- The US tried to limit China’s AI advancements through export limits on semiconductors.
- The limitations may have inadvertently fueled the innovation behind DeepSeek instead.
- China’s DeepSeek disrupted the AI industry with claims of more efficient computing at scale.
The US government has for years actively tried to curb China’s access to semiconductor chips, a key component in generative-AI models. Instead, those export limits may have fueled the innovation that led to DeepSeek‘s R1 — a large language model that’s disrupting the American AI industry and the booming economy built around it.
Brian Colello, a tech analyst for Morningstar, said the quote “constraints lead to creativity” came to mind.
“These Chinese models were processor-constrained, so it led to some creative techniques in training, and the DeepSeek model has come out with better-than-expected performance given the processors that it’s been trained on,” he told Business Insider.
DeepSeek disruption
Last week, DeepSeek, a Chinese AI startup, dropped a version of its app fueled by its R1 model. The model appeared to rival those from major US tech companies such as Meta, OpenAI, and Google — but at a much lower cost.
The Chinese company said it spent nearly $6 million on computing power to train its new system, a fraction of what US tech companies have spent on their models.
DeepSeek said its models were trained with fewer and less powerful semiconductor chips than competitors typically used.
Since 2022, US sanctions have made it illegal for the AI-chip-manufacturing leader Nvidia to sell some of its hardware to China, including its most advanced chips. The sanctions aimed to limit China’s advancements in artificial intelligence and military technology.
“Sanctions forced DeepSeek to use H800s, which were less powerful than H100s,” Patrick Moorhead, the CEO of Moor Insights & Strategy, told BI of the Nvidia chips DeepSeek used.
“In a roundabout way, sanctions initiated in the Biden administration motivated DeepSeek to be more creative in how it trained and ran models,” he added. “No one should be surprised, as ‘necessity is the mother of invention.'”
DeepSeek’s success despite sanctions echoes that of the Chinese tech giant Huawei. While Huawei has lost its access to advanced chips in recent years, it launched a line of smartphones that largely replaced Western tech with domestic hardware and software — stunning US policymakers.
Murky training and computing costs
Some experts and analysts who spoke with BI expressed skepticism over DeepSeek’s claims about the cost of the models and the number and type of chips they were built on. It remains unclear exactly which semiconductors were used to train and deploy DeepSeek.
Still, some analysts said the startup showed it’s possible to do more with less when it comes to AI.
The Deutsche Bank analysts Adrian Cox and Galina Pozdnyakova wrote about DeepSeek in a research note published Monday: “They’ve had to squeeze more value out of their software and methods such as chain-of-thought reasoning and using several models at once, instead of just throwing more computing power at the problem.”
Chris Miller, the author of the 2022 book “Chip War,” told BI that the DeepSeek models were impressive but that costs in AI had come down dramatically since 2023, so he did not find the company’s R1 research paper especially surprising.
He added that the idea that DeepSeek was working on a “shoestring budget” was not true, saying the company used a “very narrow definition of training costs.” Miller said it’s “pretty clear that the training cost is an order of magnitude higher” than DeepSeek suggested.
Ineffective chip restrictions
Alexandr Wang, the CEO of Scale AI, said during a CNBC interview on Sunday that DeepSeek had 50,000 H800s, which Miller said would be a “substantial number.” While that number is still much lower than what US firms have, Miller said, it’s likely a lot more than US export officials would’ve wanted a single Chinese firm to accumulate.
Zongyuan Zoe Liu, a senior fellow for China studies at the Council on Foreign Relations, told BI that the developments at DeepSeek suggested AI development in China “seems to be at least on par with the US.”
But she said “we’re still at the beginning of the race” for AI dominance.
“It certainly serves as a good reminder for American policymakers that technology restriction may not work, depending upon what the end goal is,” she added.
Several experts said that they thought the latest developments with DeepSeek could lead to even more semiconductor sanctions on China but that those would not necessarily stop further innovation.
“The US could put sanctions on China all day long,” Colello said, “but there’s always the threat: What if China comes up with some breakthrough anyway?”
Tom Carter contributed reporting.