Technology stocks are hammered today over AI concerns and a failure of US export bans.
China’s DeepSeek AI Models Raise Doubts
The Wall Street Journal reports China’s DeepSeek AI Models Raise Doubts Over U.S. Tech Dominance
Global chip stocks slumped Monday after Chinese artificial-intelligence company DeepSeek said it had developed AI models that nearly matched American rivals despite using inferior chips, raising questions about the need to spend huge sums on advanced gear provided by Nvidia and other tech giants to train AI models.
DeepSeek said last week that the performance of its latest R1 model was on par with OpenAI’s o1-mini model that the ChatGPT maker released in September. The announcement came after DeepSeek said in a late-December report that it used a cluster of more than 2,000 Nvidia chips to train its other V3 model, compared with the tens of thousands of chips that are normally used for training models of a similar size.
The company said training one of its latest models cost $5.6 million, compared with the $100 million to $1 billion range cited last year by Dario Amodei, chief executive of AI company Anthropic.
“There is a new AI challenger in town and investors are spooked at what they’ve discovered,” AJ Bell investment director Russ Mould wrote in a note to clients. “Its assistant is free to use and runs off lower-cost chips and less data—implying a major challenger to the established AI names in the West.”
DeepSeek’s perceived success risks intensifying the AI war between the U.S. and China, Quilter Cheviot’s Ben Barringer said in a note to investors, particularly after the recent announcement of Stargate—a joint venture between OpenAI, SoftBank Group, Oracle and MGX to build data centers in the U.S. for OpenAI.
News Corp, owner of Dow Jones Newswires and The Wall Street Journal, has a content-licensing partnership with OpenAI.
Is AI Spending Justified?
AI spending is just one of the concerns that Sends U.S. Tech Stocks Reeling.
The sudden popularity of a Chinese artificial intelligence model called DeepSeek pummeled stocks Monday, with the tech-focused Nasdaq composite index down 2.7 percent in morning trading.
Several tech companies that have banked on a surge of AI interest sold off, with U.S. chipmaker Nvidia down more than 12 percent at the opening bell. Microsoft lost around 3.5 percent, Meta Platforms lost 1.6 percent, and Oracle dropped more than 7 percent.
Analysts said the sell-off underscores anxieties about whether the massive spending on artificial intelligence ― and the specialized chips, data centers and related power infrastructure ― are justified. Nvidia has exploded in value in recent years because it dominates the market for the chips at the center of the global AI race. It’s now one of three companies with a market capitalization above $3 trillion.
But some analysts have raised questions about how soon those investments will translate into straightforward moneymaking use cases for U.S. companies. The emergence of a low-cost Chinese competitor adds to skepticism over those investments.
The emergence of the model from China has escalated concerns about the U.S. adversary’s tech prowess amid a global race to develop advanced artificial intelligence. China’s advancements in the technology are viewed as a national security threat to the United States, and the success of DeepSeek has only exacerbated those concerns.
DeepSeek is a China-based start-up that last week launched a free AI assistant that it says can operate at a lower cost than American AI models like ChatGPT. The company was founded in 2023 by Liang Wenfeng, co-founder of the hedge fund High-Flyer. By Monday, it had rocketed to the top of downloads in the Apple Store.It also reported outages and limited registrations due to “large-scale malicious attacks” amid the uptick in interest.
President Donald Trump has said that the United States needs to remain competitive with China in developing artificial intelligence, and he appeared focused on the technology during his first week in office. He told reporters Saturday that he was considering using emergency powers to provide the “tremendous energy” that U.S. companies need to develop AI models.
We’re already leading,” Trump said on Air Force One. “Very shortly, we’re going to be leading by a lot.”
Longtime technology investor Marc Andreessen, a Trump ally, called DeepSeek’s AI model “one of the most amazing and impressive breakthroughs I’ve ever seen” and “a profound gift to the world” in a post on X.
Leading tech firms have spent billions building out artificial intelligence technology for sale to large businesses. Meta chief executive Mark Zuckerberg said in a Facebook post last week that his company plans to invest between $60 billion and $65 billion on AI and build a massive data center. OpenAI, Oracle and SoftBank recently announced a Trump-supported joint venture, called Stargate, that seeks to spend up to $500 billion building out data centers to support AI projects.
China’s DeepSeek Turns AI on its Head
Sherwood reports China’s DeepSeek turns AI on its Head
DeepTrouble
The weekend buzz around the large language model — the fact that it “thinks” before it speaks, shows its workings, and matches OpenAI’s most powerful model, the o1, on a range of metrics — seems to have left much of Silicon Valley wowed and worried, in almost equal measure.Per DeepSeek’s own figures, the R1 model outperforms the OpenAI o1 on a variety of key tests, shining particularly brightly in math, where it beats the latest model from Sam Altman’s company on three different tests. While it’s less consistent on coding and language tests — it fared particularly badly on the “SimpleQA” (not shown in chart above), a test evaluating the simple factual accuracy of the info that LLMs spit out — the differences are fairly slim, making the cost-effective R1 look impressive.
