A little-known Chinese artificial intelligence startup shook the tech world this weekend by releasing an OpenAI-like assistant, which shot to the No.1 ranking on Apple’s app store and caused American tech giants’ stocks to tumble. From Washington’s perspective, the news raised an immediate policy alarm: It happened despite consistent, bipartisan efforts to stifle AI progress in China. Both President Donald Trump and former President Joe Biden have made global competition — and, specifically, boxing out China — the centerpiece of their tech and AI policies. Somehow, a small and nimble Chinese company just jumped ahead anyway. In tech terms, what freaked everyone out about DeepSeek’s R1 model is that it replicated — and in some cases, surpassed — the performance of OpenAI’s cutting-edge o1 product across a host of performance benchmarks, at a tiny fraction of the cost. The business takeaway was straightforward. DeepSeek’s success shows that American companies might not need to spend nearly as much as expected to develop AI models. That both intrigues and worries investors and tech leaders. Some, like Microsoft CEO Satya Nadella, celebrated what they saw as the commodification of AI — a future where a wide range of companies can deploy the technology way more cheaply. Investors, clearly, had doubts: Quite a few tech stocks took a plunge, especially the highly valued chipmaker Nvidia, the biggest beneficiary of the current AI infrastructure rush. The policy implications, though, are more complex. Washington’s rampant anxiety about beating China has led to policies that the industry has very mixed feelings about. On one hand, most tech firms hate the export controls that stop them from selling as much to the world’s second-largest economy, and force them to develop new products if they want to do business with China. If DeepSeek shows those rules are pointless, many would be delighted to see them go away. On the other hand, anti-China, protectionist sentiment has encouraged Washington to embrace a whole host of industry wishlist items, from a lighter-touch approach to AI rules to streamlined permitting for related construction projects. Does DeepSeek mean those, too, are failing? Or does it trigger a doubling-down? DeepSeek’s success truly seems to challenge the belief that the future of American AI demands ever more chips and power. That complicates Trump’s interest in rapidly building out that kind of infrastructure in the U.S. Why pour $500 billion into the Trump-endorsed “Stargate” mega project — and why would the market reward companies like Meta that spend $65 billion in just one year on AI — if DeepSeek claims it only took $5.6 million and second-tier Nvidia chips to train one of its latest models? (U.S. industry insiders dispute the startup’s figures and claim they don’t tell the full story, but even at 100 times that cost, it would be a bargain.) “The leading American companies are all part of this phenomenon whereby AI models continuously get smaller, faster, cheaper,” said Gregory Allen, director of the Wadhwani AI Center at the Center for Strategic and International Studies think tank. He points out that DeepSeek techniques, powered by more impressive equipment, could also turbocharge American AI development: “If Deepseek is kind of illustrative of the floor of capabilities being raised — what do you get for not that much money — in parallel, there is this phenomenon of the ceiling being raised.” Tech companies, of course, love the recent bloom of federal support, and it’s unlikely they’ll drop their push for more federal investment to match anytime soon. Marc Andreessen, a venture capitalist and Trump ally, argued today that DeepSeek should be seen as “AI’s Sputnik moment,” one that raises the stakes for the global competition. That would strengthen the case that some American AI companies have been pressing for the new administration to invest government resources into AI infrastructure (OpenAI), tighten restrictions on China (Anthropic) and ease up on regulations to ensure their developers build “artificial general intelligence” before their geopolitical rivals. David Sacks, fellow VC and Trump’s AI czar, invoked the model to pump up Trump’s approach and rebuke the Biden administration for adding too much friction to AI development. “DeepSeek R1 shows that the AI race will be very competitive and that President Trump was right to rescind the Biden EO,” Sacks posted on X. “I’m confident the U.S. can lead, but we can’t afford to be complacent.” It’s too early to tell how Congress will respond, but the past offers some clues. In many ways, DeepSeek mirrors another Beijing tech breakthrough that blindsided both policymakers and Silicon Valley: when Huawei in 2023 seemingly sidestepped U.S. sanctions to drop a smartphone powered by a 7-nanometer chip, the most advanced China had ever produced at the time. Just like that launch, which happened as then-Commerce Secretary Gina Raimondo was visiting China, Beijing may be sending a well-timed message to the Trump administration. In an interview last year, Deepseek’s CEO said the number one challenge facing his company is export controls. Trump has ordered agencies to study existing restrictions for possible loopholes and report back by April 1. “The story is the same in both cases, which is that the technology is legitimate and impressive, but the timing is politically motivated,” said Allen. “[This] is an attempt to try and change the political narrative in a new presidential administration — to try and say export controls have already failed, we should give up on them. That is the plan here.” That strategy backfired during the Huawei chip saga. If anything, China hawks in Congress were further galvanized by that and other Chinese tech breakthroughs slipping past U.S. restrictions, often reacting by calling for a tougher crackdown. Allen sees reason to not let up on China this time either. Deepseek’s innovation mainly comes from a technical mechanism, he said: “They have this architecture that allows them to extract more IQ points per chip.” But when that multiplier is applied across China’s limited chip count versus a massive supply, the result is a far greater intelligence advantage for the U.S. By that logic, so long as the U.S. keeps denying China access to coveted American tech and fixing its workarounds, it will have the upper hand. “Their improvements — architectural and algorithmic — are available to Western countries. And our advantages of superior chip architecture remain denied to them,” said Allen. Deepseek’s development could also mean U.S. export controls need more time to bite, not necessarily that they have failed. “Right now, particularly with the stock market behaving the way it is, we're really at peak freak out,” Bill Drexel, a fellow at the Center for a New American Security, told POLITICO Tech host Steven Overly. “[China] has gone a lot faster than we expected. They've closed more gaps than we anticipated, but it’s still going to be difficult for them to compete in the years ahead.”
| 
