Speaking at the House GOP Conference meeting, Trump said the release of DeepSeek from a Chinese company should remind U.S. “industries that we need to be laser-focused on competiting to win” amid a global AI race.

Trump said he views it as “a positive” that China had potentially developed a faster and cheaper method of building AI models.

The president said that the U.S. could “hopefully” come up with the “same solution” instead of “spending billions and billions” of dollars on AI development.

DeepSeek wipes $600 billion off Nvidia’s value

Chinese AI startup DeepSeek rocked the U.S. tech sector after it released two AI models, V3 and R1, that appeared to rival those built by leading AI labs at a fraction of the price.

The company is also subject to U.S. export controls, which should have limited its access to Nvidia and other U.S. chipmakers’ top-of-the-line AI chips.

DeepSeek’s R1, which by some estimates was developed at just 3% to 5% of the cost of OpenAI’s newest reasoning model, undermined the dominant Silicon Valley hypothesis that more compute and more advanced chips will deliver better AI models.

The suggestion that AI models that rival Big Tech’s efforts could be built with fewer or less advanced GPUs spurred a market selloff that hit Nvidia particularly hard, wiping $600 billion off the chipmaker’s value in the worst single-day rout ever recorded.

Nvidia has since responded, calling DeepSeek “an excellent AI advancement and a perfect example of Test Time Scaling.”

“DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely available models and compute that is fully export control compliant,” an Nvidia spokesperson said. “Inference requires significant numbers of NVIDIA GPUs and high-performance networking.”

Does DeepSeek spell doom for Nvidia?

Trump has recently backed the $500 billion Stargate Project, co-led by OpenAI and SoftBank, which aims to build out the U.S.’s AI infrastructure.

Silicon Valley has been pumping an eyewatering amount of money into AI development; just days after the Stargate announcement, Meta also said it was increasing its capex spend to $65 billion.

However, suggestions that DeepSeek could spell doom for Nvidia or other billion-dollar AI infrastructure projects could prove shortsighted.

Fortune notes that DeepSeek’s impact could actually increase demand for advanced AI chips—both from Nvidia and rival chipmakers.

This is because of the Jevons Paradox, a phenomenon in which improvements in resource efficiency lead to an overall increase in resource consumption rather than a decrease.

Microsoft CEO Satya Nadella has also pointed to this phenomenon as a potential win for the U.S. AI sector.

However, DeepSeek’s open-source AI bombshell does appear to have re-invigorated the U.S. tech sector’s push toward “AGI” in the global AI race.

OpenAI CEO Sam Altman took to X on Monday to say he would move up some of OpenAI’s new releases in response to DeepSeek’s “impressive model.”

Mark Zuckerberg’s Meta is reportedly also scrambling to assert how DeepSeek was built by assembling four war rooms of engineers to determine how it trained a model that could outperform even the next version of Meta’s Llama AI.