Fact Check: Elon Musk glorifies China’s growth, DeepSeek on podcast? No, Musk’s voice is AI-generated

Hyderabad: A video of American businessman Elon Musk—popular now for the role he played in winning Donald Trump the US presidential seat—has been circulating on social media. In the clip, Musk appears on screen for the first 0.9 seconds, followed by an alleged audio recording in which he criticises Britain and speaks on China’s growth.

In the audio, he is allegedly heard saying, “It is not really about history; it is about reality. Britain fought nearly everyone but is not hated anymore. China, on the other hand, is simply minding its own business, growing rapidly, and yet, suddenly everyone is mad. Why? Because when China grows, it messes with other people’s profits.

Take Artificial Intelligence, China’s DeepSeek model blew up in days on Wall Street, sending markets into total panic mode. Microsoft dropped 3.5 per cent, Amazon fell 0.24 per cent in the first hour, Oracle sank 8 per cent and Nvidia crashed a massive 17 per cent. In just one day, 589 billion dollars in market value vanished—the biggest single-day market wipeout ever.”

Musk is also allegedly heard speaking about manufacturing, claiming that China went from buying tunnel machines and pen tips to making them better and cheaper, crushing German and Western firms.

Comparing India’s growth and hitting at PM Modi, an X user shared the video and wrote, “Watch this and you will understand how Modi wasted 11 years of India even after getting an absolute majority. You need education and vision, setting aside caste, creed and Revadi. Jai Hind.” (Archive)

Fact Check

NewsMeter found that the claim is false, as the audio resembling Musk’s voice is AI-generated.

We conducted a keyword search for statements by Musk regarding China’s growth and criticism of Britain but found no relevant reports. However, in January 2025, Musk expressed scepticism about DeepSeek’s reported capabilities, suggesting that the company might have more Nvidia GPUs than publicly disclosed.

Upon a reverse image search of keyframes from the viral video featuring Elon Musk, we found an eight-hour-long podcast published on Lex Fridman’s YouTube channel on August 2, 2024, titled ‘Elon Musk: Neuralink and the Future of Humanity.’

The podcast featured a conversation with Elon Musk, DJ Seo, Matthew MacDougall, Bliss Chapman, and Noland Arbaugh about Neuralink and the future of humanity.

Musk appeared in the first one-and-a-half hours of the podcast, wearing the same T-shirt. The surroundings and background are also identical to those in the viral video.

The channel also provided a transcription of the podcast.

Musk discussed topics such as telepathy, the power of the human mind, the future of Neuralink, ayahuasca, merging with AI, xAI, Optimus, his approach to problem-solving, history and geopolitics, lessons of history, the collapse of empires, time, aliens and curiosity. However, we found no mention of him speaking about China, Britain or DeepSeek at any point.

Also, the podcast was released on August 2, 2024, which predates the release of DeepSeek on January 20, 2025.

We also found another podcast by Lex Fridman featuring Elon Musk, published on November 9, 2023, titled ‘Elon Musk: War, AI, Aliens, Politics, Physics, Video Games, and Humanity.’

In this podcast, using its transcript as the reference, we found that Musk spoke about China but did not criticise Britain in comparison to China, nor did he mention DeepSeek. He discussed global power dynamics, drawing parallels between China’s economic rise and historical cases where dominant powers clashed with emerging ones, referencing the Thucydides Trap. Musk warned that China’s economy could grow to be two or three times larger than that of the United States, posing a significant challenge to the existing global order.

However, where did the audio come from?

We extracted the audio from the viral clip and analysed it using AI detection tools. The Hiya Deepfake Voice Detector rated the sample as highly likely to be a deepfake, giving it an authenticity score of just 1 out of 100. Similarly, Hive Moderation assessed the audio as likely AI-generated, assigning it an aggregate confidence score of 99.9 per cent.