The Chinese company’s slimmed-down training costs, use of cheaper chips, API, and open-source model have hauled the endless drive for more chips and compute that’s driven much of the market for the last 18 months into question. Meta is planning to spend more than $60 billion on capital expenditures just this year.
At a time when people are wondering if we can trust TikTok due to Chinese government ties, many have similar questions about DeepSeek. Tech evangelist Marc Andreessen was among those singing R1’s praises over the last few days — though he may not have asked it about Tiananmen Square yet.
Another Sanction Failure
Biden placed numerous exports bans on chip technology to prevent this from happening.
Instead DeepSeek said in a late-December report that it used a cluster of more than 2,000 Nvidia chips to train its AI.
No one should be surprised by this.
To Those Hard of Learning, Here’s a Repeat Lesson on Why Sanctions Fail
On September 26, 2024, I commented To Those Hard of Learning, Here’s a Repeat Lesson on Why Sanctions Fail
Let’s discuss a claim that sanction failures are due to a lack of political will.
Robin Brooks on X: “When someone tells you that sanctions can’t and won’t work, that’s basically pro-Russian propaganda. Are we seriously to believe that nothing can be done to stop the shameful flood of transshipments to Russia via Central Asia? Come on. This is just about a lack of political will.“
I am pretty sure that “someone” is me because we have gone round and round on this.
When someone tells you that sanctions do work. Ask them for evidence.
The above post was on oil-related sanctions. The next article is how and why chip sanctions failed.
On August 26, 2024, I commented China Gains Secret Access to Nvdia Microchips by Renting Computers
The US has blocked export of Nvdia chips to China. But where there’s profit, there’s a way.
Know Your Customer’s Customer’s Customer
China sets up an AI company in Singapore. AI developers buy cloud time through a subsidiary that further masks the operation by paying in Bitcoin.
In turn, the subsidiary buys time from a company Dubai or Singapore that hosts the servers.
US politicians are outraged. But some of us are amused knowing full well that sanctions don’t work. So instead of cloud profits going to US corporations, the profits go to Saudi Arabia, Singapore, Dubai, and South Korea.
Only Amazon is forced to “know your customer”.
Where Art Thou EU?
Trump secured pledges to spend $500 billion on data centers.
However, If the benchmarks are accurate, China caught up on AI development by renting NVDIA chips.
Meanwhile, the EU is doing nothing but attempting to regulate AI to death.
For discussion, please see The EU Will Suffer Because it Is Hostile Territory for the AI industry


The Chinese product is supposedly rife with hallucinations.
so exactly the same as chat- gpt
Clearly, DeepSeek does not have any DEI hires on the staff.
never miss a change to blame black folks for all the worlds problems
AI is a tough sell because all of it’s users are brainless.
Deep Seek can’t discuss Tiananmen Square or the killing of all those sparrows by the CCP that caused millions to die.
Things will shake out. There is lots of room in AI for everyone to censor what they need to censor.
most business that are investing in AI arent doing so because they want to build bots that talk about politics. the trillion dollar wall street investment is with the hope that they can create business applications that can lead to cost cutting. deepseek works just as well and siginfigantly cheaper and open source… undercutting investers wild dreams
Oh thank God.
Value investing is back
A big aspect of this AI situation is Altman changing the now misnamed openAI to closed source private operation killing innovation. These “tech titans” should know better but nothing like short term greed interfering with long term what’s best for the US. What a massive squandering of resources.
Lina Khan and Elon Musk were correct, let’s give them the credit they deserve.
DeepSeek trade secret REVEALED: All queries are actually referred directly to classrooms of Chinese 4th graders who provide the needed info in real time. This technology is so vastly superior that all other AI attempts are doomed from the start.
Lol. Command economy at its best. What better use for 1 billion 4th graders?
The AI bubble was going to pop sooner than later. My significant other has tried several AI tools for writing. One highly rated tool gave the same character 3 different names within 10 pages. Another used the same descriptions frequently. She is attempting to feed one AI tool’s output into another. She finds there is little time savings since more is now used for distilling good ideas from mundane, as well as increased time editing in general.
Same issue with AI coding
On a lighter note we find out that the Covid did start in the chinese lab with grants from the US and Fauci. They were trying to treat chinese bats . Anyone who spoke out on the social media about these things were locked out for spewing misinformation marginalized and ridiculed . Millions of lifes lost many millions suffered extreme sickness from covid or from the vaccine affects . You wonder why it took 4 years to get the truth out . To think we are worried about a few thousand illegal invaders being rounded up and flown out to where they came from . Go figure .
