CN DeepSeek buzz puts tech stocks on track for $1 trillion wipeout

archive

(Bloomberg) — Chinese artificial intelligence startup DeepSeek rocked global technology stocks Monday, raising questions over America’s technological dominance.

Buzz grew over the weekend about DeepSeek’s latest AI model being cost-effective while running on less-advanced chips, casting doubt on the validity of the rich valuations for companies like Nvidia Corp. (NVDA), which has led the global AI stock boom as its chips have been seen as essential to the technology. Shares of the Santa Clara, California-based firm slid 10% in premarket trading on Monday.

Nasdaq 100 futures tumbled 3.4%, while contracts on the S&P 500 fell 2% as of 5 a.m. in New York. In Europe, tech stocks led market losses, with shares of chip equipment maker ASML Holding NV down 11%. The Cboe Volatility Index, known as the VIX, spiked higher. The Nasdaq 100 and Europe’s Stoxx 600 technology sub-index were together set for a market capitalization wipeout of roughly $1 trillion, if the losses hold.

“DeepSeek shows that it is possible to develop powerful AI models that cost less,” said Vey-Sern Ling, managing director at Union Bancaire Privee. “It can potentially derail the investment case for the entire AI supply chain, which is driven by high spending from a small handful of hyperscalers.”

Roughly 200,000 Nasdaq 100 futures contracts changed hands by 4:45 a.m. New York time, about four times more than the 30-day average for this time of day, according to data compiled by Bloomberg.

The AI model from DeepSeek — founded by quant fund chief Liang Wenfeng — is widely seen as competitive with OpenAI and Meta Platforms Inc.’s latest offerings. Lauded by investor Marc Andreessen as “one of the most amazing and impressive breakthroughs,” DeepSeek’s app shows its work and reasoning as it addresses a user’s written query or prompt. Released last week, the product is now at the top of Apple Inc.’s App Store rankings, with users praising its transparency.

Chinese AI-related stocks reacted positively, with mainland-listed Merit Interactive Co. among those surging by their daily limits. Merit is among those with the clearest links to DeepSeek after stating in an earlier filing that it had incorporated the homegrown AI firm’s model into marketing. In Hong Kong, the Hang Seng Tech Index climbed as much as 2% ahead of Lunar New Year holidays this week.

AI trades slumped elsewhere as investors rethought assumptions on computing power and energy. Siemens Energy AG, one of the few AI winners in Europe, slid 20%.

The DeepSeek product “is deeply problematic for the thesis that the significant capital expenditure and operating expenses that Silicon Valley has incurred is the most appropriate way to approach the AI trend,’ said Nirgunan Tiruchelvam, head of consumer and internet at Singapore-based Aletheia Capital. “It calls into question the massive resources that have been dedicated to AI.”

The decline in Nasdaq futures (NQ=F) comes at the start of a big week for earnings from major tech companies including Apple (AAPL) and Microsoft Corp. (MSFT) Profit growth is expected to have slowed while valuations remain inflated, once again causing concern over the large AI-driven rally in the sector.

The Nasdaq 100 is trading at 27 times estimated forward earnings, compared with its three-year average of 24 times. Nvidia is at 33 times, though that’s slightly down from its three-year average.

The DeepSeek release raises new doubts, challenging the notion that China’s AI technology is years behind US counterparts. Washington’s trade restrictions had kept the most cutting-edge chips out of China’s hands, but DeepSeek’s model was built using open source technology that is easy to access.

“While current leaders like Nvidia have a strong foothold, it is a reminder that AI dominance cannot be taken for granted,” said Charu Chanana, chief investment strategist at Saxo Markets. “The emergence of China’s DeepSeek indicates that competition is intensifying, and although it may not pose a significant threat now, future competitors will evolve faster and challenge the established companies more quickly. Earnings this week will be a huge test.”

—With assistance from Audrey Wan, Vlad Savov, Newley Purnell, Henry Ren and Michael Msika.

About the author...

