English

DeepSeek has shattered basis of US AI stock market surge

The bleeding on Wall Street caused by the emergence of the hitherto relatively unknown Chinese AI company DeepSeek has been staunched, at least for now, with the market regaining some of the $1 trillion loss it incurred in the selloff on Monday.

The Icon for the smartphone apps DeepSeek is seen on a smartphone screen in Beijing, Tuesday, January 28, 2025. [AP Photo/Andy Wong]

But much as some would like to dismiss the losses as a mere flesh wound there is no question that the emergence of DeepSeek, providing AI based on chips below the top range and at a lower cost than the US tech giants, has shattered the assumptions on which the AI boom of the past two years has been based.

The key assumption was that with the development of advanced AI chips by Nvidia and massive investments by the major tech giants in the establishment of data centres for AI development, US companies would enjoy global superiority and accumulate continuously elevated profits due to their monopoly position.

This was reflected in the rise of Nvidia which went from a mid-sized company involved in the development of graphics to the largest company in the world by market capitalisation—more than $3 trillion—in the space of just two years.

Nvidia chief Jensen Huang predicted that $1 trillion worth of AI processing centres, using chips from his company, would be built by the high-tech giants over the next few years.

Of course, there was always the risk that some other company would develop a competitive technology. But such risks were ignored in the market hype which saw the tech-heavy NASDAQ index rise by 92 percent in two years, increasing market capitalisation by $14 trillion.

And the last place any significant competition might have been expected to come from was a start-up Chinese firm. After all the US was the dynamic innovator and China was a mere imitator.

To reinforce its position the US, initially under the first Trump administration and to a much greater extent under Biden, imposed a series of export controls and other restrictions on China to deny it access to the top-level chips.

But using lower-level Nvidia chips to which it had access, DeepSeek was able to develop an AI system comparable to, and in some cases better than that produced by the US tech giants. And this was accomplished in little more than two years.

DeepSeek was initially developed by Liang Wenfeng as a side project for his hedge fund in 2023. He began accumulating a stockpile of lower-grade Nvidia chips and developed a strategy for overcoming the export control bans imposed by the US.

This involved recruiting graduates from top Chinese universities and tasking them with developing methods to “train” AI systems without having to use the latest banned Nvidia technology. Liang said he preferred less experienced graduates because they were not bound by the perceived way of doing things.

Whatever the immediate movements on the stock markets, the assumption that deploying massive piles of cash to create new facilities using the most advanced chips creates an unassailable position in the market has gone by the board.

As AI venture capitalist Mike Volpi told the Financial Times (FT): “If you’re Anthropic or OpenAI, attempting to be at the forefront, and someone can serve what you can at a tenth of the cost, that’s problematic.”

The problems extend well beyond the AI industry. They reach into the stock market and the financial system more broadly because the boom in AI stocks has been the driving force behind the rise of Wall Street over the recent period.

The seven largest companies, mainly hi-tech, with Nvidia at the forefront, account for at least 34 percent of the market capitalisation of the S&P 500 index, the highest figure on record according to Goldman Sachs data going back to 1980.

Nvidia alone accounted for 6.8 percent of the S&P 500 on Friday last week, up from 1.1 percent at the end of 2022 when the AI boom began. It contributed almost a quarter of the index’s total return in 2024, according to the Wall Street Journal (WSJ).

Last Monday Nvidia’s share price tumbled by almost 17 percent, leading to a loss of nearly $600 billion in its market capitalisation, the biggest single-day loss by any company in history.

The broader implications of Monday’s events were outlined by billionaire Ray Dalio, the founder of the hedge fund Bridgewater, in an interview with the FT earlier this week.

He likened the present situation to the collapse of the dotcom bubble at the turn of the century with the development of the internet.

Dalio warned that “pricing has got to levels which are high at the same time as there’s an interest rate risk, and that combination could prick the bubble.”

There are of course similarities with what took place a quarter of century ago, but the present situation is much more serious. This is because the past 25 years has seen a massive injection of money into the financial system, the result of so-called quantitative easing by the Federal Reserve and other central banks, the rise of private credit and the escalation of debt to unprecedented levels.

Dalio raised an important distinction between AI and the stock market. He said there was a “major new technology that certainly will change the world and be successful. But some people are confusing that with the investments being successful.”

In other words, while AI has the potential to bring enormous advances in the productive forces of society, its development under capitalism, amid rampant speculation, can bring about a market collapse.

