DeepSeek’s “surprising” AI claims show China “starting to gain an edge”

May Be Interested In:Jan. 6 defendant who allegedly brought guns to Obama’s house wants pardon


DeepSeek burst into public consciousness when the Chinese tech firm claimed to have made substantial progress in training its AI for a fraction of the cost of its American competitors and with a relatively low number of highly-advanced chips.

As a result, the stocks in major AI companies such as OpenAI and advanced microchip manufacturers such as Nvidia took a pounding, losing around $1tn in value overall, before recovering some of those losses once the dust had settled.

The emergence of DeepSeek came shortly after President Donald Trump unveiled his “Stargate” project to invest $500bn in advancing American artificial intelligence, an effort backed by the likes of OpenAI co-founder and CEO, Sam Altman.

But do DeepSeek’s claims about its achievements stack up? And what does it mean for the future of artificial intelligence companies? Here, Newsweek compiles informed views and insights on those key questions.

Stuart Russell: If True, DeepSeek’s Claims Show US AI Firms Wasted a Lot of Money

If the claims of DeepSeek’s excellent performance hold up, this would be another example of Chinese developers managing to roughly replicate US systems a few months after their release.

The general outlines of how OpenAI’s o1 works have been known for quite a while—even before it was released—so it’s not all that surprising that it can be roughly replicated.

What’s surprising is the claim that the total training cost was only $6 million and it was done using “only” a few thousand GPU chips. (Reports vary widely on exactly what was used.)

Both of these numbers have caused grief in the markets because US companies such as Microsoft, Meta, and OpenAI are making huge investments in chips and data centers on the assumption that they will be needed for training and operating these new kinds of systems.

(And by “huge” I mean really, really huge—possibly the biggest capital investments the human race has ever undertaken.)

If that assumption is false and it can be done much more cheaply, then those investments are mostly a waste of money, and future demand for Nvidia’s chips in particular will be much lower than predicted.

Stuart Russell is Distinguished Professor of Computer Science, University of California, Berkeley, and Smith-Zadeh Professor in Engineering; Professor of Cognitive Science; Professor of Computational Precision Health, UCSF; and Honorary Fellow of Wadham College, Oxford.

Illustration. DeepSeek AI is a Chinese artificial intelligence company and the name of its conversational agent, which uses a large language model (LLM) that it claims it developed for a fraction of the cost of…


Sipa via AP Images

Nabil Jawdat Sarhan: China Is Gaining an Edge Over U.S.

I have no reason to doubt the credibility of this information.

China has invested heavily in nurturing talent in artificial intelligence, computer science, computer engineering, and other fields.

As a researcher in AI, I’m astonished by the massive volume of Chinese publications in top research journals and conferences in the field. Moreover, Chinese companies have been successful in making competitive products at much lower costs than in the U.S.

The technology gap between the USA and China has substantially narrowed over the years, and China has started to gain an edge in some areas.

Nabil Jawdat Sarhan is associate professor in the Department of Electrical & Computer Engineering at at Wayne State University, and Director, Wayne State Computer Systems and Deep Learning Research Lab.

Mark Kon: I Find DeepSeek’s Story Believable

There is a general consensus that the computational power required by DeepSeek is lower by orders of magnitude (e.g. a factor of more than 100) than the power required by large language models such as ChatGPT.

The question of whether the low level of financing claimed by the company reflects reality, or whether there were funding sources that were not acknowledged (e.g. from the government of China), is more difficult.

I personally believe the story of the development of the DeepSeek engine. Funneling development that displayed such a low level of hardware use at a substantially high level would have been unlikely.

Additionally, if the government of China had been involved in the initial construction of the DeepSeek engine, it would probably have covered the intellectual property more carefully, and not allowed it to become open source.

Mark Kon is a professor in the Department of Mathematics and Statistics at Boston University.

Dario Amodei: DeepSeek Has Not Made a Unique Breakthrough

DeepSeek does not “do for $6M what cost US AI companies billions”. I can only speak for Anthropic, but Claude 3.5 Sonnet is a mid-sized model that cost a few $10M’s to train (I won’t give an exact number).