Abhishek Vishnoi

1737975427784.png
 
* Cantor Fitzgerald On NVIDIA: The DeepSeek “Announcement Is Actually Very Bullish With AGI Seemingly Closer To Reality And Jevons Paradox Almost Certainly Leading To The AI Industry Wanting More Compute, Not Less”
* Has DeepSeek Ended The AI Run Led By NVIDIA? A Deep Dive Into What’s Next For The Magnificent Seven

There has been a lot of noise about Nvidia and their stock dropped 15% today, but I believe they will be fine as the shovel/pickaxe seller in the short term, and it's really OpenAI and others that have to watch out.

For anyone who hasn't heard of it, Jevons paradox is the idea that falling energy/resource prices increase demand/consumption. So dropping the token cost by 95-98% isn't going to decrease the demand for AI accelerators and GPUs by that much. It will make it easier to provide higher quality cloud/search AI without burning billions of dollars.
Basically this. There will be some winners and losers but better approaches like this will spread throughout the industry, increase quality, and lower costs. Which could be bad for OpenAI but good for anyone forking off millions/billions to OpenAI.
 
Last edited:
Why does anyone just believe claims from the Chinese in regards to tech? They have no oversight or independent evaluation and have been shown to outright lie about their “advanced” technology so many times it’s a running joke.

The timing is especially suspicious since Trump announced major investment in this area. It’s like everyone forgets that the Chinese make claims like this constantly.
I noticed a serious uptick of Chinese shilling especially in the legacy media immediately following Trumps win.

There were a bizarre number of articles singing the praises of the Xiaomi SU7 as the best EV ever.

This is a car the Chinese brag about being faster than Tesla and Porsche but when actually driven on a track the brakes failed on the 2nd lap, the air bags didn't deploy and the drivers seat broken in half.
 
Basically this. There will be some winners and losers but better approaches like this will spread throughout the industry, increase quality, and lower costs. Which could be bad for OpenAI but good for anyone forking off millions/billions to OpenAI.
As to whether you should directly invest your money into deepseek?

Fuck no. If they end up with a viable product, you just won’t get paid your investment back.

Chinese companies will or can not pass basic audits and get listed as VIE’s, or variable interest entities.
 
Uh oh shit is going down

KEY POINTS
  • DeepSeek on Monday said it would temporarily limit user registrations “due to large-scale malicious attacks” on its services.
  • The Chinese AI startup recently toppled OpenAI’s ChatGPT from its title of most-downloaded free app in Apple’s App Store.
DeepSeek on Monday said it would temporarily limit user registrations “due to large-scale malicious attacks” on its services, though existing users will be able to log in as usual.

The Chinese artificial intelligence startup has generated a lot of buzz in recent weeks as a fast-growing rival to OpenAI’s ChatGPT, Google’s Gemini and other leading AI tools.

Earlier on Monday, DeepSeek took over rival OpenAI’s coveted spot as the most-downloaded free app in the U.S. on Apple’s App Store, dethroning ChatGPT for DeepSeek’s own AI Assistant. It helped inspire a significant sell-off in global tech stocks.

Buzz about the company, which was founded in 2023 and released its R1 model last week, has spread to tech analysts, investors and developers, who say that the hype — and ensuing fear of falling behind in the ever-changing AI hype cycle — may be warranted. Especially in the era of the generative AI arms race, where tech giants and startups alike are racing to ensure they don’t fall behind in a market predicted to top $1 trillion in revenue within a decade.
 
MovieBob and his cronies are having a ball.

tech3.jpg

tech.jpg

Who did they steal it from then?
I cannot register so I cannot test DeepSeek myself yet, but people who have are impressed by the cultural specificity of the AI. One reviewer told DS to "Write a Year Of The Snake Salutation in the Cantonese dialect, addressed to Hong Kong people. Do not use too many rhyming puns". And I give the output a passing grade. What's more DS offers its "thought process" of the task. The same reviewer asked DS to write a Japanese Tanka poem in Nara-period language. I cannot verify the stylistic congruence, but I'm impressed by the "thoughts process" part (e.g. DS knows that sakura flowers are often symbolic of "fleeting happiness" in Japan)

I don't think either ChatGPT nor anything by Google or Microsoft is so attuned to the idiosyncrasies of Far East languages.
 