And Fauci got a pardon. Disgraceful
So did 1600 of my favorite violent morons! Pardons for all!
Ratcliffe confirms what everyone else has known for 5 years —
The Covid virus came from Wuhan lab where they were conducting gain of function research on bat viruses while searching for a SARS Vaccine.
https://x.com/
More to come…
Not my post, but relevant and I thought a good summary:
First we have the grant applications from the NIH to EcoHealth Alliance and Dr. Peter Daszak for Gain of Function research on SARS COV1 in 2014 that continued all the way up through 2019. The sub award was to the Wuhan Institute of Virology. Specifically this grant was for conducting research to determine the spillover propensity to humanized mice (i.e. humans) using the SARS COV 1 backbone. Specifically, they cut the spikes off the original SARS COV1 virus and swapped it with a spike that was more capable of infecting humans. This is exactly the definition of gain of function research that the Obama administration put a moratorium on. Here is the link to the NIH grant application if anyone doubts (https://reporter.nih.gov/project-details/8674931)
Second, we know through our own intelligence agencies that researchers at the WIOV were infected as early as November 2019. Here is the link to a CBS news article covering this (www.nbcnews.com/…)
Third, the spike protein on SARS COV2 was a commonly known spike protein (ACE2) that was used in GoF research because of its likelihood to bind to ACE2 receptors in the lungs. The grant application clearly spells this out. Additionally, the genetic sequencing between SARS1 and SARS2 is so similar that we would have seen clusters of infections near known bat habitats if it were a natural event, rather than it conveniently popping up in and around the WIOV without any other recorded infections..
Fourth, the actions of the NIH are very telling. In May of 2020, Dr. Fauci publicly stated that the evidence “was very very strong” supporting a natural zoonotic spillover event, even though the NIH can’t produce a shred of evidence to support this claim. A good scientist evaluates all evidence before making a conclusion and putting their reputation on the line. I find it extremely odd that the NIH and Fauci would issue a definitive statement as such, without having a wealth of evidence to support it.
The CIA’s own analysis, that was conducted under the Biden administration, claims that it most likely came from the WIV. Here is a link to the New York Times if you don’t believe me (www.nytimes.com/…)
Sorry, I believe nothing coming from the NY times they have been exposed as habitual liers . They are nothing but a liberal government mouth piece . They lied about Hunters Laptop , Russia collusion , covid to name a few . NY times and W Post are nothing more than a propaganda arm of the US government .
The only thing truth that ever leaks out of the NYT or WaPo is when they accidentally make an admission against interest usually around the 19th paragraph .
Will Trump ban DeepSeek for “national security reasons?
They compete with my stuff. You bet he will.
Thanks for confirming Mr. President.
its open source. it runs locally. There is not way to “ban” it
We had a choice: harness the talents of millions of brilliant, educated Chinese for profit, or tariff the hell out of everyone and fall into obscurity. We chose wrong out of egoism and ignorance.
Idiotic post. China tariffs us more than them on most items. But faux outrage is my fave.
All goods imported into China are subject to the nation’s value-added tax (VAT) of either 13 percent or 17 percent. The 13 percent tax is available for certain goods that fall mainly within the categories of agricultural and utility items, while the 17 percent tax applies to other goods subject to the VAT tax.
If it is a VAT it also applies to Chinese domestic goods. USA has sales taxes in many states.
Only an army of H1Bs can save us! Americans are too dumb.
(double post)
Maybe NVDA isn’t really worth 30 times sales.
LOL…
And also: Because Trump wants to use military—not civilian—flights to transport deportees, “American taxpayers are paying 300% more for Trump’s immigration plan than they should.“
The important part isn’t deporting immigrants. It’s putting on a popular show for the rednecks.
The chains were my idea!
Those military planes would be doing touch and go practice if they did not deport illegal aliens.
Something that’s getting lost in yesterday’s was-it-or-wasn’t-it US-Colombia trade war: “Colombia accepted 475 deportation flights from the US from 2020 to 2024.“
Trump won this one. He was ahead of the MSM the whole time and the story is over. Normally, I scrutinize “stories” like Colombia refusing the planes because there’s always more to the story. It’s a moot point to rebut Trump because, as I said, Trump won this one.
The planes didn’t land. The stinky guy failed. That’s what I keep him around for.
Hope you have a little more cash for coffee. I know you can’t afford eggs now.
Depending on how good DeepSeek turns out to be, this could be our generations ‘Sputnik’ moment where we suddenly realize that that we do not have anywhere near the tech lead we think we do.
Whether that spurs an equivalent prioritizing of research equivalent to what happened in the space race remains to be seen.
