Palmer Luckey, the American entrepreneur who sold his AR/VR hardware company Oculus to Facebook and founded defence technology comany Anduril, is skeptical of the recent hype surrounding Chinese AI startup DeepSeek and its claims about the cost of training its AI models. In a Fox Business interview, Luckey criticized the US media for what he perceives as overhype creating reporting of DeepSeek’s claims.

“Useful idiots” in America not challenging DeepSeek claims

Luckey specifically challenged the widely reported figure of $5 million as the training cost for DeepSeek’s AI model, which is purportedly competitive with US-developed models costing billions.He pointed out that DeepSeek has not disclosed the full costs associated with developing both of its models, and that significant infrastructure costs remain unknown. He accused DeepSeek of intentionally publicizing the $5 million figure to damage US companies, with the media acting as “useful idiots” by repeating the claim without proper scrutiny. He argued that neither China, DeepSeek, nor the media have an incentive to correct the record, even as US companies like Nvidia saw their valuations plummet.
DeepSeek’s model reportedly matches the capabilities of chatbots like ChatGPT and Meta’s Llama at a significantly lower development cost. Its app also topped the Apple App Store charts, and it reportedly functions using less powerful Nvidia chips. These developments triggered a downturn in the US tech sector earlier this week.

Chinese AI does not require “losing our minds”

While acknowledging DeepSeek’s impressive progress and innovations in AI, Luckey cautioned against accepting the company’s self-reported figures at face value. He believes there are those who actively want to see US tech companies, and US President Donald Trump fail, and that some of those individuals are located within the United States.
Luckey suggested that the timing and presentation of DeepSeek’s news were deliberate, and that the resulting stock market reactions indicate it achieved its intended purpose. He concluded by emphasizing that acknowledging the competitive threat of Chinese AI does not require “losing our minds” or falling prey to what he considers to be CCP propaganda.