Also, 3.5 Sonnet was not trained in any way that involved a larger or more expensive model (contrary to some rumors). Sonnet’s training was conducted 9-12 months ago, and DeepSeek’s model was trained in November/December, while Sonnet remains notably ahead in many internal and external evals.

Thus, I think a fair statement is “DeepSeek produced a model close to the performance of US models 7-10 months older, for a good deal less cost (but not anywhere near the ratios people have suggested)”.

[…] DeepSeek-V3 is not a unique breakthrough or something that fundamentally changes the economics of LLM’s; it’s an expected point on an ongoing cost reduction curve. What’s different this time is that the company that was first to demonstrate the expected cost reductions was Chinese…DeepSeek’s total spend as a company (as distinct from spend to train an individual model) is not vastly different from US AI labs.

Dario Amodei, co-founder and CEO of AI firm Anthropic, writing in an essay published on his personal website.

DeepSeek founder Liang Wenfeng
Liang Wenfeng, founder of startup DeepSeek, delivers keynote speech during the 10th China Private Equity Golden Bull Awards on August 30, 2019 in Shanghai, China.

VCG/VCG via AP

Kristian Hammond: DeepSeek Could Be AI Disruptor—if They Open Door to Their Techniques

If they are telling the truth and the system can be built on and run on much less expensive hardware, DeepSeek will have a significant impact.

Nvidia currently is so valuable because everyone thinks that we cannot build AI without them. But DeepSeek demonstrates that we can.

Similarly, Open AI has ‘argued’ that we cannot support their software locally, but DeepSeek, again, demonstrates that we can.

And the fact that DeepSeek could be built for less money, less computation and less time and can be run locally on less expensive machines, argues that as everyone was racing towards bigger and bigger, we missed the opportunity to build smarter and smaller.

If they open the door to their techniques, it will be a disrupter as everyone will rush towards small.

Kristian Hammond is a professor of computer science at Northwestern University, where he directs the Center for Advancing Safety for Machine Intelligence.

WedBush Analysts: DeepSeek Is the Temu of the AI World

We continue to view this as a golden buying opportunity that will not change the AI spending trajectory of the AI Revolution […]

However, saying this model was built for $6 million with no Nvidia next generation hardware is likely a fictional story that is at the heart of this debate and drove a massive tech-sell off […]

To this point, any parent that has been in elementary school seeing young kids’ art projects there is only one project on back to school night that is next level and blows the others ones away by miles.

At the time you are amazed an eight-year-old built this project by themselves (and also how bad your own kid’s projects compares)…then two days later you find out that kid’s parent is a well-known architect/engineer in town and actually built the project themselves while their kid watched TV. Now you view the project dramatically differently.

[…] We have spoken to many enterprises heading down the AI path/use cases in no way does DeepSeek make them flinch on spending seven/eight figures of their IT budgets over the next year on AI initiatives.

DeepSeek does not disrupt the $2 trillion of AI Cap-Ex set for the next three years with Nvidia, Microsoft, Google, Amazon, Palantir, ServiceNow, Salesforce, Oracle, TSMC, and others front and center poised to benefit. LLM models will become cheaper and more commoditized over time…the value however is in the data, use cases, algorithm reasoning, and storage for AI.

In a nutshell, DeepSeek created an awesome LLM model (and credit to its software developers) however this Chinese AI small lab/LLM model is not bringing down the entire US tech ecosystem with it.

Just like Temu was the “Amazon model destroyer” a few years ago…Amazon’s team adjusted and now look where Temu and Amazon both are sitting.

From a research note published by equity analysts at the financial firm Wedbush Securities.

share Share facebook pinterest whatsapp x print

Similar Content

Google will no longer set diversity hiring targets
Google will no longer set diversity hiring targets
Kamala Harris
Time owner attacks Kamala Harris for refusing interview
Newborn baby
Snoo killer? Startup aims to undercut $1700 product new parents swear by
BetMGM Bonus Code
BetMGM bonus code NEWSWEEK200: Get $200 bonus for Jones-Miocic, UFC 309
Keon Coleman
Bills Expected to Get Massive Offensive Boost Ahead of Crucial Lions Match
Mysterious man wishes plane passenger happy birthday—then secret emerges

Leave a Reply

Your email address will not be published. Required fields are marked *

In-Depth Insights: News Beyond the Surface | © 2025 | Daily News