Sputnik led to a huge increase in funding for primary and technical education. I’d love to see that happen again but I don’t think we’re interested anymore.
Sanctions and tariffs work. Look at the price of the rouble and yuan since they were announced.
The goal of tariffs was to increase export – It failed
The goal of sanctions was to stop exports – failed everywhere
We just need more sanctions to reach the Golden Age!
I thought the goal of tariffs was to force deadbeat Marxist countries to take back their criminal aliens here in the US?
The goal of sanctions and tariffs is to reduce the advantages of low cost (i.e. low wage) nations from eroding the wages in the issuer nation. Most studies focus and GDP, or sales, as a measure of success. What you need to look at the pace of reduction of discretionary income for the average worker post tariffs and the time it gives the issuer nation to rebalance the economy. Low tariffs always benefits the rich people as they always benefit from labor arbitrage. Tariffs gives local substitutes a better chance and gives you time to adjust. You need to look at the bigger picture over the longer term for the median consumer, not the rich equity holder…
China is doing amazing things in biotech, too. In some cases, they are ahead of US efforts and US pharmas are now in-licensing their stuff.
Scientists from China come to the US where they provide cheap labor to academic labs, and then return to China with the know-how. Much easier to test experimental therapeutics there, and so they are in the lead in some areas. What to do?
John Thune recently lamented our diminished number of Navy ships. I think he’s trapped in the ’90’s. Or listens to Hugh Hewitt – also trapped in the ’90’s.
I have read that China is investing heavily in photonic chips, so they’re probably not worried about catching NVIDA et al with regards to traditional semiconductor chips. They’re trying to get to next level ASAP.
So are we, of course, and everyone should expect Chinese companies will periodically make these announcements that remind us that they got tech mojo.
And they avoid the woke mind virus.
China knows how detrimental wokeness can be to a country, which is why they directly support it here in the US.
They call it baizuo and laugh at us for it.
Photonic chips execute algorithms, like silicon ones.
So that has more to do with TSMC and Intel than Nvidia.
Photonic chips are what’s coming after semiconductor technology.
I would imagine China, like America, is trying to incorporate photonic chips with quantum computing.
That would be the Holy Grail.
I don’t think the NVDA bubble has popped. This is just a temporary correction.
I think the whole AI bubble has a ways to go. Hundreds of billions more need to be spent and wasted before it has played itself out and pops.
Another Y2K dot.bomb collape?
What companies are peddling today is not AI. It’s a very good parlor trick, based on statistics. Geoffrey Hinton, AI pioneer, once said that the approach we’re using today will not achieve AGI, though I can’t find that quote anymore.
Necessity, the mother of AI invention…and this is open source! Hmm, a financial guy I know says the Chinese lie about everything and thinks China is just focusing on “low end” AI. Could be true, or could be yet another “Exceptional American” whistling past the graveyard? I don’t know, but the story and post is fascinating. Thanks, Mish.
China’s AI is 3 kids in a raincoat.
If not China, there will be other competitors. Ten dedicated software developers could build a system that beats current techniques by 10x, and it will happen over the coming years. Last time I checked, the cost of supporting 10 exceptional software developers and a high powered building of compute servers is well below $500 billion.
There are politicians to be bribed out of that… the smarter ones are expensive.
I saw a quote on CNBC on this today, “I’d have to be crazy to have my developers keep paying ChatGPT when Deepseek is cheaper.”
It’s amazing what you can do with enough cheap quasi-slave labor aint it?
NVIDIA has lost $500 billion in market cap today.
600 at one point. Let the bubble burst
My SPY puts are up 24% today but still underwater. I have until Jun 30 to cash out though…..wait till Feb 1 tariff mania starts and watch it fly!
Bloated market. Antagonistic Fed moving forward. A long way down. All asset prices are absurd.
An eccentric chemistry professor in upstate NY has said the market is overvalued by 3x for the past 2 years. Nothing to do with Trump and Tariffs. Greenspam turned the whole thing into a Greater Fool Ponzi.
Powell is gone in June 2026. I’ll put someone in there that will get rates back to 0… maybe even let the stinky guy send out some personally signed checks. It makes him feel like he’s in charge.
This is why H1Bs are the path to American Greatness, just like the old days.
There is a great discussion on H1Bs between Lex Fridman and Andreesson on Fridman’s youtube channel.
Andreesson is in the club. I’m sure he comported himself properly.
It’s been in a massive speculation bubble for many years now over AI.
Now a lot of people are going to lose a lot of money. I presume you are readying your put options on NVidia.
I mentioned NVidia was going to crash last evening in one of Mish’s threads because of this (even posted a link). I had the misses sell her stake plus her Oracle stock late last night even though she didn’t want to because I knew it was going to be a bloodbath.