The US high-tech giants have taken a major hit as has the US political and military establishment which considers the suppression of Chinese technological development an existential question for US global dominance.

Accordingly measures to counter DeepSeek are already being rapidly prepared.

Responding to Monday’s events, Trump said: “The release of DeepSeek AI from a Chinese company should be a wake-up call for our industries that we need to be laser-focused on competing to win.”

Given Trump’s record during his first term when he introduced measures to try to bankrupt the Chinese telecommunications company Huawei, the threat contained in his words is clear.

Summing up the significance of DeepSeek’s advances, an FT editorial noted that “China is managing to make technological leaps in AI despite export controls by the Biden administration intended to deprive it of both powerful chips and the advanced tools needed to make them.… Far from stifling Chinese innovation, Washington may have stimulated it.”

That is, it has kicked an own goal to use an analogy from soccer.

The ability of domestically trained Chinse engineers to increase efficiency and develop workarounds “raises questions over whether the technological ‘moat’ established by high-spending US groups such as Meta, Google, OpenAI and Anthropic is as wide and impregnable as they had thought,” the editorial continued.

A recent article in the WSJ cited the remarks of Dmitri Alperovitch, a writer on US-China security issues, which point to the direction of possible counter-measures—an all-out offensive denying Chinese access to any US-made chips.

“The piecemeal half-in, half-out approach to export controls is a failure,” he said.

Michigan Republican Congressman John Moolenaar, chairman of the House Select Committee on the Chinese Communist Party, told the WSJ: “We must work to swiftly place stronger export controls on technologies critical to DeepSeek’s AI infrastructure.”

The propaganda campaign for the escalation of economic warfare against China is already in motion, like a well-oiled machine, spewing out the claim that DeepSeek’s advances are the result of theft.

David Sacks, a big investor in internet and high-tech firms, who has been appointed as Trump’s AI and crypto czar, told Fox News it was “possible” that intellectual property (IP) theft had occurred.

“There’s a technique in AI called distillation… when one model learns from another model [and] kind of sucks the knowledge out of the parent model.

“And there’s substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI models, and I don’t think OpenAI is very happy about this.”

Sacks, however, provided no evidence for this claim. In any case, all AI companies train their models through the use of information culled from the internet, if not through distillation from competitors.

The initial response of OpenAI chief Sam Altman was to call DeepSeek’s RI an “impressive” model and say that it was invigorating to have a new competitor.

Within days, however, he was singing a very different song as the implications of what had happened started to sink in.

A statement issued by OpenAI said it knew that China-based and other companies were constantly trying to distil the models of leading US AI firms.

“We engage in countermeasures to protect our IP… and believe as we go forward that is it critically important that we are working closely with the US government to best protect the most capable models from efforts by adversaries and competitors to take US technology.”

Numerous comments have pointed to the use by OpenAI of technology and intellectual property developed elsewhere and the hypocrisy of its claims that DeepSeek has stolen its IP.

One of the most strident came from American AI journalist and author Ed Zitron in a recent newsletter.

“Personally, I genuinely want OpenAI to point a finger at DeepSeek and accuse it of IP theft, purely for the hypocrisy factor. This is a company that exists purely from the wholesale industrial larceny of content produced by individual creators and internet users, and now it’s worried about a rival pilfering its own goods?”

The same issue was raised, albeit in somewhat more measured tones, by FT columnist John Thornhill.

“As with other Chinese apps, US politicians have been quick to raise security and privacy concerns about DeepSeek. And OpenAI has even accused the Chinese company of possible breaches of intellectual property rights. Given the cases against OpenAI for infringing others’ copyright, though, that might strike some as rich,” he wrote.

Beyond the issue of hypocrisy, there is a much deeper question which goes to the very nature of the capitalist profit system and its fundamental incompatibility with the development of the productive forces of society.

There is no question that AI provides the basis for tremendous economic and social advancement.

But within the social relations of capitalism, based on private ownership of the means of production and private profit, this advance, arising not from the supposed “genius” of the Sam Altmans and Elon Musks but through the collective labour of millions all over the world, is monopolised for profit.

And if that is challenged in any way, then the monopolies and the imperialist governments which serve their interests resort to economic warfare which, inevitably and necessarily, contains within it the germinating seeds of military war.

From this contradiction, between the productive forces and the social relations of capitalism, arises, as Marx explained, both the objective necessity and possibility for the overthrow of the profit system by the working class, whose labour is the source of all wealth, and the establishment of international socialism so that the productive forces can be harmoniously and cooperatively developed for the betterment of humanity.