Last edited:
Yes buy many Chinese products you know where the Chinese invest in American real estate because real estates always worth something that's why they never invest in the Chinese museum companies.
 
ALL HANDS TO THE WAR ROOM ALL HANDS TO THE WAR ROOM THE ZUCK IS IN FULL SPERG

Meta is reportedly scrambling ‘war rooms’ of engineers to figure out how DeepSeek’s AI is beating everyone else at a fraction of the price​

Mark Zuckerberg’s Meta assembled four war rooms of engineers to determine how a Chinese hedge fund managed to release an AI game-changer that may already rival its own technology, The Information reported.

DeepSeek, an AI startup backed by hedge fund High-Flyer Capital Management, this month released a version of its AI chatbot, R1, that it says can perform just as well as competing models such as ChatGPT at a fraction of the cost.
The potentially groundbreaking, open-source tech has called into question the gargantuan AI investments made by American companies and has put Meta’s AI-dedicated team on high alert.

Meta AI infrastructure director Mathew Oldham has reportedly told colleagues that DeepSeek’s newest model could outperform even the next version of Meta’s Llama AI, which Zuckerberg said could be released in “early 2025,” The Information reported Sunday. The report cited two employees with direct knowledge of Meta’s efforts to catch up.

Of the four war rooms Meta has created to respond to DeepSeek’s potential breakthrough, two teams will try to decipher how High-Flyer lowered the cost of training and running DeepSeek with the goal of using those tactics for Llama, the outlet reported, citing one anonymous Meta employee.
Among the remaining two teams, one will try to find out which data DeepSeek used to train its model, and the other will consider how Llama can restructure its models based on attributes of the DeepSeek models, The Information reported.

Meta did not immediately respond to Fortune’s request for comment.

“We regularly evaluate all competitive models in our development process and have done so since [the company’s] Gen Al [group] was formed,” a Meta spokesperson said in a statement to The Information. “Llama has been foundational in establishing the ecosystem for open-source AI models and we couldn’t be more excited to extend this leadership with the upcoming release of Llama 4.”

On Friday, Meta CEO Zuckerberg announced that the company would spend as much as $65 billion on projects related to AI in the coming year, including construction of a large data center and more AI hires.

The announcement came just days after OpenAI, in partnership with SoftBank, Oracle, and others, announced a $500 billion White House–backed AI infrastructure project called Stargate that will build dozens of new data centers across the U.S.
 
  • Thunk-Provoking
Reactions: Bland Crumbs
Meta is reportedly scrambling ‘war rooms’ of engineers to figure out how DeepSeek’s AI is beating everyone else at a fraction of the price

Everyone is wondering "How could the chinese do this? Did they steal the software? Did we have an intelligence leak? Double agents? HOW?"

When the answer is actually really fucking simple. Deep Seek was able to outdo everyone else because they don't dedicate 50% of their development time trying to work out how to lobotomize the AI so it doesn't accidentally have nono thoughts.

When you dedicate 50% of development time trying to build an AI and the other 50% figuring out how to lobotomize it, yeah its gonna hold up development time.

"But Muh tinaman square."

Yes. The difference is that the chinese have 2 rules. Don't trash talk the CCP, don't mention tianaman square.

Western AIs have to account for 500 seperate untouchable groups and historical events and willingly ignore or falsify data in regards to statistics for every single one of them.

One requires considerably less doublethink than the other.
 
Western AIs have to account for 500 seperate untouchable groups and historical events and willingly ignore or falsify data in regards to statistics for every single one of them.
There's also the little thing called competency crisis beginning to bite the ass of the line go up middle manager class who thought they could treat workers as interchangeable cogs.
 
Back