepSeek (1/25) DeepSeek (2/2) | On tackling Taiwanese tone sandhi

It could be DeepSeek accelerates AI commoditization and helps drive it further into the mass market. Even in that case, the lion's share of AI value could remain in a small, upper end of the market, where the best ideas originate and stay for a hot minute before they are commoditized.

Where’s the Value in AI?

For example, when training its V3 model, DeepSeek reconfigured Nvidia's H800 GPUs: out of 132 streaming multiprocessors, it allocated 20 for server-to-server communication, possibly for compressing and decompressing data to overcome connectivity limitations of the processor and speed up transactions. To maximize performance, DeepSeek also implemented advanced pipeline algorithms, possibly by making extra fine thread/warp-level adjustments.

“Microsoft Corp. and OpenAI are investigating whether data output from OpenAI’s technology was obtained in an unauthorized manner by a group linked to Chinese artificial intelligence startup DeepSeek, according to people familiar with the matter.” The story goes on to say that “Such activity could violate OpenAI’s terms of service or could indicate the group acted to remove OpenAI’s restrictions on how much data they could obtain, the people said.”

The venture capitalist and new 'Old Donald' administration member David Sacks, meanwhile, said that there is “substantial evidence” that DeepSeek “distilled the knowledge out of OpenAI’s models.”

Bloomberg reports that the U.S. government is investigating whether DeepSeek acquired Nvidia's restricted GPUs for AI workloads through intermediaries in Singapore, bypassing U.S. export restrictions.

DeepSeek has not disclosed the specific hardware used to train its R1 model. However, it previously indicated that it used a limited number of H800 GPUs — 2,048 — to train its V3 model with 671 billion parameters in just two months, or 2.8 million GPU hours

Although China has boosted investment in advanced tech to diversify its economy, DeepSeek is not one of the big Chinese firms that have been developing AI models to rival US-made ChatGPT.

Experts say the US still has an advantage - it is home to some of the biggest chip companies - and that it's unclear yet exactly how DeepSeek built its model and how far it can go.

“Our principle is not to lose money, nor to make huge profits … our starting point is not to take advantage of the opportunity to make a fortune, but to be at the forefront of technology and promote the development of the entire ecosystem.”

By 2021, he was reported to have bought 10,000 of the chips, seemingly for his personal hobby. Only a handful of large Chinese tech firms have similar reserves of Nvidia semiconductors. “Many people would think that there is an unknown business logic behind this, but in fact, it is mainly driven by curiosity,” Liang said in 2023.

Some analysts and investors have expressed scepticism about DeepSeek’s market-rattling claims.

“Microsoft Corp. and OpenAI are investigating whether data output from OpenAI’s technology was obtained in an unauthorized manner by a group linked to Chinese artificial intelligence startup DeepSeek, according to people familiar with the matter.” The story goes on to say that “Such activity could violate OpenAI’s terms of service or could indicate the group acted to remove OpenAI’s restrictions on how much data they could obtain, the people said.”

The venture capitalist and new 'Old Donald' administration member David Sacks, meanwhile, said that there is “substantial evidence” that DeepSeek “distilled the knowledge out of OpenAI’s models.”

But the team behind the system, called DeepSeek-V3, described an even bigger step. In a research paper explaining how they built the technology, DeepSeek’s engineers said they used only a fraction of the highly specialized computer chips that leading A.I. companies relied on to train their systems.

When Chinese artificial intelligence firm DeepSeek shocked Silicon Valley and Wall Street with its powerful new A.I. model, Marc Andreessen, the Silicon Valley investor, went so far as to describe it as “A.I.’s Sputnik moment.” Presumably, Mr. Andreessen wasn’t calling on the federal government to start a massive new program like NASA, which was our response to the Soviet Union’s Sputnik satellite launch; he wants the U.S. government to flood private industry with capital, to ensure that America remains technologically and economically dominant.

As an antitrust enforcer, I see a different metaphor. DeepSeek is the canary in the coal mine. It’s warning us that when there isn’t enough competition, our tech industry grows vulnerable to its Chinese rivals, threatening U.S. geopolitical power in the 21st century.

Although it’s unclear precisely how much more efficient DeepSeek’s models are than, say, ChatGPT, its innovations are real and undermine a core argument that America’s dominant technology firms have been pushing — namely, that they are developing the best artificial intelligence technology the world has to offer, and that technological advances can be achieved only with enormous investment — in computing power, energy generation and cutting-edge chips. For years now, these companies have been arguing that the government must protect them from competition to ensure that America stays ahead.

It should be no surprise that our big tech firms are at risk of being surpassed in A.I. innovation by foreign competitors. After companies like Google, Apple and Amazon helped transform the American economy in the 2000s, they maintained their dominance primarily through buying out rivals and building anticompetitive moats around their businesses.

Over the last decade, big tech chief executives have seemed more adept at reinventing themselves to suit the politics of the moment - resistance sympathizers, social justice warriors, MAGA enthusiasts - than on pioneering new pathbreaking innovations and breakthrough technologies.